UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effect of a dopamine antagonist and an agonist on rats’ perception of reward quantity : an examination… Martin-Iverson, Mathew Thomas 1985

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1985_A1 M37_8.pdf [ 9.74MB ]
Metadata
JSON: 831-1.0096734.json
JSON-LD: 831-1.0096734-ld.json
RDF/XML (Pretty): 831-1.0096734-rdf.xml
RDF/JSON: 831-1.0096734-rdf.json
Turtle: 831-1.0096734-turtle.txt
N-Triples: 831-1.0096734-rdf-ntriples.txt
Original Record: 831-1.0096734-source.json
Full Text
831-1.0096734-fulltext.txt
Citation
831-1.0096734.ris

Full Text

THE EFFECT OF A DOPAMINE ANTAGONIST AND AN AGONIST ON RATS' PERCEPTION OF REWARD QUANTITY: AN EXAMINATION OF THE ANHEDONIA HYPOTHESIS by MATHEW THOMAS MARTIN-IVERSON B . S c . U n i v e r s i t y of A l b e r t a , 1980 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n THE FACULTY OF GRADUATE STUDIES (Neurosc ience ) We a c c e p t t h i s t h e s i s as conforming to the r e q u i r e d s t a n d a r d THE UNIVERSITY OF BRITISH COLUMBIA Augus t , 1985 © Mathew Thomas M a r t i n - I v e r s o n , 1985 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of ^yy/^/o^/ccJ Sr/d^r^2j\ fcvck/h.//^ The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date /fa ?J . /4f.5 / f i l l Supervisor: Dr. Hans C. Fibiger ABSTRACT A procedure was developed to determine the e f f e c t of a dopamine (DA) antagonist (haloperidol) and a DA agonist (d-amphetamine) on rats' perceptions of the hedonic value of food. Eighteen rats were trained to discriminate between two quantities of sweet food p e l l e t s (1 and 4), i n a forced-choice two-lever successive discrimination procedure. To control for non-specific perceptual e f f e c t s of the treatments, the rats were also trained to discriminate between 1 and 4 tones. It was established that rats attended to the value of food, as well as the proportional differences i n quantity, when discriminating food quantities. This was accomplished by a l t e r i n g the value of the food i n two ways. F i r s t l y , "hunger" was altered by changing the degree of food deprivation during testing. Secondly, unsweetened food p e l l e t s were introduced as probe cues. These two methods of a l t e r i n g the value of food p e l l e t s were u t i l i z e d while quantity generalization gradients were determined, by presenting animals with 1,2., 3 and 4 numbers of stimuli as probe cues. Two measures were derived from these generalization gradients: the point of subjective equality (PSE), which i s the calculated number of s t i m u l i that would maintain responses equally d i s t r i b u t e d between the two levers, and the slope of the gradient. The PSE primarily r e f l e c t s perceptual processes, while the slopes of the gradients provide an index of performance impairment. It was observed that decreasing the value of food by i i i e ither decreasing food deprivation or reducing the sweetness of the food p e l l e t s resulted i n the rats perceiving a given quantity of food as larger than before these treatments (decreased the food PSE). Neither a l t e r i n g food deprivation nor introducing novel tone probes had an e f f e c t on the numerical attributes of tones, as r e f l e c t e d by the tone PSE. Haloperidol (0.030, 0.50 and 0.083 mg/kg, i.p.) produced a s t a t i s t i c a l l y s i g n i f i c a n t , but s l i g h t dose-dependent performance d e f i c i t , as r e f l e c t e d by the slope of the generalization gradients. I t did not af f e c t the perception of food p e l l e t quantities at any dose, as r e f l e c t e d by the food PSE. Haloperidol decreased the number of tones a given quantity was perceived as by rats (increased the tone PSE). Amphetamine (0.25, 0.50 and 1.0 mg/kg, i.p.) decreased the perception of a given quantity of food (increased the food PSE) i n a dose-dependent manner, without a s i g n i f i c a n t e f f e c t on performance. Thus, amphetamine enhanced the hedonic value of food. Amphetamine also increased rats' perceptions of a given number of tones (decreased the tone PSE). It therefore appears that while d-amphetamine can enhance the perceived hedonic value of food, haloperidol has no e f f e c t on rats' perceptions of the hedonic value of food. Furthermore, evidence that DA systems are involved i n the mechanism of an "in t e r n a l clock" or "counter" was obtained. i v TABLE OF CONTENTS Abstract i i Table of Contents i v L i s t of Tables v i i i L i s t of Figures ix Acknowledgment x i i INTRODUCTION 1 Theories of Appetitive Rewards 1 Theories of Extinction 7 H i s t o r i c a l Developments i n the Physiology of Reinforcement and Motivation 9 Motivation: Role of Dopamine 10 ICSS: Role of Dopamine 11 Reinforcing Actions of DA agonists 18 Neuroleptics: The Anhedonia Hypothesis 20 S i m i l a r i t y of Effects of Neuroleptics to Reward Omission 23 Neuroleptic Effects on Incentive Motivation 27 Neuroleptic Effects on Stimulus Control 30 Neuroleptic Effects on Timing and Counting 3 2 Neuroleptic Effects on Rate-Independent Measures of Reinforcer E f f i c a c y 34 Response-Specific Effects of Neuroleptics 39 Neuroleptic Effects on Responses D i f f e r i n g i n E f f o r t : A Test of the Perceived E f f o r t Hypothesis 40 Summary 41 Rationale 43 V METHODS 46 Subjects 46 Apparatus 46 Drug Treatments 47 Order of Experiments . .. 48 S t a t i s t i c s 48 EXPERIMENT 1: Preferences for d i f f e r e n t quantities of food p e l l e t s as assessed by a free-operant choice procedure 49 Experiment 2: E f f e c t of pre-loading rats with food on preferences for d i f f e r e n t quantities of food p e l l e t s 55 Experiment 3: Food and tone quantity discriminations 62 Experiment 4: The e f f e c t of manipulations of motivational state on perceptions of quantities of food and tone cues 85 Experiment 5: A free-operant choice procedure assessing preferences for food d i f f e r i n g i n taste but not quantity 112 Experiment 6: E f f e c t of body weight on preferences for sweetened versus unsweetened food p e l l e t s .... 118 Experiment 7: Generalization of food quantity discriminations to food d i f f e r i n g only i n taste and hedonic value 125 Experiment 8: Generalization of tone quantity discrimination to a novel tone 139 Experiment 9: Reinforcement properties of tone st i m u l i 143 v i Experiment 10: Role of p e l l e t dispenser " c l i c k s " i n p e l l e t discriminations 148 Experiment 11: The e f f e c t of a dopamine agonist and a dopamine antagonist on the generalization gradients of food p e l l e t s and tones 153 GENERAL DISCUSSION 188 Rationale 188 Summary of Experiments 190 Experiment 1 190 Experiment 2 190 Experiment 3 191 Experiment 4 192 Interpretation of Generalization Gradients 192 I n i t i a l Predictions 194 Obtained Forms of Generalization Gradients 194 Effects of Food Deprivation on Quantity Perceptions 196 Model of Food Quantity Perceptions 197 Experiment 5 199 Experiment 6 200 Experiment 7 201 Experiment 8 202 Experiment 9 202 Experiment 10 203 v i i Experiment 11 204 Conclusions 205 Theoretical Implications 211 The Progressive Nature of Neuroleptic-Induced Suppression of Behavior: Some Speculations 219 BIBLIOGRAPHY 223 v i i i L i s t of Tables TABLE I: E f f e c t of haloperidol and amphetamine on the PSE derived from food p e l l e t and tone quantity generalization gradients 161 TABLE I I : E f f e c t of haloperidol and amphetamine on the slopes derived from food p e l l e t and tone quantity generalization gradients 162 i x L i s t of Figures Figure 1: Preferences between 4 and 1 food p e l l e t s as assessed by a free-operant choice procedure i n rats 52 Figure 2: E f f e c t of pre-loading rats with d i f f e r e n t quantities of food on preferences for 4 p e l l e t s over 1 p e l l e t 57 Figure 3: E f f e c t of pre-loading rats with d i f f e r e n t quantities of food on t o t a l responses 59 Figure 4: Acquisition and asymptotic performance of two forced-choice quantity discriminations ... 68 Figure 5: Response latencies for quantity discriminations with cues consisting of 1 or 4 food p e l l e t s or tones 75 Figure 6: Response bias as indicated by the proportion of responses to the lever cued by the 4 p e l l e t stimulus during i n t e r - t r i a l i n t e r v a l s 76 Figure 7: Food p e l l e t generalization gradients determined while rats were under one of three food deprivation conditions 93 Figure 8: Tone generalization gradients determined while rats were under one of three food deprivation conditions 94 Figure 9: PSE derived from generalization gradients while rats were under one of three food deprivation conditions 95 X Figure 10: Slopes derived from generalization gradients while rats were under one of three food deprivation conditions 98 Figure 11: Response latencies on t r i a l s conducted while rats were on one of three food deprivation conditions 100 Figure 12: Responses during i n t e r - t r i a l i n t e r v a l s while rats were on one of three food deprivation conditions 103 Figure 13: Percent choice for sweet p e l l e t s over unsweetened p e l l e t s 115 Figure 14: E f f e c t of reducing quantity of sweet p e l l e t s on choice 121 Figure 15: E f f e c t of a l t e r i n g food deprivation condition on choice for sweetened and unsweetened food p e l l e t s 122 Figure 16: Generalization gradients obtained from presentation of various quantities of sweet and unsweetened food p e l l e t s 130 Figure 17: Response latencies during t r i a l s i n which various quantities of sweet and unsweetened food p e l l e t s were presented 132 Figure 18: Generalization gradients obtained from f i r s t 2 sessions of sweet and unsweetened p e l l e t probes 134 x i Figure 19: Generalization gradients obtained from the presentation of various quantities of regular and novel tones 142 Figure 20: Responses to levers during two 30 min free-operant tests for reward properties of tones 146 Figure 21: E f f e c t of haloperidol on food p e l l e t generalization gradients 157 Figure 22: Ef f e c t of haloperidol on tone generalization gradients 158 Figure 23: E f f e c t of haloperidol on response latencies on food generalization t r i a l s 165 Figure 24: E f f e c t of haloperidol on response latencies on tone generalization t r i a l s . . 169 Figure 25: E f f e c t of haloperidol on i n t e r - t r i a l i n t e r v a l responses 172 Figure 26: E f f e c t of d-amphetamine on food p e l l e t generalization gradients 174 Figure 27: E f f e c t of d-amphetamine on tone generalization gradients 175 Figure 28: E f f e c t of d-amphetamine on response latencies on food p e l l e t t r i a l s 178 Figure 29: E f f e c t of d-amphetamine on response latencies during tone generalization t r i a l s .... 181 Figure 30: E f f e c t of d-amphetamine on i n t e r - t r i a l i n t e r v a l responses 182 x i i ACKNOWLEDGMENTS I would l i k e to express my deepest appreciation for my advisor, Dr. Chris Fibiger, who has taught me the fin e a r t of c r i t i c a l reading of the l i t e r a t u r e , the f i n e r art of writing, and perhaps the f i n e s t art of knowing what experiments are l i k e l y worthwhile to pursue. I am gr a t e f u l to Dr. Don Wilkie, who came up with an excellent suggestion when approached with the problem of di s t i n g u i s h i n g hedonic from motor processes, and for his helpf u l comments i n r e l a t i o n to the parameters for the discrimination procedure. My thanks and appreciation are extended to Jonathan Druhan, whose stimulating conversation was largely responsible for any coherence that may be i n the inter p r e t a t i o n of the present r e s u l t s . I would also l i k e to thank Carolyn Szostak and Janet Finlay for th e i r a i d when I needed i t most, and John Maclennan, who often retrieved my mind from the Great Beyond during the wee, small hours of the morning. F i n a l l y , I thank the remaining members of my thesis advisory committee, Drs. E. McGeer, A.G. P h i l l i p s and J. Steeves. 1 INTRODUCTION Reinforcement i s a central concept i n the s c i e n t i f i c study of behavior. The physiological mechanisms that underlie reinforcement have therefore been the subject of a great deal of in t e r e s t . Research with e l e c t r i c a l brain stimulation and with drugs that either f a c i l i t a t e or antagonize central catecholamine transmission have provided highly suggestive, but controversial, clues to the i d e n t i f i c a t i o n of possible b i o l o g i c a l substrates of reinforcement. Some of these clues involve physiologic and pharmacologic manipulations that appear to mimic natural reinforcers ( e l e c t r i c a l brain stimulation and catecholamine agonists). Others have come from pharmacologic manipulations (neuroleptics) during operant testing that produce behavioral results that resemble to some extent omission of reward (extinction). Before reviewing t h i s research, a b r i e f survey of theories of reinforcement and extinction i s necessary to understand the behavioral processes that neurobiologists have attempted to understand within a neurological framework. Theories of Appetitive Rewards Reinforcement i s an operation that increases the pr o b a b i l i t y of the occurrence of ce r t a i n classes of behavior. A reward i s an event that increases the pr o b a b i l i t y of a behavior that bears a p a r t i c u l a r rel a t i o n s h i p with that event. This relationship may be purely temporal: the reward may be contiguous with the 2 behavior. A l t e r n a t i v e l y , and more commonly within laboratory situations, the reward bears a causal r e l a t i o n to the behavior: the reward i s contingent upon the instrumental response. Why the pr o b a b i l i t y of a behavior i s increased by a contiguous and/or contingent event i s not clear. Extreme response theories of reinforcement are based on the assumption that when any two behaviors occur i n sequence, the p r o b a b i l i t y that they w i l l occur i n the same sequence i s increased, regardless of the b i o l o g i c a l value of either behaviors (Denny and Adelman, 1955). Less extreme response theories are predicated on the notion that only a pa r t i c u l a r class of behavior has the property of increasing the p r o b a b i l i t y of the occurrence of other behaviors. S h e f f i e l d (1966) proposed that consummatory behaviors have th i s property. A more b i o l o g i c a l l y - o r i e n t e d response theory of reinforcement postulates that some l e v e l of a c t i v a t i o n of the neural substrates underlying consummatory behaviors i s s u f f i c i e n t to increase the p r o b a b i l i t y of behavior (Glickman and S c h i f f , 1967), even i f i t i s i n s u f f i c i e n t to e l i c i t the response i t s e l f . Early motivational theories of reinforcement assumed that there are i n t e r n a l states (drives), related i n some way to the homeostatic needs of an animal, that " i r r i t a t e " the animal (Hull, 1943). Behaviors that reduce t h i s i r r i t a t i o n are more l i k e l y to be repeated the next time a similar i r r i t a t i o n occurs. A more recent motivational theory i s the central motive state (CMS) concept of Bindra (1969). In t h i s 3 theory, physiological states interact with incentive s t i m u l i to produce an increased attention and s e n s i t i v i t y to pa r t i c u l a r classes of stimuli and an increased p r o b a b i l i t y to emit c e r t a i n broad classes of species-typical behaviors. This CMS i s assumed to occur by actions of the incentive stimuli and the physiological state on a common, c e n t r a l l y located, set of neurons. A key feature of thi s theory i s that a f f e c t (emotion) and motivation are considered to r e f l e c t a common process. Furthermore, a CMS can be e l i c i t e d by a previously neutral stimulus, providing that the incentive stimulus (reward) i s contingent upon the neutral stimulus, and an animal i s repeatedly exposed to t h i s contingency. Researchers interested i n the physiology of rewards have been attracted by stimulus theories of reinforcement i n recent years. An early proponent of a stimulus theory was Young (1966, 1977). He assumed that stimuli produce a f f e c t i v e responses i n animals. A p a r t i c u l a r response i s , to some degree, either p o s i t i v e (approach), neutral or negative (withdrawal). A po s i t i v e a f f e c t i v e response i s determined by two orthogonal properties of a stimulus: a c t i v a t i o n (the "excitement" produced by the stimulus) and hedonia (the evaluation of a stimulus along a pleasant-unpleasant continuum). Cabanac (1971) has suggested that the value (pleasantness) of a stimulus i s related to i t s b i o l o g i c a l u t i l i t y , or at least to i t s expected b i o l o g i c a l u t i l i t y . Young proposed that the evaluation of a stimulus i s the 4 r e s u l t of competition between excitatory and i n h i b i t o r y inputs into a u n i f i e d neural system. As w i l l be seen, many s c i e n t i s t s investigating the neurological substrates of reinforcement are attempting to i d e n t i f y either t h i s neural system, or those underlying central motive states. Doubt has been cast on the v a l i d i t y of the concept of unitary reinforcement systems. Evidence has accumulated that has established the existence of b i o l o g i c a l constraints on the e f f i c a c y of rewards. For example, Shettleworth (1973) observed that food would reinforce s p e c i f i c food-acquisition related behaviors i n the golden hamsters, but not behaviors without relevance to foraging, such as grooming. Likewise, Sevenster (1973) found that the opportunity to attack a male stickleback would reinforce rod-biting i n male sticklebacks, but that the opportunity to court a female stickleback, while r e i n f o r c i n g some behaviors, was an i n e f f e c t i v e reinforcer of rod-biting. Whether an event i s a reward or not may depend upon i t s functional relevance with regard to the s p e c i f i c instrumental behavior. Thus, rather than ac t i v a t i n g a central neural system that increases the pr o b a b i l i t y of a preceding behavior, "rewards" may be events with p a r t i c u l a r inherent relationships with s p e c i f i c behaviors. Another concept of reinforcement evolved from response theories of reward f i r s t introduced by Premack (1959, 1965, 1971). His o r i g i n a l hypothesis was that any p a r t i c u l a r behavior has an assigned value that may be assessed by 5 measuring the amount of time an animal engages i n that behavior. A behavior with greater value than a second behavior w i l l act as a reinforcer for that behavior. Timberlake and A l l i s o n (1974) elaborated upon th i s i n i t i a l hypothesis and proposed a "response deprivation" theory of instrumental conditioning. In t h i s scheme, the r e l a t i v e value of a behavior does not determine whether or not i t i s rewarding. Rather, an instrumental contingency that reduces the duration of the contingent behavior below i t s optimal l e v e l , determined by pre-contingency baseline testing, produces reinforcement. Thus, a less valued behavior may reinforce a more valued behavior i f the contingency i s arranged such that baseline lev e l s of the more valued behavior produces less than optimal durations of the less valued response. Heth and Warren (1978) provided data to support an extension of the response deprivation hypothesis to include stimulus deprivation. An event, whether a stimulus or a behavior, i s rewarding i f the contingency between the instrumental response and the event reduces the duration of the event with baseline levels of the instrumental response. Hanson and Timberlake (1983) developed a general behavior-regulation model of instrumental behavior related to Timberlake's e a r l i e r response deprivation hypothesis. A major assumption of t h i s model i s that events (exposure to a stimulus or performance of a behavior) are regulated independently i n r e l a t i o n to some set-point. This set-point 6 can be determined by the measurement of the events i n question i n the r e l a t i v e absence of environmental constraints (contingencies). The imposition of a contingency between the instrumental response and the rewarding event "cross-couples the environmental e f f e c t s of regulatory systems underlying the instrumental and contingent responses, thereby challenging t h e i r set points". Thus, reinforcement i s interpreted as a compromise between the deviations from set points forced by the constraint function (environmental contingency). While Hanson and Timberlake's model appears to be successful at predicting le v e l s of instrumental responding under t y p i c a l operant conditions, i t does not address the problem of "response relevancy". However, unlike theories of a unitary reinforcement process, the fact that p a r t i c u l a r responses and "rewards" are incompatible i s not a serious , threat to the v a l i d i t y of the behavior-regulation model. A key feature of the model i s that events are assumed to be independently regulated with respect to t h e i r own set-points. It may be that the establishment of behavioral set-points i s not an independent process. Thus, the set-point for courting i n sticklebacks may be inversely related to the set-point for aggressive behavior. Hanson and Timberlake were not concerned with the processes underlying the establishment of behavioral set-points. However, the inves t i g a t i o n of the neurological substrates of reward may 7 be conceived as an attempt to determine the processes underlying the establishment of such set-points. Theories of E x t i n c t i o n Extinction refers to the gradual decrease i n the p r o b a b i l i t y of instrumental behavior after omission of reward. Three main types of theories have been proposed to account for t h i s phenomenon: generalization decrement, response interference, and i n h i b i t o r y learning. Generalization decrement theory assumes that the omission of reward during extinc t i o n changes the stimulus conditions from those under which learning o r i g i n a l l y occurred (Capaldi, 1966). During a c q u i s i t i o n , responding occurs i n the presence of memories of obtaining rewards. I n i t i a l l y during extinction, memories of previous rewards are active, and responding i s near the l e v e l observed when rewards are not omitted. However, as the extinction session progresses, memories of rewards are gradually replaced by memories that do not have anything to do with rewards. As extinct i o n sessions have no reward s t i m u l i , and as memories of previous rewards are replaced, the stimulus s i t u a t i o n becomes gradually more d i f f e r e n t from that e x i s t i n g during a c q u i s i t i o n , and t h i s produces a generalization decrement, reducing the p r o b a b i l i t y of responding. The main support of t h i s theory i s that extinction rate i s elevated by changing stimulus conditions (see Mackintosh, 1974 for a discussion of t h i s evidence). 8 Interference theories of extinction postulate that extinct i o n i s associated with some sets of responses that i n i t i a l l y compete with, and eventually replace, the instrumental behavior. The most popular interference theory suggests that omission of reward e l i c i t s emotional responses ( f r u s t r a t i o n - r e l a t e d responses), and these behaviors compete with instrumental behaviors (Amsel, 1958). The main problem with t h i s theory i s that f r u s t r a t i v e behaviors are transient; they usually occur only during the i n i t i a l periods of extinct i o n . If f r u s t r a t i v e behaviors are the only competing responses, then as the extinction session progresses, instrumental behavior should re-appear. Inhibitory learning as an explanation of extincti o n has a long history (Pavlov, 1927; Hu l l , 1943), but these early formulations have l i t t l e to recommend them (cf. Mackintosh, 1974). Perhaps the best concept of i n h i b i t o r y learning i s that of "disconfirmation" of expectancies (Mackintosh, 1974). Animals are assumed to learn the r e l a t i o n between th e i r instrumental behavior and rewards: they expect that t h e i r behavior w i l l produce a s p e c i f i c outcome. During extinction, t h i s expectancy i s no longer confirmed, and so animals learn the new relat i o n s h i p between t h e i r behavior and rewards, and come not to expect the reward. As discussed by Mackintosh (1974), i t i s l i k e l y that a l l of these theories of extincti o n apply to some degree. Thus, any int e r p r e t a t i o n of an a r t i f i c i a l state resembling reward omission should address a l l of these p o s s i b i l i t i e s . 9 H i s t o r i c a l Developments i n the Physiology  of Reinforcement and Motivation The search for a neurological substrate of reinforce-ment began with the discovery that rats w i l l learn to respond for e l e c t r i c a l stimulation of the brain (Olds and Milner, 1954). The fact that animals would work for e l e c t r i c a l stimulation of s p e c i f i c s i t e s i n the brain lent credence to a view of a unitary reinforcement process. Concomitantly, observations that lesions of the ventromedial area of the hypothalamus produce obesity i n rats (Hetherington and Ranson, 1942), while lesions of the l a t e r a l hypothalamus produce aphagia (Anand and Brobeck, 1951), were integrated by S t e l l a r (1954) into a model of two in t e r a c t i n g "centers", one regulating s a t i e t y and the other subserving hunger. This model aroused inte r e s t i n the neurological mechanisms of s p e c i f i c "drives". Thus, two independent l i n e s of research were begun i n the early n i n e t e e n - f i f t i e s : one investigating the neural substrates of a unitary reinforcement process, and the other devoted to elucidating brain mechanisms responsible for motivational responses to s p e c i f i c challenges to an animal's homeostasis. During the next decade, these two separate l i n e s of research came to focus upon the role of a single neurotransmitter, the catecholamine noradrenaline (NA). Grossman injected drugs d i r e c t l y into the l a t e r a l hypothalamus and found that NA e l i c i t e d feeding (Grossman, 1960, 1962a), and that adrenergic antagonists i n h i b i t e d food 10 consumption (Grossman, 1962b), possibly by acting on the substrates of hunger. Stein (1962, 1964) observed that amphetamine, which was known to release NA, increased i n t r a c r a n i a l s e l f - s t i m u l a t i o n (ICSS), and t h i s e f f e c t was shared with NA i t s e l f when d i r e c t l y infused into the brain (Wise and Stein, 1969). I t was also noted that both amphetamine administration and ICSS produce a release of endogenous NA (Stein and Wise, 1969). These data indicated that central NA neurons were a l i k e l y substrate for the rewarding properties of e l e c t r i c a l brain stimulation. Thus, the same neurotransmitter which appeared to mediate a hunger drive was also a l i k e l y candidate for the central substrate for reinforcement. Furthermore, the same brain s i t e seemed to be involved: the l a t e r a l hypothalamus. Motivation:  Role of Dopamine The advent of a neurotoxin that s p e c i f i c a l l y destroys monoaminergic neurons (Ungerstedt, 1968) altered opinions of what constituted the c r i t i c a l neural pathways underlying the presumably motivational impairments (aphagia and adipsia) c h a r a c t e r i s t i c of l a t e r a l hypothalamic lesions. Application of 6-hydroxydopamine (6-OHDA) to the ascending dopamine (DA) projection that courses through the l a t e r a l hypothalamic area r e s u l t s i n behavioral impairments that are i n many respects indistinguishable from the l a t e r a l hypothalamic syndrome (Ungerstedt, 1971). Thus, the motivational d e f i c i t s appeared to be related to a second catecholamine system, the 11 n i g r o s t r i a t a l DA pathway, rather than to NA terminals within the l a t e r a l hypothalamus. Analysis of the behavioral impairments associated with e l e c t r o l y t i c l a t e r a l hypothalamic lesions or 6-OHDA lesions of the n i g r o s t r i a t a l projection led Marshall and Teitelbaum (1977) to a t t r i b u t e the d e f i c i t s to an uncoupling of exteroceptive s t i m u l i from t h e i r normal behavioral sequelae, rather than to interference with a s p e c i f i c drive. This interpretation i s consistent with a s h i f t i n view i n the psychological l i t e r a t u r e from drive reduction theories to incentive motivational theories, i n which a n t i c i p a t i o n of rewards, instigated by exteroceptive s t i m u l i having both response a c t i v a t i n g and d i r e c t i n g properties, are conceived as playing an important role i n determining behavior (Bindra, 1969; Beck, 1978; K i l l e e n , 1982). The fact that appropriate responses can be e l i c i t e d i n lesioned animals given s u f f i c i e n t l y arousing situations suggests that the role of DA i n motivation may be related to one or both of the two functions that Bindra (1969) ascribed to his central motive state: s e l e c t i v e attention and motor f a c i l i t a t i o n . ICSS: Role of Dopamine Some early mapping studies of p o s i t i v e s i t e s of ICSS revealed that areas near to DA perikarya or terminal f i e l d s support ICSS (Crow, 1972; P h i l l i p s and Fibiger, 1973; P h i l l i p s , Brooke and Fibiger, 1975; P h i l l i p s , Carter and Fibiger, 1976a; Mora, P h i l l i p s , Koolhaas and R o l l s , 1976). 12 While t h i s evidence i s purely c o r r e l a t i v e , and therefore does not es t a b l i s h a role for DA i n ICSS, such a role has been indicated by research u t i l i z i n g s e l e c t i v e lesioning and pharmacological techniques. Catecholamine depletions produced with i n t r a v e n t r i c u l a r i n j e c t i o n s of 6-OHDA attenuate s e l f - s t i m u l a t i o n from the l a t e r a l hypothalamus ( P h i l l i p s and Fibiger, 1976). Moreover, both neuroleptics, which block DA receptors, (Fouriezos, Hansson . and Wise, 1978) and i n t r a v e n t r i c u l a r i n j e c t i o n s of 6-OHDA attenuate ICSS from the l a t e r a l hypothalamus (Fibiger, Carter and P h i l l i p s , 1976). 6-OHDA lesions of the substantia nigra, which sends a DA projection to the caudate-putamen attenuate ICSS supported by electrodes placed within the caudate-putamen, even when NA-containing terminals are protected from the ef f e c t s of the neurotoxin ( P h i l l i p s , Carter and Fibiger, 1977). Lesions of the ascending DA projections a r i s i n g from the ventral tegmental area of Tsai (VTA) produced by 6-OHDA i n rats, without concomitant damage to NA pathways, also suppress ICSS supported by electrodes within the VTA ( P h i l l i p s and Fibiger, 1978). Furthermore, ICSS supported by electrodes i n the locus coeruleus, hippocampus and o l f a c t o r y bulb are not affected by NA depletions produced with 6-OHDA (Clavier, Fibiger and P h i l l i p s , 1976; P h i l l i p s , Van Der Kooy and Fibiger, 1977). As DA depletions and neuroleptic treatments produce motor impairments and sedative e f f e c t s , i t i s c r i t i c a l to determine whether suppression of ICSS i s due to motor 13 e f f e c t s or r e f l e c t s a reduction i n reward value. Addressing t h i s possible confound, P h i l l i p s , Carter and Fibiger (1977) investigated the e f f e c t s of u n i l a t e r a l substantia nigra lesions induced with 6-OHDA i n rats pretreated with desimipramine on ICSS supported by electrodes i n the caudate-putamen either i p s i - or c o n t r a - l a t e r a l to the le s i o n s i t e . Reductions i n ICSS were observed i n rats with electrode placements i p s i - but not c o n t r a - l a t e r a l to the les i o n s i t e , r u l i n g out non-specific response impairments as a cause of the decrease i n ICSS. Fouriezos and Wise (1976) observed that administration of the DA receptor blocker, pimozide, produces a gradual decrease i n ICSS that bears a remarkable resemblance to that observed during extinction. Because of th i s resemblance, and because response rates are high at the beginning of a session, they suggested that motor or performance d e f i c i t s cannot account for the decreased ICSS. Further work has revealed that ICSS suppressed by pimozide can be re-instated by presentation of a stimulus that had previously been paired with reinforcement (Franklin and McCoy, 1979). This e f f e c t i s reminiscent of the increase i n responding during ext i n c t i o n by the presentation of secondary reinforcers (Bugelski, 1938). i. The e x t i n c t i o n - l i k e e f f e c t s of pimozide occur with doses that do not produce motor incapacitation and are task-s p e c i f i c ( G a l l i s t e l , Boytim, Gomita and Klebanoff, 1982), suggesting that response suppression i s the r e s u l t of an a l t e r a t i o n i n behavioral processes related to reward omission. These data might be interpreted as r e s u l t i n g from a generalization decrement, produced by the e f f e c t s of pimozide causing a major stimulus change. However, while possibly contributing to the e x t i n c t i o n - l i k e pattern of responding, assuming that the drug e f f e c t s remain r e l a t i v e l y constant i t cannot account for the progressive decrease i n responding. Pimozide has not been reported to e l i c i t p a r t i c u l a r behaviors that might compete with the instrumental responses, thus the remaining p o s s i b i l i t y i s that expectancies of reward are not confirmed i n the presence of the neuroleptic. The findings discussed above are also consistent with a view that neuroleptics, l i k e 6-OHDA lesions of the l a t e r a l hypothalamus, decrease the a b i l i t y of stimuli to activate appropriate responses. This could occur by decreasing either of the two functions of a central motive state: motor f a c i l i t a t i o n or an increase i n s e n s i t i v i t y to relevant s t i m u l i . Responding may be high i n the i n i t i a l segment of testing sessions because of the added a c t i v a t i o n of sensory stimulation due to handling and to environmental changes (Fouriezos and Wise, 1976). The e f f e c t s would be s p e c i f i c to the task associated with neuroleptic treatment because only the incentive s t i m u l i or the responses relevant to that task would be affected. Results obtained from rate-free techniques of s e l f -stimulation measurement have also indicated that DA plays a role i n the rewarding effects of e l e c t r i c a l stimulation of some brain s i t e s . Zarevics and Setler (1979) allowed rats to self-stimulate with electrodes implanted i n the l a t e r a l hypothalamus, but decreased the magnitude (amperage) of the stimulation with each successive lever response. Responding to a second lever would reset the stimulation to i t s o r i g i n a l magnitude. The stimulation amperage at which rats would respond to the reset lever provides a reasonably rate-free assessment of reward value. A DA receptor antagonist (pimozide) was found to increase t h i s threshold. Low doses of haloperidol (0.01-0.04 mg/kg) were found to have the same ef f e c t s as reductions i n stimulation i n t e n s i t y on another r e l a t i v e l y rate-free procedure (Wauquier, Clincke and Fransen, 1983). Pimozide was reported to increase the number of l a t e r a l hypothalamic stimulation pulses required to sustain responding at 50% of maximal rate (reward summation function) at doses which had minimal ef f e c t s on the asymptotic rate ( S t e l l a r , K e l l y and Corbett, 1983), providing further evidence of reward-attenuation. These investigators observed a s i m i l a r e f f e c t on the reward summation function when pimozide was infused d i r e c t l y into the nucleus accumbens septi (NAS), a major terminal f i e l d of DA-containing neurons with perikarya i n the VTA. G a l l i s t e l and Karras (1984) also reported an increase i n the reward summation function by neuroleptics, and they further demonstrated that amphetamine decreased the reward summation function. Another r e l a t i v e l y rate-independent procedure i s the conditioning of taste preferences with ICSS (Ettenberg, 1980). Pimozide was found to attenuate the conditioning of taste preferences with ICSS, and t h i s e f f e c t was not related to decreased responding for e l e c t r i c a l brain stimulation (Ettenberg and White, 1981). Thus, i t appears that ICSS at least from some s i t e s involves a DA component. These phenomena are not e a s i l y accounted for by a sensorimotor integration hypothesis, as responding i s either maintained at the same l e v e l i n the drug state (by increasing reinforcement magnitude), or the rate of at least one response measure i s a c t u a l l y increased (responding on reset l e v e r s ) . In the reset lever procedure i n p a r t i c u l a r , stimulus control over responding appears i n t a c t . The rate-enhancing actions of amphetamine on ICSS are c l e a r l y dependent upon i t s DA agonist e f f e c t s (Fibiger, 1978; Wise, 1978). Lesions of the ascending DA pathways with 6-OHDA i p s i l a t e r a l to an ICSS-supporting electrode i n the substantia nigra abolish the ICSS f a c i l i t a t i n g e f f e c t of amphetamine, while c o n t r a l a t e r a l lesions do not (Clavier and Fibiger, 1977). S p e c i f i c NA depletions do not a f f e c t the action of amphetamine on ICSS supported by electrodes i n the locus coeruleus, a nucleus containing NA but not DA c e l l bodies (Breese and Cooper, 1975). While the above evidence establishes that DA-releasing neurons are sometimes involved i n ICSS, debate has centered upon whether or not DA neurons are a c r i t i c a l l i n k i n a l l pathways subserving ICSS. Wise and Bozarth (1981, 1984), for example, have argued for the view that a c t i v a t i o n of DA neurons may be a c r i t i c a l l i n k i n the neural c i r c u i t r y underlying the actions of a l l p o s i t i v e r e i n f o r c e r s . Recent mapping of ICSS-supporting s i t e s within structures containing DA terminal f i e l d s have found that p o s i t i v e ICSS regions do not correspond with l o c a l DA terminal d i s t r i b u t i o n s (Prado-Alcala and Wise, 1984; Prado-Alc a l a , Streather and Wise, 1984). A rather precise evaluation of the e l e c t r o p h y s i o l o g i c a l c h a r a c t e r i s t i c s of the axons mediating ICSS ( G a l l i s t e l , Shizgal and Yeomans, 1981) appears to rule out dopaminergic axons as the relevant f i r s t order axons activated by rewarding stimulation. Thus, the DA hypothesis of ICSS does not require that ICSS-supporting regions contain DA axons, perikarya or terminals, as long as they are connected i n series with the stimulated axons. However, i t appears that increases i n DA metabolism do not always accompany ICSS (Mitch e l l , Nicolaou, Arbuthnott and Yates, 1982). It i s not yet clear i f there exists a single neuroanatomical system through which a l l reinforcers must ultimately act, or i f DA neurons form a c r i t i c a l l i n k i n that system. Evidence for p a r a l l e l reward pathways, not a l l of which involve DA components, has been reviewed recently ( P h i l l i p s , 1984). 18 Reinforcing Actions of DA Agonists Further evidence for the role of central DA-releasing terminals i n reinforcement was obtained following the development of the intravenous drug self-administration procedure. I t was found that animals would learn and sustain operant responses for intravenous injections of a variety of psychostimulants and other drugs that f a c i l i t a t e DA transmission (Pickens and Thompson, 1968; Johanson and Schuster, 1975; Wise, Yokel and deWit, 1976; Spyraki and Fibiger, 1981; Nielsen, Duda, Mokler and Moore, 1984; C o l l i n s , Weeks, Cooper, Good and Russell, 1984). Furthermore, self-administration of these drugs i s attenuated by neuroleptic treatment (Yokel and Wise, 1976; Ettenberg et a l . , 1982), or by 6-OHDA lesions of the NAS when the NA-releasing terminals were spared (Roberts, Corcoran and Fibiger, 1977; Lyness, Fried l e and Moore, 1979; Roberts, Koob, Klonoff and Fibiger, 1980). While the suppression of psychostimulant s e l f -administration produced by neuroleptics can be attributed to response impairments, the e f f e c t of low doses of neuroleptics have been interpreted as evidence against t h i s view. Neuroleptics given i n small doses increase the response rate for self-administration of psychostimulants (Yokel and Wise, 1973; De Wit and Wise, 1977). Response rates for psychostimulants also increase as the dose i s decreased (Pickens and Thompson, 1968; Wilson, Hitomi and Schuster, 1971). The e f f e c t of low doses of neuroleptics i s 19 therefore s u p e r f i c i a l l y similar to that of decreasing reinforcement magnitude, and i s not predicted by a non-s p e c i f i c motor impairment hypothesis. I t i s possible, however, that low doses of neuroleptics p r e f e r e n t i a l l y suppress stimulant-induced behaviors that compete with lever-responding, thereby increasing response rate (Woolverton, 1981). More recently, psychostimulants have proven to be e f f e c t i v e primary reinforcers i n the conditioned place preference procedure (Spyraki, Fibiger and P h i l l i p s , 1982a, 1982b; Martin-Iverson, Ortmann and Fibiger, 1985). However, the role of DA i n the r e i n f o r c i n g actions of psycho-stimulants as assessed by place preference conditioning i s not clear. While amphetamine-induced place preferences are attenuated by neuroleptic administration (Spyraki et a l . , 1982a), those conditioned with cocaine (Spyraki et a l . , 1982b), methylphenidate or nomifensine (Martin-Iverson, et a l . , 1985) are not affected by comparable doses of neuroleptics. H i l l (1970) provided evidence that psychomotor stimulants enhance the effectiveness of conditioned reinforcers ( i e . those reinforcers established by arranging a predictive r e l a t i o n s h i p between a previously neutral stimulus and a reward). This finding has been confirmed i n a number of tests by Robbins (Robbins, 1975; Robbins, Watson, Gaskin and Ennis, 1983; Taylor and Robbins, 1984) and Beninger (Beninger, Hanson and P h i l l i p s , 1980). However, the 20 in t e r p r e t a t i o n of these results i s not immediately c l e a r . They may r e f l e c t a f a c i l i t a t i o n by psychostimulants of a catecholamine-mediated process underlying conditioned reinforcement. A l t e r n a t i v e l y , the f a c i l i t a t i o n of conditioned reinforcement may occur because of an association of the previously neutral s t i m u l i with the d i r e c t r e i n f o r c i n g actions of psychostimulants. Given the apparent potency of the d i r e c t reinforcement properties of psychostimulants, and i n the i n t e r e s t s of parsimony, the l a t t e r view appears to be the most acceptable. Neuroleptics: The Anhedonia Hypothesis Evidence also points to a role of central DA systems i n the mediation of natural rewards. Pimozide produces an e x t i n c t i o n - l i k e pattern of behavior i n rats trained on either a lever-pressing task (CRF schedule) or i n a straight a l l e y running task for food reward (Wise, Spindler, DeWit and Gerber, 1978). Responding progressively decreased across succeeding tests with pimozide pre-treatment. The progressive nature of the d e f i c i t could not be attributed to drug accumulation, as animals treated i n t h e i r home cages for 3 days, but not p r i o r to the test sessions, when given pimozide before a test on the fourth day did not exhibit response rates d i f f e r e n t from those from animals receiving pimozide on the f i r s t test day. Furthermore, rats given 3 days of t e s t i n g with reward omission, and then rewarded on 21 the fourth test day i n the presence of pimozide, exhibited response rates similar to rats that were given another extinct i o n session on the fourth day, or 4 days of reward + pimozide. Thus, the e f f e c t s of non-reward could transfer to the pimozide treatment. The authors interpreted these findings as a r e s u l t of a "blunting" of the hedonic properties of rewards (anhedonia). It i s uncertain what i s meant by "blunting" hedonic properties of rewards. If "blunting" i s si m i l a r to a reduction i n reward magnitude, then i n a straight a l l e y run-way task a decrease i n speed of running and latencies to leave the s t a r t box would be expected (Crespi, 1942), and such e f f e c t s should be less extreme than those observed during reward omission. This i s exactly what Wise et a l . (1978) reported. However, i f neuroleptics produce an e f f e c t s i m i l a r to a reduction i n reward magnitude, then i t might be expected that a decrease i n response rate would be greatest on the f i r s t day of drug treatment, due to negative contrast e f f e c t s (Crespi, 1942; Mackintosh, 1974). This e f f e c t was c l e a r l y not observed. Moreover, reductions i n reward magnitude do not r e s u l t i n progressive decreases i n response rates i n rats trained on a lever-press task with a CRF schedule (Martin-Iverson and Fibiger, unpublished data). In fact, response rate either increased or remained constant after decreases i n reward magnitude, depending upon the procedure used to reduce reward magnitude (reward quantity or t a s t e ) . 22 Rats trained on i n t e r v a l schedules exhibit response rates proportional to reward magnitude, but a progressive decrease i n response rate i s not observed (Meltzer and Brahlek, 1968). Pigeons do not appear to vary response rate i n operant tasks as a function of reward magnitude (Neuringer, 1967). If reward i s reduced by using an intermittent reinforcement schedule (reducing the reward per response), response rates increase. Of course, s u f f i c i e n t "blunting" of hedonic value may re s u l t i n blocking of hedonia, and the results of the lever-press task resemble, at least s u p e r f i c i a l l y , reward omission. Why a neuroleptic may blunt hedonia i n one task and block i t i n another i s not clea r . The fact that rats continued to eat those p e l l e t s that they did earn suggested to Wise and his colleagues that a s a t i a t i o n i n t e r p r e t a t i o n i s inappropriate (a conclusion further supported by the observation that responding for non-satiating rewards i s also attenuated by neuroleptics -Wise, Spindler and Legault, 1978; Fouriezos and Wise, 1976). However, i f the hedonic properties of the p e l l e t s are blocked by pimozide, why do the rats eat them? If the hedonic properties are blunted, but not blocked, then why do the rats progressively reduce t h e i r response rates? A possible explanation i s that pimozide reduces the hedonic value to a much greater extent than the reductions i n reward magnitude t y p i c a l l y used i n reinforcer magnitude studies. S u f f i c i e n t value i s maintained to e l i c i t some feeding, but 23 not enough for animals to expend the extra e f f o r t to work for the food. The progressive nature of the response suppression may then be accounted for by ' v i r t u a l ' reward omission. The problem with t h i s account i s that i t assumes a r e s u l t that experimentation has not established - that a small amount of reward does not sustain responding. S i m i l a r i t y of E f f e c t s of  Neuroleptics to Reward Omission If Wise's interpretation i s correct, then combining neuroleptic treatment with extinction should have no greater e f f e c t than either alone. However, i f neuroleptics produce response suppression as a r e s u l t of some mechanism other than anhedonia, then neuroleptics may decrease responding during extin c t i o n . Such a depression of the rate of extinction was observed ( P h i l l i p s and Fibiger, 1979). Furthermore, response attenuation was observed before any reinforcers had been received when a v a r i a b l e - i n t e r v a l (VI) 4 minute reinforcement (food) schedule was used ( P h i l l i p s and Fibiger, 1979). Neither of these findings are predicted by the anhedonia theory of neuroleptic e f f e c t s , since the a n t i c i p a t i o n of reward i s assumed not to be altered u n t i l experience indicates otherwise. Another factor that decreases response p r o b a b i l i t y during extinc t i o n can be invoked to explain decreases i n response rate p r i o r to experience with the f i r s t reward. As discussed e a r l i e r , the s i m i l a r i t y of the stimulus conditions 24 during extinc t i o n to those during a c q u i s i t i o n determines, to some extent, the extinction response rate. Stimulus-like e f f e c t s of pimozide may produce a reduction i n responding early i n the session because of a generalization decrement. These experiments do not rule out the p o s s i b i l i t y that anhedonic e f f e c t s of neuroleptics occur i n conjunction with other, possibly motor, ef f e c t s which might contribute to a reduction i n responses early during a VI schedule. It i s also possible that neuroleptics reduce both primary and secondary reinforcement. In the i n i t i a l stages of extinction, only the primary reinforcers are absent. Therefore, pimozide + reward omission may produce a greater suppression of responding than reward omission alone because of the attenuation of secondary reinforcement. Against t h i s i n t e r p r e t a t i o n i s another i n t e r e s t i n g phenomenon observed during the VI-4 min schedule for food reward: responding tended to occur i n bursts aft e r each reinforcement. This suggests that the incentive e f f e c t s of food reinforcement were not impaired by neuroleptic treatment. Other evidence against the anhedonia theory of neuroleptics was provided by Tombaugh, Anisman and Tombaugh (1980). In t h i s series of experiments, i t was demonstrated that pimozide produced a more rapid decline i n responding than reward omission on each of three d i f f e r e n t p a r t i a l reinforcement schedules (variable i n t e r v a l , fixed i n t e r v a l and fi x e d r a t i o ) . Another report indicated that pimozide did not produce the same pattern of responding as extinction during FI schedules for either food or e l e c t r i c a l brain stimulation reward with respect to either post-reinforcement pauses or ICSS durations (Greenshaw, Sanger and Blackman, 1981). Faustman and Fowler (1981) observed that haloperidol produced responses of a s i g n i f i c a n t l y longer duration than those r e s u l t i n g from non-reward i n rats on a f i x e d - r a t i o schedule, but the e f f e c t s of extinc t i o n and neuroleptics on response duration were i d e n t i c a l i n a CRF schedule (Faustman and F u l l e r , 1982). Thus, neuroleptic treatment was not the same as extinct i o n with p a r t i a l reinforcement schedules. This, of course, can be explained by the occurrence of both anhedonic and motor e f f e c t s , and i s not a' serious argument against the view that neuroleptics can produce anhedonia. A serious challenge to the anhedonia hypothesis has been presented, however. One of the strongest points i n favor of the anhedonia theory i s that t r a n s f e r r i n g from extinction to pimozide produced a decrease i n behavior as though animals had been given another day of extinction. Mason, Beninger, Fibiger and P h i l l i p s (1980) and Tombaugh et a l . (1980) found that the reverse transfer did not occur: animals given pimozide before a number of sessions i n which the response contingency operated and then transferred to an extinct i o n condition (without pimozide) exhibited high rates of responding equivalent to that seen on the f i r s t day of extinc t i o n . That the transfer from one condition to the other i s asymmetrical suggests that the effects of pimozide are not i d e n t i c a l to non-reward. Similar results have been obtained by other investigators (Gerber, Sing and Wise, 1981; Beninger, 1982). However, another interpretation of the lack of symmetrical transfer has been proposed by Gerber et a l . (1981), who argued that symmetry of transfer may not occur i f pimozide has discriminable e f f e c t s other than those producing non-reward. Transfer occurs from extinct i o n to pimozide because the action of pimozide has a l l the properties of reward omission. Transfer from pimozide to extinction does not occur because pimozide produces multiple e f f e c t s , some of which are not produced by non-reward. Mason et a l . (1980) provided further evidence against the anhedonia hypothesis. While extinction resulted i n a decrease i n responses on a DRL ( d i f f e r e n t i a l reinforcement for low rates of responding) task, pimozide treatments did not. Furthermore, rats trained on a runway task i n which reinforcement was omitted from half of the t r i a l s produced a p a r t i a l reinforcement extinct i o n e f f e c t (PREE): rats ran down the a l l y faster during extinction then those trained without reinforcement omission. Pimozide, when given to rats before half of the ac q u i s i t i o n t r i a l s f a i l e d to produce a PREE. Thus, pimozide did not a f f e c t behavior i n a manner analogous to reward omission, even when the c r i t i c a l test was conducted during a drug-free period, i n which response ef f e c t s could not confound the r e s u l t s . A possible problem with t h i s experiment i s state-dependency. The hypothetical non-reward produced with pimozide may have been dissociated 27 from the extinction testing because of associations with drug cues. More recently, i t was reported that, following a sim i l a r procedure as that used'by Mason and his colleagues, haloperidol was successful at inducing resistance to extinction indistinguishable from PREE (Camp and Ettenberg, 1984). The reasons for the discrepancy between the two experiments i s not clear. I t may be that the e f f e c t s of pimozide are state-dependent, but those of haloperidol are not, but t h i s i s simply conjecture at th i s stage. The importance of c l a r i f y i n g t h i s discrepancy i s great; the p a r t i a l reinforcement extinction e f f e c t allows the t e s t i n g of the e f f e c t s of a neuroleptic i n the absence of any performance d e f i c i t s . Neuroleptic E f f e c t s on  Incentive Motivation Mason and his colleagues rep l i c a t e d the summation of the e f f e c t s on non-reward and pimozide, but further observed that the response suppression i n t h i s case was greater than with pimozide alone (Mason et a l . , 1980). This rules out an explanation based on attenuation of secondary reinforcement, because i f both primary and secondary reinforcement are blocked during pimozide treatment, then removing the food should have no e f f e c t on behavior. There i s additional evidence i n d i c a t i n g that the incentive properties of s t i m u l i are not affected by 28 neuroleptics. Pimozide, while impairing the a c q u i s i t i o n of a lever-press task u t i l i z i n g a retractable lever, f a i l e d to suppress responding to a retractable bar after a few t r a i n i n g sessions (Tombaugh, Tombaugh and Anisman, 1979). The authors attributed t h i s to the a c t i v a t i o n a l influences of the sensory s t i m u l i associated with intrusion of the bar. In addition, rats given doses of pimozide s u f f i c i e n t to produce catalepsy could be induced to respond by "priming" them with food p e l l e t s . In f a c t , hungry rats treated with pimozide tend to remain near a location i n a complex maze i n which they obtained food more so than do vehicle treated rats (Irwin, Tombaugh, Zacharko and Anisman, 1983). This suggests that neuroleptics might ac t u a l l y increase the re i n f o r c i n g value of food, p a r t i c u l a r l y i t s secondary re i n f o r c i n g properties. Furthermore, Szostak and Tombaugh (1981) observed that while pimozide impaired performance of task involving a fi x e d consecutive number schedule of reinforcement, such an impairment was absent i f the a v a i l a b i l i t y of reinforcement was signaled by a stimulus. An experiment designed s p e c i f i c a l l y to test whether pimozide attenuates secondary reinforcement was conducted by Tombaugh, Grandmaison and Zito (1982). Rats were exposed to a l i g h t i n the presence of food aft e r treatment with pimozide or vehicle. They were then tested, drug-free, for approach behavior directed towards the l i g h t . Rats exposed to the l i g h t - f o o d pairs while under the influence of pimozide approached the l i g h t as often as those which were 29 given the light-food pairings without pimozide treatment. That secondary reinforcement developed i n these rats indicates that primary reinforcement was also not impaired. A further experiment by these investigators revealed that pimozide does not block place preference conditioning induced with food. This i s i n contradiction to another report of attenuation of food-induced place preferences produced by haloperidol (Spyraki, Fibiger and P h i l l i p s , 1982c). Reasons for t h i s discrepancy may be related to the fact that the l a t t e r investigators associated food with a place to which animals had a strong i n i t i a l bias, while the former used apparatus that was r e l a t i v e l y neutral. More importantly, Spyraki et a l . tested animals that were satiated, while Tombaugh et a l . tested rats while hungry. I t i s possible that the differences i n state between conditioning phases and test, while i n s u f f i c i e n t to produce state-dependent learning when only one state variable i s changed, i s s u f f i c i e n t to produce state-dependent learning when both drive and drug condition are d i f f e r e n t . In addition, food was presented to the animals i n d i f f e r e n t fashions i n the two experiments. While Spyraki et a l . exposed t h e i r rats to a large quantity of food throughout the conditioning sessions, Tombaugh et a l . presented small quantities of food at discrete i n t e r v a l s , accompanied by the noise of a p e l l e t dispenser. Further evidence against neuroleptic action on incentive motivation has come from the ICSS l i t e r a t u r e . 30 Wasserman, Gomita and G a l l i s t e l (1982) found that while pimozide produced an e x t i n c t i o n - l i k e suppression of responding i n a run-way task reinforced with e l e c t r i c a l brain stimulation, the enhancement of running speed by 'priming* e l e c t r i c a l brain stimulation was not attenuated. An investigation into the e f f e c t s of apomorphine on s e l f -stimulation suggested that apomorphine mimicked the rewarding, but not the priming, properties of e l e c t r i c a l brain stimulation (Leith, 1983). Thus, the case against neuroleptics' blocking incentive or secondary r e i n f o r c i n g properties i n addition to, or instead of, primary reinforcement, i s strong. This i s important, because Wise has modified his o r i g i n a l hypothesis to one that states that the motivational arousal properties of both primary and secondary reinforcers are l o s t . This re-formulation i s quite d i f f e r e n t than the f i r s t hypothesis, not only i n including secondary reinforcement (which, i n l i g h t of the preceding evidence, appears untenable) but i n s h i f t i n g the emphasis from hedonia to a c t i v a t i o n (motivation). The evidence just discussed indicates that the arousal properties of reinforcers are not altered by neuroleptics; the evidence appears more supportive of the o r i g i n a l anhedonia theory. Neuroleptic E f f e c t s  on Stimulus Control Tombaugh, Ritch and Shepherd (1980) observed that pimozide did not a l t e r discrimination accuracy on a simultaneous discrimination task, although response rates were reduced. As the discrimination was maintained by reinforcement, one would think that no blunting of reinforcement was evident. Unfortunately, the investigators did not demonstrate the e f f e c t of extinction on the discrimination task, so d i r e c t comparisons between the e f f e c t s of the neuroleptic and non-reward are not possible. The evidence i s suggestive, however, that reward was not completely attenuated. Furthermore, as discrimination accuracy was not affected, the authors argued that i t i s u n l i k e l y that secondary reinforcement was reduced, as t h i s might be expected to reduce stimulus control. This claim may not be wholly j u s t i f i e d ; although discriminative s t i m u l i do acquire incentive properties, and such properties may indeed f a c i l i t a t e or energize responding, the control of those stimuli over behavior i s not the r e s u l t of conditioned reinforcement. Discriminative stimuli are e f f e c t i v e at c o n t r o l l i n g behavior because of the information concerning the environmental contingencies that they impart. I t i s t h i s property that results i n the development of conditioned r e i n f o r c i n g e f f e c t s by discriminative s t i m u l i (Mackintosh, 1974). Such secondary reinforcement q u a l i t i e s may aid i n retarding extinction, but do not underlie stimulus c o n t r o l . This experiment does e s t a b l i s h that neuroleptics do not seriously impair the a b i l i t y of an animal to perceive s t i m u l i , and to d i r e c t t h e i r behavior accordingly. The main f a u l t of the previous experiment was r e c t i f i e d i n an experiment by Beninger (1982). He trained three groups of rats on a simple discrimination task and then observed the e f f e c t s of non-reward, pimozide + reward and pimozide + non-reward on discriminative performance. While a l l three conditions resulted i n a progressively greater response suppression, pimozide tended to reduce responding to a greater extent than extinction. Accuracy of responding was not affected by any of the treatments. Therefore, the lack of e f f e c t of pimozide on response accuracy i s si m i l a r to that observed with reward omission. The conclusion that neuroleptics do not i n t e r f e r e with stimulus control or associative learning i s supported by a variety of experiments i n d i c a t i n g no, or only very s l i g h t , loss of discrimination accuracy aft e r neuroleptic treatments (Tombaugh, 1981; Szostak, Tombaugh and Tombaugh, 1981). A common observation i n these discrete t r i a l situations i s that neuroleptics increase the latency to respond and decrease response rate. However, t h i s response e f f e c t i s not correlated with an a l t e r a t i o n i n accuracy. Neuroleptic E f f e c t s  on Timing and Counting An i n t e r e s t i n g finding of the experiment by Szostak and Tombaugh (1981) i s that pimozide appears to cause rats to perceive a s p e c i f i c duration as being shorter than i t a c t u a l l y i s , or i f they count responses, perceive a s p e c i f i c 33 number of responses as being less than they a c t u a l l y are. Other investigators have found that methamphetamine increases both the duration of an i n t e r v a l and the number of discrete s t i m u l i perceived by rats (Maricq, Roberts and Church, 1981; Maricq and Church, 1983; Church and Meek, 1984) and that haloperidol decreases these perceptions (Maricq and Church, 1983; Church and Meek, 1984). Church and Meek (1984) have discussed a variety of evidence that suggests animals have a single central mechanism that i s responsible for both timing and counting, and that these are general processes not s p e c i f i c for a p a r t i c u l a r stimulus modality. Such an "inte r n a l clock" may provide a pacemaker which regulates the sequencing and timing of behavior. As neuroleptics appear to decrease the rate of t h i s i n t e r n a l clock, the often observed increase i n response latencies and decrease i n response rates may r e f l e c t a 'slowing' of t h i s pacemaker. If t h i s i s so, then the actions of neuroleptics may not be s t r i c t l y motor, but rather r e f l e c t an e f f e c t on a process that contributes to the organization of behavior over time. It also appears possible that i f reinforcement magnitude i s assessed by a counting or timing mechanism, the observed 'anhedonia* may r e s u l t from a decreased perception of amount of a reinforcer. Recent evidence reported by Roberts and Holder (1985) has indicated that the timing of an event i s affected by the signal value of that event. During reward omission i n a two-lever duration discrimination, animals exhibited a bias 34 towards responding to the short-duration stimulus-cued lever. Thus, reward omission produces the same e f f e c t as haloperidol on timing behavior. Furthermore, Holder and Roberts (1985) observed that the perceived duration of a tone was correlated with the strength of a response conditioned to the tone. These authors conceive of the function of an i n t e r n a l clock to be the prediction of the temporal occurrence of s i g n i f i c a n t events. However, i f prediction i s the function of the clock, i t i s d i f f i c u l t to explain why a l t e r i n g the si g n i f i c a n c e of a stimulus would a l t e r the timing of i t s duration; accurate prediction requires a clock that maintains a constant rate. The results of these experiments examining the rela t i o n s h i p between signal value and timing indicate that rats time the duration of s i g n i f i c a n t events i n proportion to t h e i r value: the more valuable an event the longer i t appears to be. Thus, the present evidence suggests that an i n t e r n a l clock's function i s not the prediction s i g n i f i c a n t events, but the assessment of i t s value. If t h i s i n t e r p r e t a t i o n i s correct, then the above evidence suggests that neuroleptics decrease the value of rewards and DA agonists increase reward value. Neuroleptic E f f e c t s on Rate-Independent  Measures of Reinforcer E f f i c a c y A major problem with many of the experiments discussed above u t i l i z i n g free-operant procedures i s that rate of responding has been taken as an index of r e i n f o r c i n g value. 35 Rate of responding i s not r e f l e c t i v e of the reinforcement magnitude. Indeed, reducing the reinforcement per response by imposing p a r t i a l reinforcement schedules increases response rate. Evenden and Robbins (1983) investigated the e f f e c t of amphetamine, a-flupenthixol (a neuroleptic), extinction and s a t i a t i o n on c e r t a i n rate-free measures of reward and reward omission. Rats given a choice to respond to either of two levers (each associated with equivalent p a r t i a l reinforcement schedules) exhibit a tendency to repeat responses to the lever associated with the most recent reward (win-stay). Furthermore, i f a lever press has not been rewarded, rats tend to s h i f t responding to the alternate lever ( l o s e - s h i f t ) . These response strategies are independent of rate. Omission of reward results i n an increase of switching from one lever to another (lose-shif t ) . S a t i a t i o n or treatment with a-flupenthixol did not produce any e f f e c t on win-stay or l o s e - s h i f t behaviors. Amphetamine, however, decreased both win-stay and l o s e - s h i f t response strategies. Evenden and Robbins interpreted the e f f e c t of amphetamine as i n d i c a t i v e of a loss of the d i r e c t i v e a b i l i t y of reinforcement. In contrast, amphetamine had no influence on rate of responding, while both s a t i a t i o n and a-flupenthixol decreased response rate. Extinction i n i t i a l l y increased response rate, followed by a decrease. Thus, the neuroleptic did not exhibit features of reward omission, either i n terms of response choice or response rate. Indeed, i t appeared that a DA agonist "uncoupled" the behaviors of rats from the ef f e c t s of reinforcement contingencies, but only those behaviors not d i r e c t l y relevant to obtaining reward. Heyman (1983) used a procedure similar to the reward summation function for ICSS to assess the ef f e c t s of pimozide and amphetamine on reinforcement. Animals were trained on a five-component multiple v a r i a b l e - i n t e r v a l schedule and then were tested after treatments with various doses of pimozide or amphetamine. He found that pimozide decreased both the asymptotic response rate and increased the reinforcement rate necessary to maintain responding at half-maximum rates, while amphetamine decreased the reinforcement rate necessary to maintain responding at h a l f -maximum l e v e l . Heyman interpreted these findings as establishing that pimozide reduces both the hedonic properties of food and motor capacity, and that amphetamine increases the hedonic aspect of food without influencing response capacity. Morley, Bradshaw and Szabadi (1984) examined the ef f e c t s of pimozide i n a si m i l a r procedure,, with two important exceptions: only two d i f f e r e n t VI schedules were used, and they were imposed during d i f f e r e n t sessions. Sessions with a high rate of reinforcement were conducted f i r s t . These authors found that pimozide decreased responding to the same degree regardless of reinforcement rate. This led them to conclude that reinforcement e f f i c a c y was not affected by neuroleptics. However, as the order of the schedules was not counterbalanced, i t i s possible that 37 sequence e f f e c t s may have produced th e i r r e s u l t s . For example, since rats had had previous experience with pimozide before testing with the low reinforcement rate, progressive e f f e c t s of pimozide may have resulted i n a greater response suppressant e f f e c t than may have been observed i f rats had not had previous pimozide treatments. The results of t h i s experiment have questionable v a l i d i t y ; Heyman's data, i n which multiple VI schedules were presented within the same session, and some e f f o r t was made to counter-balance the order of the VI schedules, appear more v a l i d . Heyman's results are open to an alternative i n t e r p r e t a t i o n . The* idea that the rate of reinforcement necessary to maintain responding at half-maximal lev e l s r e f l e c t s reinforcement e f f i c a c y i s acceptable. The questionable aspect of Heyman's interpretation i s r e l a t i n g reinforcement e f f i c a c y to hedonia. Sinnamon (1982) presented a reward-effort model of reinforcement i n which an animal evaluates . the anticipated hedonic value of obtaining a reinforcer i n l i g h t of the required e f f o r t necessary to obtain i t . The eff e c t s of neuroleptics on reinforcement e f f i c a c y , as found by Heyman with food reinforcement, and by others mentioned previously using rate-free measures of ICSS, may r e s u l t either from a reduction i n hedonic value or from an increase i n the perceived amount of e f f o r t required to obtain the rein f o r c e r . Neuroleptics may increase perceived e f f o r t by increasing the threshold l e v e l of 38 a c t i v a t i o n of motor systems underlying behavior, or by operating on the evaluative process d i r e c t l y , enhancing the " e f f o r t input" component of the evaluative process. Complementary to t h i s , amphetamine may either decrease threshold a c t i v a t i o n levels of motor systems or reduce the " e f f o r t input" into the evaluative process. Within the confines of t h i s model, neuroleptics may also reduce reinforcement e f f i c a c y by increasing the stimulus thresholds of whatever parameter of food s t i m u l i that are assigned value (such as sweetness intensity) or by d i r e c t l y reducing the value assigned to food i n the evaluative process. Whether DA systems are involved i n willingness to expend e f f o r t , the amount of e f f o r t required to perform a p a r t i c u l a r response, sensory thresholds, hedonia (the value assigned to a behavior or stimulus, which i s presumably altered by such factors as hunger and stimulus q u a l i t i e s ) , or some combination of these factors i s not c l e a r . Xenakis and Sclafani (1981, 1982) found that pimozide decreased consumption of sucrose solutions i n a manner simi l a r to that observed after adulteration of the solutions with quinine or a f t e r d i l u t i n g the solutions. While th i s may r e s u l t from e f f e c t s on any of the processes just discussed, the fact that l i c k i n g rate for sucrose was suppressed i n a greater proportion than l i c k i n g rates for water, regardless of t h i r s t l e v e l , suggests that either sweetness or the hedonic value assigned to sucrose solutions was reduced, and not only willingness to exert e f f o r t . That l i c k i n g for both sucrose and water are reduced to some extent indicates that perceived e f f o r t may also be increased by neuroleptics. However, Gramling, Fowler and C o l l i n s (1984) observed that rats l i c k i n g sucrose solutions a f t e r pimozide treatments did not resemble rats undergoing extinction i n either rate or pattern of responding. Furthermore, the duration of l i c k s was increased by pimozide over that observed during vehicle or extinction, suggesting a motor d e f i c i t . It i s possible that the greater suppression of sucrose l i c k i n g than l i c k i n g for water produced by pimozide (Xenakis and Scl a f a n i , 1981) r e f l e c t s rate-dependent differences i n the action of neuroleptics on l i c k i n g behavior. A l t e r n a t i v e l y , l i c k i n g patterns may d i f f e r , depending upon the solution, and such differences i n response patterns could be d i f f e r e n t i a l l y sensitive to disruption by neuroleptics. Response-Specific E f f e c t s  of Neuroleptics Ettenberg, Koob and Bloom (1981) observed that a-flupenthixol, i n doses that are s u f f i c i e n t to attenuate responding on a lever-press task, does not reduce response rate when nose-poking i s the operand. They suggested that since neuroleptic suppression i s behavior-dependent, the anhedonia hypothesis i s based on an a r t i f a c t . They take th e i r evidence to suggest that neuroleptics produce a motor d e f i c i t which depends on the type of behavior. This in t e r p r e t a t i o n i s supported by the observation that a c q u i s i t i o n of an operant task i s not impaired by pimozide, provided the response requirement i s minimized (Tombaugh, Szostak and M i l l s , 1983). However, within the framework of Sinnamon's reward-effort model, i t can be seen that i f a response takes l i t t l e e f f o r t , decreases i n reinforcement value w i l l have only s l i g h t e f f e c t s on behavior. Assuming that neuroleptics produce a t o t a l block of primary reinforcement, Ettenberg et a l . ' s observation i s not explicable, unless neuroleptics produce t h e i r e f f e c t by increasing perceived e f f o r t . In t h i s case, an increase i n the perceived e f f o r t of nose-poking may not have much of an e f f e c t on behavior i f nose-poking takes very l i t t l e e f f o r t to begin with. An increase i n very l i t t l e e f f o r t i s s t i l l only s l i g h t e f f o r t , and no decrease i n responding would be predicted, unless higher doses are used. Thus, the nose-poking and minimal response requirement data i s understandable i f one accepts the view that neuroleptics increase the perceived e f f o r t of behavior. Neuroleptic E f f e c t s on Responses  D i f f e r i n g i n E f f o r t : A Test of the  Perceived E f f o r t Hypothesis Asin and Fibiger (1984) conducted a test of the "perceived e f f o r t " hypothesis of neuroleptics. This model predicts that a neuroleptic, such as haloperidol, w i l l produce greater response suppression with a task requiring greater e f f o r t . They therefore examined the e f f e c t of haloperidol on the response rate of rats trained to press either a lever with l i t t l e force requirement ( l i g h t lever) or one with a r e l a t i v e l y large force requirement (heavy le v e r ) . Haloperidol resulted i n equivalent decreases i n response rate between rats trained on one of the two types of levers. However, th i s r e s u l t could not be attributed to non-reward either; animals undergoing extinction exhibited less response suppression when trained on the heavy lever than those trained to respond to the l i g h t lever. This finding may be interpretable i f one assumes that haloperidol produced both non-reward and an increase i n perceived e f f o r t . Non-reward results i n a greater response rate on heavy levers than on l i g h t levers, while the "perceived e f f o r t " hypothesis predicts the opposite. If both factors are involved i n the actions of neuroleptics, then the d i f f e r e n t i a l e f f e c t s of "reward omission" and "perceived e f f o r t " on responding to the heavy lever may cancel each other out, r e s u l t i n g i n a response pattern indistinguishable from that observed i n rats trained to press a l i g h t lever. Summary The evidence just surveyed regarding the role of DA transmission i n mediating the properties of rewards i s not conclusive. DA depletions decrease ICSS supported by electrodes i n some brain s i t e s , but not i n others. In addition, the decrease i n responding observed after DA depletions i s often incomplete or transient. Increases i n DA 42 metabolism are reported during ICSS supported by electrodes i n some brain regions, but not i n others. Neuroleptics produce a progressive decrease i n responding for both e l e c t r i c a l brain stimulation and for natural rewards i n a manner that resembles the e f f e c t s of reward omission. However, the e f f e c t s of neuroleptics do not resemble those of reward omission i n a variety of measures and i n a variety of procedures. Neuroleptics reduce and DA agonists increase reinforcement e f f i c a c y i n rate-independent procedures with both e l e c t r i c a l brain stimulation and natural rewards. However, e f f e c t s of drugs on reinforcement e f f i c a c y may r e s u l t from actions on either reward or motor processes, or on the coordination of the two. DA agonists increase responding for ICSS, and t h i s e f f e c t cannot always be attributed to non-specific response ac t i v a t i o n or arousal. Furthermore, DA agonists are s e l f -administered, produce conditioned place preferences and enhance the a c q u i s i t i o n of secondary reinforcement by previously neutral e f f e c t s . DA receptor antagonists can attenuate self-administration of DA agonists and conditioned place preferences produced with amphetamine. However, place preferences with a variety of DA agonists are not attenuated by DA antagonists, and DA depletions have f a i l e d to convincingly attenuate conditioned place preferences. Central DA depletions attenuate self-administration of DA agonists. 43 Neuroleptics do not appear to reduce the incentive value of s t i m u l i , nor do they a f f e c t stimulus control. Both DA agonists and antagonists a f f e c t timing and counting behaviors i n a manner consistent with increasing and decreasing stimulus value, respectively. However, the evidence r e l a t i n g stimulus value with perceptions of duration and number i s presently sparse. Neuroleptics reduce responding for sweet solutions to a greater degree than responding for water, but the e f f e c t s on response patterns for sweet solutions are not s i m i l a r to those observed with reward omission. Rationale There i s a need for a procedure that can determine unequivocally whether or not animals' perceptions of the value of reinforcers i s decreased by neuroleptics, and enhanced by amphetamine. Thus, a procedure i s required i n which e f f o r t i s not a relevant factors. To t h i s end, an attempt was made to t r a i n animals to discriminate between d i f f e r e n t magnitudes of reinforcement. Providing that rats attend to the hedonic value of food when discriminating between d i f f e r e n t magnitudes, then alte r a t i o n s i n t h e i r perceptions of the hedonic properties of food can be determined by investigating the e f f e c t s of drugs on the point of subjective equality (PSE), the magnitude of reinforcement which i s as close to a large magnitude as i t i s to a small magnitude of reinforcement. For technical reasons, reward quantity was chosen as the dimension of reward magnitude. Church and Meek (1984) successfully trained rats to discriminate between two d i f f e r e n t quantities of s t i m u l i . By introducing intermediate quantities of st i m u l i as generalization probes, they obtained generalization gradients that allowed determination of the PSE, which r e f l e c t s the perception . of number of st i m u l i . The goals of the present research were to determine whether or not rats perceptions of food p e l l e t quantity r e f l e c t the hedonic value of those food p e l l e t s , and, i f so, whether a DA antagonist or agonist would a l t e r those perceptions. Experiment 1 determined whether rats evaluate four food p e l l e t s as being of greater hedonic value than one food p e l l e t . Experiment 2 assessed the e f f e c t of motivational changes (d i f f e r e n t l e v e l s of food deprivation) on the r e l a t i v e hedonic value of 4 and 1 food p e l l e t s . In Experiment 3, rats were trained to discriminate between two quantities of food p e l l e t s , as well as two quantities of tones, so that perceptions of reinforcers with d i f f e r e n t hedonic value could be compared with perceptions of neutral st i m u l i that d i f f e r i n number. Experiment 4 determined the influence of motivation on perceptions of stimuli d i f f e r i n g i n hedonic value and number, as well as on st i m u l i that d i f f e r only i n number. Experiment 5 assessed the r e l a t i v e hedonic value of food p e l l e t s that d i f f e r i n sweetness, but not i n number or n u t r i t i o n a l value. Experiment 6 evaluated the e f f e c t of motivational manipulations on the r e l a t i v e hedonic value of food d i f f e r i n g only i n sweetness and hedonic value. Experiment 7 examined whether a s p e c i f i c number of food p e l l e t s with r e l a t i v e l y low hedonic value, i s perceived as equivalent to food p e l l e t s of higher hedonic value but i d e n t i c a l number, or to food p e l l e t s of equivalent hedonic value but lower i n quantity. Experiment 8 determined i f the res u l t s of Experiment 7 could be attributed to stimulus novelty e f f e c t s , and further examined i f the discrimination of d i f f e r e n t quantities of tones was based on counting (or timing) or on s p e c i f i c stimulus-response associations. Experiment 9 determined whether tones were act u a l l y hedonically neutral. Experiment 10 assessed whether food quantity discriminations could be attributed to the sound of the food dispenser, rather than to q u a l i t i e s of food. F i n a l l y , i n Experiment 11, the ef f e c t s of 3 doses of haloperidol and 3 doses of d-amphetamine on the perception of food quantities and tone quantities was assessed, to determine i f a neuroleptic s e l e c t i v e l y decreases and a stimulant s e l e c t i v e l y increases rats' perceptions of the magnitude of a rewarding stimulus (food quantity). 46 METHODS  Subjects Male Long Evans rats (Charles River), maintained at 85% of ad libitum body weight, except as noted, were used i n a l l experiments. Rats that were subjects i n experiments i n excess of 3 months duration were p e r i o d i c a l l y (approximately every 2-3 months) placed on ad libitum access to food i n order to e s t a b l i s h appropriate body weights. Water was continuously available, except during experimental contingencies. Rats were housed i n d i v i d u a l l y , and were maintained on a 12 hr l i g h t cycle ( l i g h t on from 8:00 -20:00 hr). Apparatus Experimental apparatus consisted of either 2 or 4 operant chambers, depending upon the experiment. Each chamber was equipped with 2 levers positioned 4 cm to the l e f t and r i g h t of a central food cup (1.5 cm from floor) and 7 cm above the f l o o r , a p e l l e t dispenser, a cue l i g h t located 12 cm above the food cup and a sonalert tone generator (3200 hz, producing a clear tone the volume of which was dampened with a r e s i s t o r to a l e v e l c l e a r l y audible, but not uncomfortably loud, to the experimenter). For some experiments, 2 of the operant boxes were modified with the addition of a p e l l e t dispenser feeding into the central food cup, and a tone generator (SMB, Star Micronics, 47 emitting a range of frequencies [300-500 Hz] which produced a "buzzing" sound), were added. The food p e l l e t s used for a l l experiments were standard (sweetened) "Dustless" precision p e l l e t s (Bioserve Inc.), except where noted otherwise. The operant chambers (BRS-LVE) were contained within sound-attenuating boxes, and the a i r supply provided a background noise that i s o l a t e d the animals from disruptive noises. In addition, experiments were ca r r i e d out i n an iso l a t e d room. A l l experimental contingencies were controlled by MANX state set software (GC Controls) operated by a NOVA IV/X minicomputer (Data General) connected to the operant boxes by a MANX interface (GC Controls). Data ac q u i s i t i o n (MANX state set language) and r e t r i e v a l (Fortran IV) computer programs were written by the author. Drug Treatments Haloperidol (McNeil), i n the form of Haldol (dissolved i n solution with l a c t i c acid) was further d i l u t e d with d i s t i l l e d water and injected (i.p.) 25 min before beginning of experimental contingencies. D-amphetamine sulphate (Smith Klein & French) was dissolved i n s t e r i l e saline (0.9%), and injected (i.p.) 5 min before i n i t i a t i o n of experimental contingencies. Vehicle was either d i s t i l l e d water or saline, depending upon the subsequent drug conditions. Drugs were prepared on each day of use. 48 Order of Experiments The experiments are described i n what i s hoped to be a l o g i c a l sequence. However, as many of the experiments involved the same subjects over a rather long period of time, i t i s important to note i n what sequence these experiments were act u a l l y conducted. The discrimination and related experiments were conducted i n the following order: Experiment 3, 11, 4, 7, 8, 9 and 10. S t a t i s t i c s Analysis of variance of the results of some experiments was conducted using a computer GANOVA program (copywrite by M.L. Brecht and J.A. Woodward, 1983). This program provided a multivariate s t a t i s t i c as an estimate of Wilk's lambda (Rao's R, d i s t r i b u t e d as exact F, described i n the text simply as R) i n the cases of repeated measures with more than 2 repeated measures, as ANOVA i s only v a l i d under conditions of compound symmetry within variance and covariance matrices. The advantages of a multivariate estimate of p i n a univariate repeated measures design are discussed by V i t a l i a n o (1982), McCall and Appelbaum (1973) and Kirk (1968). Planned i n d i v i d u a l comparisons were made, u t i l i z i n g the GANOVA program. In some experiments, i n which response choice was assessed i n r e l a t i o n to a n u l l hypothesis of random response d i s t r i b u t i o n , t-tests were conducted, estimating the p r o b a b i l i t y that choices were greater than that expected by chance (Minitab computer package, available on the UBC Amdahl computer system). 49 EXPERIMENT 1. Preferences f o r D i f f e r e n t Quantities of Food P e l l e t s as Assessed by a Free-Operant Choice Procedure Rats may discriminate between two d i f f e r e n t quantities of food p e l l e t s by attending to any of a number of d i f f e r e n t properties of the food. If the discrimination along the dimension of quantity of food i s to be a useful index of the reward value, then d i f f e r e n t quantities must d i f f e r i n t h e i r hedonic value. Therefore, an experiment was undertaken to examine whether or not rats prefer four food p e l l e t s over one p e l l e t . A free-operant choice procedure was followed i n which, afte r equivalent experience with rewarded presses to each of two levers i n i s o l a t i o n , rats had simultaneous access to both levers, responses to either of which resulted i n food reward. Presses on one lever were followed by the presentation of 4 food p e l l e t s , and responses to the alternate lever were reinforced by 1 food p e l l e t . After te s t i n g for i n i t i a l preferences, the contingencies were reversed, such that responses to that lever which was previously reinforced by 1 p e l l e t was then rewarded by 4 p e l l e t s , and vice versa. The contingency reversal was imposed to control for response biases. Procedure After 1 week acclimatization to the colony room, 12 rats with ad libitum food access were weighed for 3 consecutive days, and the 85% ad libitum body weights were 50 calculated from the weights recorded on the t h i r d day. Food was removed from the cages following the t h i r d weighing, and each rat was fed a dish of Bioserve "Dustless" precision p e l l e t s the following day, to accustom them to the food to be used i n the experiments. Each subsequent day, the rats were fed 5-8 g of rat chow u n t i l t h e i r weights approached the calculated deprivation l e v e l , at which point the amount of food given was adjusted to maintain the weights at that l e v e l . At t h i s point, rats began experimental sessions. The f i r s t 10 sessions consisted of lever press t r a i n i n g . Only 1 lever was i n s t a l l e d i n each of 2 operant chambers during a p a r t i c u l a r session. Six rats had access to the l e f t lever on the f i r s t day, and six to the r i g h t lever. On the second day, the f i r s t s i x rats had access to only the right lever, and the second set of rats were given access to the l e f t lever. On each subsequent day, the lever to which each rat had access to was alternated. Of each set of six rats, three were reinforced with 4 p e l l e t s a f t e r each response to the f i r s t accessible lever, and with 1 p e l l e t after each response to the alternate lever on alternate days. Thus, equal numbers of rats began t h e i r t r a i n i n g with each of the two levers, and equal numbers of rats began th e i r t r a i n i n g with rewards of 4 or 1 food p e l l e t s . This counterbalancing procedure was followed to eliminate possible biasing e f f e c t s related to i n i t i a l exposure to p a r t i c u l a r contingencies or to i n i t i a l lever preferences. On 51 the f i r s t 2 days, food p e l l e t s were placed on the lever to f a c i l i t a t e i n i t i a l a c q u i s i t i o n of the lever press response. Each session continued u n t i l 50 responses were made, immediately aft e r which point the animals were removed from the operant chamber. Thus, at the end of the t r a i n i n g phase of the experiment, each rat had made 250 responses to one lever, each of which was rewarded by 4 p e l l e t s , and 250 responses to the alternate lever, each of which was reinforced by 1 p e l l e t . Six test sessions, each 15 min i n duration, were then conducted on succeeding days. During the f i r s t two sessions both levers were accessible simultaneously, with contin-gencies si m i l a r to those present during t r a i n i n g . Conversely, during the l a s t 4 sessions the response-contingencies were reversed such that responses to the levers previously rewarded by de l i v e r y of 4 p e l l e t s were reinforced with 1 p e l l e t and vice versa. Results Data from each test session for each rat were expressed as the responses to the 4 p e l l e t lever as a percent of t o t a l responses to both levers. The mean % choice of the 4 p e l l e t lever for each of the f i r s t 2 choice test sessions and the 4 subsequent contingency reversal sessions are displayed i n Figure 1. A l l twelve rats chose to respond to the lever which was rewarded by 4 p e l l e t s (4 p e l l e t lever) for 75% or more of the t o t a l responses choices on both of the f i r s t 2 52 100-1 REVERSAL CONSECUTIVE DAILY 15 MEN TEST SESSIONS FIGURE 1. Preferences between 4 and 1 food pellets as assessed by a free-operant choice procedure In rats. Responding on one of two levers was rewarded by 4 pellets and responding to the alternate lever was rewarded by 1 pellet. Rats were given two initial test sessions followed by four test sessions in which the lever contingency was reversed (REVERSAL). 53 test days. T-tests comparing the mean of the group with the hypothesis (u. > 50) were conducted for each session. Both of the f i r s t 2 sessions resulted i n s i g n i f i c a n t preferences for responding to the 4 p e l l e t lever (t = 31.4, p < 0.0001; t = 21.9, p < 0.0001, respectively, df = 11). On the f i r s t contingency reversal session, no s i g n i f i c a n t preference for the 4 p e l l e t lever was observed (t = 1.61, p > 0.1; n u l l hypothesis: u. < 50). Responding to the 1 p e l l e t lever, responses to which had previously been rewarded with 4 p e l l e t s , was high during the i n i t i a l periods of the session, but decreased as the session progressed. Conversely, responding to the 4 p e l l e t lever (previously reinforced with 1 pe l l e t ) was r e l a t i v e l y low at the beginning of the session and progressively increased. By the second contingency reversal test, rats exhibited s i g n i f i c a n t preferences for the 4 p e l l e t lever (t = 2.98, p < 0.01), and th i s preference became more pronounced on the remaining 2 contingency reversal tests (t = 9.63, p < 0.0001; t = 10.78, p < 0.0001, respectively; see figure 1). A l l rats exhibited a preference for the 4 p e l l e t lever of more than 70% by the fourth test session. Discussion The rats exhibited strong preferences for responding to the lever that was rewarded with 4 p e l l e t s . This preference was maintained by every rat following a contingency reversal 54 after a s u f f i c i e n t number of test sessions had been conducted. The reward-dependent lever ^ preference i s p a r t i c u l a r l y notable i n the l i g h t of observations that rats (observed i n other experiments) often develop strong biases towards responding to a p a r t i c u l a r lever, usually the ri g h t lever (furthest from the door of the operant chamber). The strong preference i n i t i a l l y observed, and the r a p i d i t y with which the rats 'tracked' the change i n contingency, provide compelling evidence that some property or properties of food quantity (perhaps c a l o r i c value or some other post-ingestive influence) can contribute to the hedonic value of food. A l t e r n a t i v e l y , the greater hedonic value of the larger quantity of food may be the r e s u l t of an evaluation of the reward gained versus the e f f o r t expended. 55 EXPERIMENT 2 E f f e c t of Pre-loading Rats with Food on Preferences f o r Different Quantities of Food P e l l e t s The results of Experiment 1 indicate that 4 food p e l l e t s have a higher hedonic value than 1 food p e l l e t . It i s of i n t e r e s t to determine whether or not the observed preferences may be influenced by the motivational state of the rats. Cabanac (1971) observed that how pleasant i n i t i a l l y hungry human subjects rated a sucrose solution or orange smell depended upon whether or not they had just ingested a quantity of sucrose or glucose solution. Both sucrose solution and orange smell were rated as being less pleasant af t e r pre-loads. Progressively lower pleasantness ratings were given to sucrose solutions i f the subjects swallowed the sample solutions, but not i f the samples were sp i t out. He coined the term 'negative a l l i e s t h e s i a ' to refer to the phenomenon of reduced pleasantness ratings applied to s t i m u l i following decreases i n the b i o l o g i c a l s i g n i f i c a n c e of the s t i m u l i . Negative a l l i e s t h e s i a has been observed i n humans with both food (Cabanac, 1971; Esses and Herman, 1984) and temperature (Cabanac, 1971; A t t i a , 1984). D i f f e r e n t i a l hedonic values of foods that d i f f e r only i n quantity and related variables may be the r e s u l t of an evaluation of the b i o l o g i c a l u t i l i t y of the food: the a b i l i t y to reduce hunger i n t h i s case. If t h i s i s the case, then pre-loading the rats with some quantity of food to 56 reduce hunger may reduce preferences between foods that d i f f e r s i n c a l o r i c value, but not i n taste. A l t e r n a t i v e l y , i f the value of d i f f e r e n t quantities of food obtained by exerting an i d e n t i c a l degree of e f f o r t i s related to an evaluation of gain versus e f f o r t , then a l t e r i n g motivational l e v e l would not be expected to change preferences. Procedure The same rats and the same procedure used i n Experiment 1 were u t i l i z e d i n t h i s experiment. The present experiment d i f f e r e d from Experiment 1 i n that each session was preceded by exposure to some quantity of the standard laboratory rat chow normally fed i n the home cage. Thus, 60 min pri o r to each of 3 days of testing, the rats were fed 5.0, 7.5 or 10.0 g of food. Following these sessions, rats were placed on an ad libitum feeding regimens and 3 additional sessions were conducted on succeeding days. Results Figure 2 depicts the res u l t s from the l a s t session of Experiment 1 (0 g pre-load) along with the means of the percent choice of the 4 p e l l e t lever during sessions preceded by food pre-loads. No s i g n i f i c a n t e f f e c t of pre-loading rats with food or maintaining them on ad libitum food access was observed on the mean percent choice of the 4 p e l l e t lever (R(6,6) = 2.70, p > 0.1). Univariate i n d i v i d u a l comparisons of each pre-load session with the session 57 100-. 0.0 5.0 7.5 10.0 AO LIB. FOOD ACCESS CONSECUTIVE DAILY 15 MIN TEST SESSIONS FIGURE 2. Effect of pre-loading rats with different quantities of food (grams) or ad libitum (AD LIB.) access to food on responses to a lever rewarded by 4 pellets, relative to responses to an alternate lever rewarded by 1 pellet. Rewards to either lever were delivered on a erf schedule. 58 without pre-load did not reveal any s i g n i f i c a n t e f f e c t s on choices. However, as can be observed i n Figure 3, pre-loading the rats with food did have a s i g n i f i c a n t e f f e c t on the t o t a l number of responses emitted per session (R(6,6) = 24.14, p < 0.002). Univariate planned comparisons revealed that the three test sessions conducted while animals were on ad libitum food access d i f f e r e d s i g n i f i c a n t l y from the session without pre-load ( F ( l , l l ) = 22.6, p < 0.001; F ( l , l l ) = 23.6, p < 0.001; F ( l , l l ) = 55.4, p < 0.0005, respe c t i v e l y ) . Sessions preceded by pre-loads of 5.0, 7.5 or 10.0 g of food did not d i f f e r s i g n i f i c a n t l y from control ( F ( l , l l ) = 1.35, p > 0.1; F ( l , l l ) = 0.46, p > 0.1; F ( l , l l ) = 2.12, p > 0.1, re s p e c t i v e l y ) . Thus, while ad libitum feeding reduced the total, number of responses emitted, i t had no e f f e c t on the choice of responding for 4 p e l l e t s over 1 p e l l e t . Discussion No e f f e c t of pre-loading rats with food p r i o r to t e s t i n g for preferences for 4 or 1 food p e l l e t s i n a free-operant choice procedure was observed (see Figure 2). While inconsistent with the o r i g i n a l b i o l o g i c a l u t i l i t y model of Cabanac (1971), Cabanac et a l . (1971) provided evidence suggesting that people maintained at body weights below th e i r normal weights do not exhibit negative a l l i e s t h e s i a . 59-' 100-, 0.0 5.0 7.5 10.0 AD LIB. FOOD A C C E S S CONSECUTIVE DAILY 15 MIN TEST SESSIONS FIGURE 3. Effect of pre-loading rats with different quantities of food (grams) or ad libitum (AD LIB.) access to food on total responses to both of two levers, responses to which were differentially rewarded with either 4 or 1 pellet. Rewards to either lever were delivered on a CRF schedule, v. Significantly lower than 0.0 gram condition, p < 0.001 . 60 However, t h i s finding was observed i n only 2 subjects, who happened to be the experimenters and therefore not unbiased or unaware of either the independent or dependent variables. These are l i k e l y important factors, as S t e l l a r (1977), u t i l i z i n g a similar procedure, observed that disguising the taste of the pre-load affected his a b i l i t y to r e p l i c a t e Cabanac's o r i g i n a l observation. Esses and Herman (1984) were unable to r e p l i c a t e the f a i l u r e to observe negative a l l i e s t h e s i a i n dieters (n = 18). The present findings are consistent with the economic reward-effort model of Sinnamon (1982), but do not provide unequivocal support. A variety of interpretations of the lack of e f f e c t of pre-loading with food on food quantity preferences are available. I t i s possible that pre-loading with food a l t e r s the hedonic properties of both 4 and 1 food p e l l e t s to the same degree. In t h i s case, the r e l a t i v e hedonic values do not change, although there might well be al t e r a t i o n s i n the absolute hedonic value of the food p e l l e t s . A t h i r d possible explanation i s that the food pre-loading regimen did not s i g n i f i c a n t l y a l t e r motivational l e v e l . This does not appear to be l i k e l y . The amount of food given was within the range of the normal d a i l y intake of these rats (under deprivation conditions), and the 7.5 and 10.0 g amounts were such that an hour was required for the food to be a l l consumed. Furthermore, i t i s d i f f i c u l t to believe that hunger was not decreased by the three days of ad libitum access to food. The finding that the t o t a l responses decreased with food pre-load also suggests that motivation was reduced, although the degree of response attenuation was not as great as might be expected. However, i t i s possible that the sweetness of the food p e l l e t s used i n the choice procedure provided a strong incentive stimulus that would tend to counteract the motivational loss due to a decrease i n hunger. Yet another explanation i s that of 'habit' , or ' f i x a t i o n ' , as discussed by Tolman (1932). It i s possible that the rats continued to respond out of habit, rather than because of current hedonic values of the food. In l i g h t of the rather s l i g h t reductions i n t o t a l responses observed after what were l i k e l y major changes i n the state of hunger, i t appears probable that one or both of the l a t t e r interpretations provide the best explanations for the present r e s u l t s . If the e f f o r t required to obtain the food was a major determining factor of choice, then larger reductions i n t o t a l responses would be expected when hunger was reduced. 62 EXPERIMENT 3 Food and Tone Quantity Discriminations A major d i f f i c u l t y plaguing research into the neurological substrates of reinforcement has been the problem of assessing the hedonic properties of s t i m u l i i n non-human animals. Hedonic properties of s t i m u l i have t r a d i t i o n a l l y been inferred by preference tests (Young, 1968, 1977). This procedure does not appear p a r t i c u l a r l y useful for physiological studies of the mechanisms underlying hedonic processes because there i s no a p r i o r i reason to assume that physiologic manipulations would influence the hedonic value of one stimulus but not another. Another method of measuring the hedonic value of a stimulus may be to approximate the rating scales employed i n human research. Although rats cannot provide verbal reports of subjective experiences, i t may be possible to t r a i n them to discriminate between s t i m u l i on the basis of the d i f f e r i n g hedonic value of those s t i m u l i . The process of over-shadowing suggests that rats would code the presentation of food i n terms of i t s most s a l i e n t a t t r i b u t e : i t s value. The present experiment investigated whether rats would learn to discriminate between 2 d i f f e r e n t quantities of food (4 and 1 food p e l l e t s ) , shown i n Experiment 1 to d i f f e r i n hedonic value. The a c q u i s i t i o n and maintenance of t h i s discrimination was compared to that of (presumably) 63 neutral s t i m u l i , tones, also d i f f e r i n g along the dimension of quantity. Tones were adopted as the stimulus modality for the control discrimination t r i a l s for 3 reasons: p i l o t experiments indicated that the a c q u i s i t i o n of quantity discriminations with tones was faster and reached a higher and more stable asymptote than with l i g h t s ; tones would provide a better control than l i g h t s because the presentation of food p e l l e t s was associated with the sound of c l i c k s , the number of c l i c k s being d i r e c t l y equal to the number of food p e l l e t s delivered; i t appeared l i k e l y that tones would be more hedonically neutral than l i g h t s , as rats are nocturnal animals and tend to exhibit aversion to bright l i g h t . This experiment had two goals: to t r a i n rats on quantity discriminations to be used i n further experiments, and to investigate possible differences between the a c q u i s i t i o n and maintenance of discrimination performance when a rewarding stimulus i s used as a discriminative cue (food p e l l e t s for hungry rats) as opposed to an assumed hedonically neutral cue (tones). 64 Procedure Rats were brought to 85% of t h e i r ad libitum weights, as described i n Experiment 1. The i n i t i a l body weights, as determined on the t h i r d day of weighing, ranged from 240-285 g. As these rats began t r a i n i n g at a r e l a t i v e l y young age, the calculated 85% ad libitum weight was incremented by 20 g a week for the f i r s t 3 weeks and 10 g for the fourth week, to allow for growth. After the f i r s t month, t r a i n i n g was stopped and the rats were returned to ad libitum food access for one week, to allow determination of a new 85% ad libitum body weight that would more accurately r e f l e c t an appropriate deprivation l e v e l for adult rats. Thereafter, the rats were p e r i o d i c a l l y removed from t r a i n i n g or t e s t i n g and returned to ad libitum feeding regimens for re-assessment of deprivation body weight levels as discussed i n the General Methods section. A l l rats were trained to press both of 2 levers by allowing access to only 1 lever during a session that terminated a f t e r 100 responses were emitted, followed on a subsequent day by r e s t r i c t e d access to the alternate lever. To f a c i l i t a t e a c q u i s i t i o n of the lever response, double-sided s t i c k y tape was applied to the lever and p e l l e t s were attached to t h i s tape. Each rat had emitted 100 rewarded (1 p e l l e t ) lever presses on each lever at the beginning of the discrimination a c q u i s i t i o n phase of the experiment. 65 During the a c q u i s i t i o n phase of th i s experiment, rats were required to press one lever (lever A) afte r presentation of 1 p e l l e t , and the alternate lever (lever B) after presentation of 4 p e l l e t s . The correct response was rewarded by the delivery of one p e l l e t , regardless of the number of p e l l e t s delivered as a cue. Equating the reinforcement across a l l types of t r i a l s was considered necessary to reduce the degree of differences i n incentive associated with d i f f e r e n t cues. Interspersed among food p e l l e t discrimination t r i a l s were tone discriminations, i n which one tone (3200 Hz) signaled reward a v a i l a b i l i t y upon responses to one lever (lever B) and 4 tones cued reinforcement upon responses to lever A. Note that i n no case did the correct lever response following the presentation of 4 tones correspond with that cued by 4 p e l l e t s . This requirement was adopted to reduce the pr o b a b i l i t y that rats would learn a general counting strategy (Church and Meek, 1983), i r r e s p e c t i v e of the type of cue stimulus. Furthermore, the l i k e l i h o o d that rats would attend to the sound cues associated with p e l l e t delivery ( i e . p e l l e t dispenser c l i c k s ) may have been lessened by tr a i n i n g rats to press a d i f f e r e n t lever after 4 tones than afte r 4 food p e l l e t s . The sequence of a p a r t i c u l a r t r i a l (using the presentation of 4 tones as an example) was as follows: 66 1) Illumination of the houselight served as a t r i a l cue (for a l l kinds of t r i a l s ) , ending with the termination of the t r i a l . 2) Two sec after illumination of the t r i a l cue, a tone would sound for 0.75 sec, followed 0.55 sec l a t e r by the next tone, and so forth, u n t i l 4 tones were sounded. These times were chosen to cl o s e l y approximate the durations of the times the food p e l l e t dispenser was on and off during presentation of 4 food p e l l e t s . 3) After the l a s t tone ceased, the t r i a l continued u n t i l a response was made. During i n i t i a l a c q u i s i t i o n t r a i n i n g , a correction procedure was operative, such that both correct and incorrect responses were recorded, but the t r i a l d id not end u n t i l a correct response was made. After a l l rats performed p e l l e t discriminations at 75% correct or better for 3 successive days, the correction procedure was dropped. After t h i s point the f i r s t response after cessation of the cue was recorded, and the t r i a l ended. Each correct response was immediately followed by the delivery of 1 p e l l e t . During the l a s t 10 sessions, only 75% of the correct responses were rewarded by 1 food p e l l e t (determined by a pr o b a b i l i t y function). The correction procedure operated during the f i r s t 10 sessions, the next 20 sessions operated with 100% reinforcement of correct responses, and the l a s t 10 sessions operated with a 75% reinforcement schedule. 4) At the end of the t r i a l , the houselight was shut of f , and a 20 sec i n t e r - t r i a l i n t e r v a l (ITI) began, during 6 7 which responses to either lever were recorded, but had no other programmed consequences. T r i a l types occurred i n a s p e c i f i e d sequence, where F = food p e l l e t , T = tone, subscript^ = 1 cue, and subscript4 = 4 cues: Fi_ F 4 T 4 F 4 Ti_ F1 11 T 4 T 4 F 4 F1 T1 T 4 Ti_ F 4 F]_ F 4  F l T l T4* T n i s sequence was repeated u n t i l 80 t r i a l s were completed. T r i a l s did not begin u n t i l 5 min after the rats were f i r s t placed into the operant boxes. Each program began pr i o r to placing the rats i n the boxes with a u t i l i t y test subprogram that checked that the house-light, tone and p e l l e t dispenser were a l l functioning properly. Rats were tested at approximately the same time each day, 5 days a week. Half of the rats were trained with the delivery of 4 p e l l e t s cuing responses to the right lever as correct, and the remaining half were trained with the 4 p e l l e t cue signaling reward aft e r a response to the l e f t lever. Besides recording correct and incorrect responses, and ITI responses, the latency of each response following the termination of each cue was recorded. Results Figure 4 depicts the mean percent correct responses for each t r i a l type. Analysis of variance with 3 dependent factors (blocks of 2 sessions, stimulus modality [food p e l l e t s or tones] and quantity [1 or 4 stimuli]) was used to analyze separately the mean % correct of each of 4 consecutive phases of 5 blocks of 2 days. The f i r s t phase 68 O Ld cn o o h-z L d o LY. Ld D_ < Ld 100-9 0 -8 0 -7 0 -6 0 -5 0 -4 0 3 0 -2 0 -10 0 Legend • 1 P E L L E T 4 P E L L E T S O 1 TONE 4 T O N E S i—i—i—i—i—i—i—i—i—i—i—i—i—i—r i — i — i — i — r ~ 1 2 3 4 5 6 7 8 9 10 11 12 1 3 1 4 1 5 16 1 7 1 8 19 2 0 CONSECUTIVE BLOCKS OF 2 SESSIONS FIGURE 4. Acquisition and asymptotic performance of two forced-choice quantity discriminations. Rats were trained to press either of two levers, depending upon the number of pellets (1 OR 4) or the number of tones (1 OR 4) delivered as a cue. Correct responses were rewarded with 1 food pellet, regardless of the cue type. 69 consisted of the data c o l l e c t e d during the correction procedure (acq u i s i t i o n phase), the 100% reinforcement schedule operated during the next 2 phases (asymptotic response phases), and the data represented i n the l a s t phase were c o l l e c t e d during the 75% reinforcement schedule (75% reinforcement phase). In reference to Figure 4, phase 1 i s represented by the f i r s t 5 blocks of two sessions, phase 2 and 3 refer to the next 10 blocks and phase 4 consists of the l a s t 5 blocks. As can be seen i n Figure 4, while a l l rats performed the discrimination task at chance leve l s on the f i r s t block, i n i t i a l a c q u i s i t i o n rates were cue-dependent, the e f f i c a c y of the cues at e l i c i t i n g correct responding being i n the order of F4 > Fi_ > T4 > T]_. While the i n t e r a c t i o n among blocks, stimulus modality and stimulus quantity during the a c q u i s i t i o n phase was not s i g n i f i c a n t , both the blocks x stimulus modality i n t e r a c t i o n and the blocks x stimulus quantity were s i g n i f i c a n t (R(4,15) = 28.4, p < 0.0001; R(4,15) = 9.63, p < 0.001, re s p e c t i v e l y ) . The i n t e r a c t i o n between stimulus modality and quantity was not s i g n i f i c a n t . Individual comparisons were planned between stimulus types and between quantities for each block of 2 days (Fi_xTi_, F4XT4, Fj_xF 4, Ti_xT4, F 1xT4, F 4xTi_). No s i g n i f i c a n t differences were revealed between any of the comparisons on the f i r s t block (F(l,18) = 0.08, 0.04, 1.22, 0.82, 1.86, 0.38, respectively, p > 0.1 i n a l l cases). By the second 70 block, rats made s i g n i f i c a n t l y more correct responses after presentation of the Fi_ cue than aft e r the T]_ cue (F(l,18) = 4.49, p < 0.05) but did not d i f f e r s i g n i f i c a n t l y from the T 4 cue (F(l,18) = 0.80, p > 0.1). Correct responding after the F 4 cue was s i g n i f i c a n t l y higher than after the F]_ (F(l,18) = 5.86, p < 0.025), T 4 (F(l,18) = 12.91, p < 0.0025) or T1 (F(l,18) = 60.3, p < 0.0001) cues, and correct responding to the T 4 cue increased r e l a t i v e to correct responses after the T]_ cue (F(l,18) = 5.67, p < 0.05). By the t h i r d block, correct responding after the F 4 cue was no longer s i g n i f i c a n t l y higher than a f t e r the Fi_ cue (F(l,18) = 1.22, p > 0.1), but correct responding after both food cues was s i g n i f i c a n t l y higher than aft e r either tone cue (F^xT]_: F = 17.9, p < 0.001; F 4 X T 4 : F = 15.8, p < 0.005; Fi_xT 4: F = 8.14, p < 0.02; F 4xT 1: F = 42.4, p < 0.0001, df=(l,18) i n a l l cases). Rats made more correct responses aft e r the T 4 cue than a f t e r the cue (F(l,18) = 8.76, p < 0.01). By the fourth block, the same relationships held, except that correct responding to the T]_ cue had reached the same l e v e l as responding to the T 4 cue (F^xF 4:F = 1.16, p > 0.1; Fi_xT^: F = 19.5, p < 0.001; F 4xT 4: F = 30.8, p < 0.0002; Fi_xT 4: F = 31.2, p < 0.0002; F4xT1: F = 42.8, p < 0.0001; T^xT^: F = 0.004, p > 0.1, df = (1,18) i n a l l cases). During the f i n a l block of the i n i t i a l a c q u i s i t i o n phase, both 1 and 4 food p e l l e t cues were followed by high levels of accurate responding (> 80%), that were not s i g n i f i c a n t l y d i f f e r e n t from each other (F(l,18) = 0.26, p > 0.1). Accuracy of 71 responding after tone cues was not very high (>60% but less than 70%), but also did not d i f f e r s i g n i f i c a n t l y from each other (F(l,18) = 0.79, p > 0.1). The response accuracy following either food cue was s i g n i f i c a n t l y higher than the accuracy associated with either tone cue (F]_xTi_: F = 7.98, p < 0.02; F 4 X T 4 : F = 30.6, p <• 0.0002; Fi_xT 4: F = 16.9, p < 0.001; F 4 x T i : F = 67.8, p < 0.0001, df = (1,18) i n a l l cases). S i g n i f i c a n t improvement i n the performance of the discriminations was evident during the i n i t i a l blocks of the next phase ( f i r s t part of asymptotic phase) for a l l of the cues (mean of f i r s t block (across a l l cue types and numbers) = 77.04, mean of l a s t block = 85.16), i n d i c a t i n g that asymptotic responding was not achieved i n the f i r s t part of t h i s phase. This i s also evident from the analysis of variance, which indicated that none of the i n t e r a c t i o n terms were s i g n i f i c a n t , but there was a s i g n i f i c a n t main e f f e c t of blocks (R(4,15) = 6.49, p < 0.005). The main ef f e c t s of both stimulus modality and quantity were s i g n i f i c a n t (F(l,18) = 62.6, p < 0.0001; F(l,18) = 9.21, p < 0.001, res p e c t i v e l y ) . Thus, responding to the food cues throughout t h i s phase was s i g n i f i c a n t l y better than responding to the tone cues; responding to 4 presentations of the s t i m u l i was s i g n i f i c a n t l y better than responding to single presentations of the s t i m u l i . During the next 5 blocks of the 'asymptotic' phase, none of the terms involving blocks were s i g n i f i c a n t , 72 i n d i c a t i n g that the rats were responding at asymptotic l e v e l for a l l cues. There was a s i g n i f i c a n t stimulus modality x quantity i n t e r a c t i o n (F(l,18) = 4.96, p < 0.05). Planned comparisons indicated that responding to the food cues was better than responding to the tone cues (F(l,18) = 35.4, p < 0.0005) and that responding aft e r 4 presentations of the stimuli was more accurate than aft e r 1 presentation (F(l,18) = 13.64, p < 0.005). The F]_ cue e l i c i t e d better performance than the T1 cue (F(l,18) = 19.3, p < 0.001); responding to the F 4 cue was more accurate than to the T4 cue (F(l,18) = 18.79, p < 0.001); more correct responses were made after the F 4 cue than the Fi_ cue (F(l,18) = 5.51, p < 0.05), and responding to the T4 cue was more accurate than to the T]_ cue (F(l,18) = 9.02, p < 0.01). Thus, a higher asymptotic response l e v e l was reached for the food stimuli than for the tone s t i m u l i , and more accurate asymptotic responses were emitted aft e r the 4 quantity s t i m u l i than the single stimulus cues. The order of cues with respect to l e v e l of accuracy was F4 > Fi_ = T4 > Ti_. The relationships between cues observed i n the asymptote phase were maintained during the 75% reinforcement phase. None of the terms involving blocks nor the stimulus modality X quantity int e r a c t i o n were s i g n i f i c a n t . However, the e f f e c t s of both stimulus modality (F(l,18) = 37.0, p < 0.0005) and stimulus quantity (F(l,18) = 12.76, p < 0.005) were s i g n i f i c a n t . The order of the e f f i c a c y of the d i f f e r e n t cues at e l i c i t i n g accurate responding was i d e n t i c a l to that 73 observed during the asymptote phase (F]_xTi_: F = 17.00, p < 0.001; F4XT4: F = 22.02, p < 0.0005; F2XF4: F = 8.60, p < 0.01; Ti_xT 4: F = 7.13, p < 0.025, df = 1,18 i n a l l cases). The latencies to respond aft e r presentation of the l a s t cue i n each cue series were analyzed during the 75% reinforcement phase. As can be observed i n Figure 5, there were major differences i n response latencies, depending on cue type and number. Latencies were ordered, from longest to shortest i n the following sequence: F4 >> F]_ > T]_ > T4. The i n t e r a c t i o n of stimulus , modality and quantity was s i g n i f i c a n t (F(l,18) = 49.27, p < 0.0001). Planned comparisons revealed that latencies after the F4 cue were s i g n i f i c a n t l y longer than after the Fi_ cue (F(l,18) = 32.5, p < 0.0005), the Fi_ latencies were s i g n i f i c a n t l y longer than the T1 latencies (F(l,18) = 11.23, p < 0.005), and the T1 cue latencies were s i g n i f i c a n t l y longer than those following the T 4 cue (F(l,18) = 31.06, p < 0.0005). None of the terms involving blocks was s i g n i f i c a n t . The responses that were emitted during the i n t e r - t r i a l i n t e r v a l s (ITI) during the 75% reinforcement phase were analyzed as a measure of non-specific response bias. The number of responses to the lever signaled by the F4 cue was divided by the sum of the responses directed to both levers for each animal and each block of 2 days. A value of 0.5 r e f l e c t s equivalent numbers of responses to both levers, while larger values are i n d i c a t i v e of a bias towards the F 4 lever, and values less than 0.50 indicate a greater number 74 of responses to the F-]_ lever. Figure 6 depicts the results of t h i s response bias index. There was a s l i g h t trend for rats to respond more to the F i _ lever than the F 4 lever, and t h i s trend was s i g n i f i c a n t l y d i f f e r e n t from chance on blocks 1 and 3 (block 1: t = 2.25, p < 0.05; block 2: t = 1.82, p > 0.05; block 3: t = 2.14, p < 0.05; block 4: t = 1.63, p > 0.1; block 5: t = 1.40, p > 0.1). 75 O CO, 00 UJ o 00 z o CL-IO 4-3-2-1-Legend • 1 PELLET • 4 PELLETS O 1 TONE • 4. TONES 1 2 3 4 CONSECUTIVE BLOCKS OF 2 SESSIONS FIGURES. Response latencies for quantity discriminations with cues consisting of 1 or 4 food pellets or 1 or 4 tones. Latencies were obtained from trials during the last phase of acquisition in which reinforcement for correct responses were delivered on a 75% schedule. 76 X bJ Q in < m CO z o Q_ 00 UJ or 1 0 .9-0 .8 -0.7-0.6-0 . 5 -0.4 0.3 0.2 0.1 0 o-1 2 3 4 CONSECUTIVE BLOCKS OF 2 SESSIONS - O — ' —o-FIGURE 6. Response bias as indicated by the proportion of responses to the lever cued by the 4 food pellet stimulus during inter—trial intervals. Response bias was calculated during the last phase of acquisition. 77 Discussion This experiment demonstrated that rats can learn quantity discriminations, with both a presumably neutral stimulus and with a r e i n f o r c i n g stimulus. Furthermore, as Experiment 1 established that d i f f e r e n t quantities of food p e l l e t s l i k e l y have d i f f e r e n t hedonic values, the p o s s i b i l i t y that the hedonic values of the 2 quantities of food p e l l e t s served as the discriminative cue may be entertained. In p a r t i a l support of t h i s interpretation, the present results indicate that the food discrimination was not based on a mechanism i d e n t i c a l to that of the tone discrimination. F i r s t l y , the a c q u i s i t i o n of the food discrimination was much faster than that of the tone discrimination (see Figure 1). Secondly, correct responding to the (presumably) hedonically more valuable food cue was acquired more quickly than to the less preferred food cue. Thirdly, the asymptotic l e v e l of responding for the food discrimination was s i g n i f i c a n t l y higher than that of the tone discrimination. In addition, i t i s apparent from the latencies to respond after presentation of food cues that rats ate the p e l l e t s p r i o r to making a response (see figure 5). Therefore, taste or immediate post-ingestive cues were l i k e l y those occurring i n closest temporal proximity to the response. 78 Although the above considerations are consistent with the view that the hedonic attributes of the food s t i m u l i served as the discriminative cue, they are also consistent with alternative interpretations. Acquisition of the food quantity discrimination may have been faster, and the asymptotic response l e v e l may have been higher, than observed with the tone discrimination because of attentional factors. Food p e l l e t s provide a plethora of multi-modal cues: the sound of the p e l l e t dispenser, the sight of the food (although minimized by dim i l l u m i n a t i o n ) , the smell and taste, and the f e e l of the p e l l e t s i n the mouth, and the difference i n the time taken to d e l i v e r 1 versus 4 p e l l e t s a l l could have provided cues that may have enhanced discrimination performance. However, only the sound of the tones, and the difference i n the time taken to present the cues could have conceivably f a c i l i t a t e d performance of the tone discrimination. Some of the factors l i s t e d above can be ruled out on the basis of Occam's Razor. The temporal cues could not provide s u f f i c i e n t information to allow the rats to respond c o r r e c t l y to cues of both stimulus modalities, as the long duration food cue signaled reinforcement a v a i l a b i l i t y on a lever d i f f e r e n t from that cued by the long duration tone stimulus ( i e . the stimulus composed of 4 successive tones). It remains possible that the rats timed both tones and p e l l e t s , but were able to d i f f e r e n t i a t e between the duration of one stimulus modality from that of the other. This 79 implies that the rats must attend to some other property of the stimulus to make the correct response. Although the presentations of p e l l e t s and tones were made equivalent i n terms of duration of the i n t e r v a l during which the st i m u l i were presented, the difference i n latencies to respond act u a l l y resulted i n longer durations for the p e l l e t cues than for the tone cues. Thus, i t i s possible that the discriminations were made on the basis of absolute durations, rather than durations of 1 stimulus r e l a t i v e to the duration of 4 s t i m u l i . However, the complexity of the processing involved i n t h i s kind of a duration discrimination makes thi s p o s s i b i l i t y u n l i k e l y . The p e l l e t discrimination task may have been aided i f the rats counted the " c l i c k s " of the p e l l e t dispenser, as i t delivered the food. However, counting sounds would probably have resulted i n an i n h i b i t o r y generalization gradient from the p e l l e t discrimination to the tone discrimination, and vice versa. Thus, t h i s strategy would l i k e l y have been detrimental to the ac q u i s i t i o n of both discrimination tasks. The p o s s i b i l i t y remains, however, that the rats learned the food discrimination faster and achieved a higher performance l e v e l with food cues because of the saliency of these cues, r e s u l t i n g from the multi-modal nature of the food p e l l e t s . The fact that rats performed better after presentation of 4 tones than after the presentation of 1 tone, both i n ac q u i s i t i o n and during asymptotic responding, suggests that superior performance after the F4 than after 80 the Fi_ cue may have been due to the enhanced saliency of a greater quantity of s t i m u l i , rather than because of 4 food p e l l e t s have a greater hedonic valency than 1 food p e l l e t . Inspection of the latencies to respond, however, reveal that superior performance i n response to the T4 cue, r e l a t i v e to the T]_ cue, may be the r e s u l t of a mechanism separate from that contributing to the e f f i c a c y of the F4 cue. Whereas the latencies to respond to the F4 cue are by far the longest of a l l the cues, latencies to respond to the T4 cue are the shortest. Informal observations of the responding of the animals during sessions revealed that on tone t r i a l s , the animals began responding immediately after the presentation of the f i r s t tone. On T4 t r i a l s , the rats tend to switch from pressing the lever cued by the Tj_ stimulus over to the lever cued by the T 4 cue during the tone presentations. Thus, the rats would often respond to the lever during the tone presentations, r e s u l t i n g i n latencies of less than a second. That the rats performed better on the T4 than on the T]_ t r i a l s i s r e a d i l y understood as a r e s u l t of f a i l i n g to be rewarded by pressing the T]_ lever a f t e r the f i r s t tone, and then switching to the correct lever before the end of the stimulus presentation period. On presentation of the f i r s t tone the rats are not able to determine which lever-response i s associated with reward; but high levels of accuracy could have been achieved by adopting a strategy of always responding to the T]_-correct lever f i r s t , and switching to the alternate lever i f 81 the reward i s not immediately forthcoming. On the other hand, the rats did not often press either of the levers during food cue presentations. Responses were apparently not made u n t i l a f t e r the food was consumed. Thus, the mechanism underlying the r e l a t i v e e f f i c a c y of the F4 cue appears to be quite d i f f e r e n t from that producing the r e l a t i v e e f f i c a c y of the T4 cue. These observations suggest that the reasons for improved performance after the F 4 cue were not the same as for that following the T4 cue. Furthermore, the phenomenon of overshadowing leads one to expect that the predominant (most salient) stimulus a t t r i b u t e of the food s t i m u l i would acquire predictive strength, at the expense of less s a l i e n t stimulus a t t r i b u t e s , obviating any e f f e c t of multiple stimulus a t t r i b u t e s . However, i t appears that taste i s often associated with potentiation of conditioning of redundant cues, rather than overshadowing (Lett, 1982; Spear and Kucharsky, 1983). Evidence supporting the existence of within-compound stimulus learning has also been reported (Rescorla and Durlach, 1981). Thus, the view that the multi-stimulus properties of food cues are responsible for t h e i r e f f i c a c y as discriminative s t i m u l i cannot be discounted. Another possible i n t e r p r e t a t i o n of the superior a c q u i s i t i o n rate and asymptotic response l e v e l of food quantity discriminations also invokes the concept of stimulus saliency. I t i s l i k e l y that food i s inherently a more s a l i e n t stimulus than a sound, at least to a hungry rat. Thus, performance of a food discrimination may be 82 better than that of a tone discrimination because of the saliency of food, regardless of the multi-modal nature of the food cue. That the saliency of food cues may be responsible for the differences between the performances of food and tone quantity discriminations i s not anathemic to a view that the value of the food was a c r i t i c a l element of the discrimination, however. I t i s e a s i l y argued that just those features that make food a s a l i e n t stimulus are those that are responsible for i t s hedonic properties. I t i s also possible that the incentive e f f e c t s of food cues are responsible for the performance differences observed i n the discrimination tasks. Presentation of food cues may provide an incentive to respond as well as the information required to make the correct response. A l l cues were equivalent i n terms of the amount of reward that they predicted the animals would have access to, and as such reward expectancies (incentive) may have been similar for a l l types of cues. However, a number of factors might mitigate against an equivalency of incentive. The incentive values of the discriminative s t i m u l i may have been related to the learned associative strength between cues, then incentive value may be related to the asymptote response accuracy for each cue. If t h i s were so, then incentive value would have d i f f e r e d between the cues i n the order of F4 > F]_ = T 4 > T]_. Furthermore, the food cues bore a natural r e l a t i o n s h i p to the rewards (both being food p e l l e t s ) , which may have increased the incentive value of food. However, i f 83 the s i m i l a r i t y between cue and reward were a major determining factor i n the difference i n cue e f f i c a c y , then the F]_ cue would have e l i c i t e d more accurate responding than the F4 cue. This was not the case. Of course, i t i s possible that q u a l i t a t i v e , but not quantitative, s i m i l a r i t y was the relevant r e l a t i o n s h i p contributing to incentive. In t h i s case, the food cues maintained a higher l e v e l of accuracy than the tones because of s i m i l a r i t y to the reward, while the greater stimulus quantities may improved performance due to attentional factors. A l t e r n a t i v e l y , food cues may have acquired less incentive value than the tone cues, because of transient negative contrast e f f e c t s . Thus, the reward may have been less s i g n i f i c a n t to the rats after just consuming 1 food p e l l e t , and even less so a f t e r consuming 4 p e l l e t s , than during tone t r i a l s i n which no p e l l e t s were consumed immediately p r i o r to the reward. That transient negative contrast e f f e c t s occur has been well documented (vide Mackintosh, 1974). The latency data i n the present experiment may support t h i s i n t e r p r e t a t i o n (Figure 5), but as latencies to respond after cue p e l l e t consumption was not separated from latencies to respond after cue delivery, t h i s data i s inconclusive. Major differences i n the a c q u i s i t i o n and performance of the food and tone quantity discriminations were observed, and these differences are consistent with the view that the hedonic attributes of food may have provided the relevant discriminative cue. Nevertheless, a variety of alternative 84 explanations can account for these differences. The appropriate account cannot be selected on the basis of the present data. Furthermore, i t should be noted that while there were major quantitative differences i n the discrimination accuracies, depending on cue type and number, the differences were not q u a l i t a t i v e (although q u a l i t a t i v e differences i n latencies were observed). Thus, i t cannot be discounted that discrimination of cues of d i f f e r e n t modalities were based on i d e n t i c a l mechanisms, but d i f f e r e d only i n degree. 85 EXPERIMENT 4 The E f f e c t of Manipulations of Motivational State on Perceptions of Quantities of Food and Tone Cues The purpose of thi s experiment was to determine i f the the hedonic value of food plays a role i n rat's discriminations of d i f f e r e n t food quantities. Hedonic values of sweet solutions and water baths of p a r t i c u l a r temperatures, as assessed by verbal reports i n human subjects, depend upon motivational state (Cabanac, 1971; Cabanac et a l . , 1971; Esses and Herman, 1984), although there i s some doubt as to the r e l i a b i l i t y and inter p r e t a t i o n of t h i s phenomenon ( S t e l l a r , 1977). A major assumption of incentive theorists i s that deprivation enhances the incentive values of food (cf. Beck, 1978). The hedonic values of sweet tast i n g substances are increased by food deprivation (vide Pfaffman, 1977; Beck, 1978). Therefore, i f the rats i n Experiment 3 had discriminated between d i f f e r e n t quantities of food p e l l e t s on the basis of d i f f e r e n t i a l hedonic value, then that discrimination may be altered by changes i n l e v e l of food deprivation. The r e s u l t s of Experiment 2 indicate that preferences for 4 p e l l e t s over 1 p e l l e t , when measured i n a free operant choice procedure, are not modified by decreases i n motivation. As discussed previously, t h i s r e s u l t may r e f l e c t the fact that while the hedonic values of both quantities of food were altered, the r e l a t i o n between then did not. In 86 Experiment 3, assuming for the moment that the rats discriminated the two food quantities on the basis of hedonic value, such a discrimination may have been between 2 absolute l e v e l s of hedonia, or may have been performed on the basis of r e l a t i v e value. It appears possible that animals may attend to the r e l a t i v e hedonic value of food, and that t h i s r e l a t i v e value would not be changed by alt e r a t i o n s i n motivation. However, i t has been demonstrated that the incentive value of rewards i s a p o s i t i v e l y accelerated function of magnitude, including quantity (Leventhal, Morrel, Morgan and Perkins, 1959). This suggests that e f f e c t s of deprivation on incentive value would not be equivalent for d i f f e r e n t quantities, but would tend to have greater e f f e c t s on greater quantities. Therefore, i f rats discriminate d i f f e r e n t quantities of food on the basis of hedonic value, then performance of a food quantity discrimination should be influenced by l e v e l of deprivation. As discussed e a r l i e r (Experiment 2: Discussion), i t i s possible that 'habit' may over-ride effects related to changes i n hedonic processes. Young (1977) has provided evidence that strongly supports the view that "dietary habits tend to form so that they meet metabolic and n u t r i t i o n a l needs but established habits p e r s i s t even though opposed to bodily needs". Considering the high l e v e l of accuracy and the length of the testing of the food discrimination i n Experiment 3, i t appeared l i k e l y that the discrimination procedure would be i n s e n s i t i v e to changes i n 87 motivation. Therefore, a v a r i a t i o n of the discrimination procedure of Experiment 3 was used i n which rats were given presentations of intermediate quantities of food and tone cues, interspersed among presentations of the "regular" quantities. Responses following the usual 1 or 4 quantities were reinforced following a 75% reinforcement schedule, while intermediate quantities (2 or 3) were never reinforced. The t e s t i n g of such stimulus generalization gradients has a number of advantages. F i r s t l y , the generalization s t i m u l i are not followed by d i f f e r e n t i a l reinforcement, making subsequent responses more susceptible to the influence of variables a l t e r i n g perceptual processes. Secondly, c e r t a i n key features of generalization gradients (plotted as % of the responses directed to one lever, chosen a r b i t r a r i l y , against the number of st i m u l i presented) can be used to d i f f e r e n t i a t e between perceptual and performance e f f e c t s . The number of st i m u l i associated with 50% of the responses directed to each lever, referred to as the point of subjective equality (PSE), i s an index of perception that i s free of response a r t i f a c t s , as long as responding i s not suppressed to such a degree that data i s obtained from i n s u f f i c i e n t numbers of t r i a l s to accurately calculate the PSE. That i s , changes i n the accuracy of responding a f f e c t i n g a l l s t i m u l i quantities w i l l s h i f t the slope of the l i n e , without affecting, the abscissa value i n t e r s e c t i n g at the ordinate value of 50%. On the other hand, changes i n 88 perception, without performance e f f e c t s , w i l l r e s u l t i n a change i n the PSE without a change i n slope of the l i n e . If there i s a change i n both slope and PSE, then i t i s clear that both perceptual and performance factors have altered, with the caveat that very extreme performance disruptions (near complete f a i l u r e to respond) disallow conclusions concerning perception to be made. Thus, an a r t i f a c t - f r e e measure of perception of quantity i s obtainable, without the extensive t e s t i n g under d i f f e r e n t reinforcement schedules required by signal detection theory, and which i s l i k e l y to be susceptible to alterations i n pertinent variables. In t h i s experiment, the e f f e c t of three levels of food deprivation (75%, 85% and 95% of ad libitum body weight) on the generalization gradients of quantities of food and tones were investigated. If the perception of food quantities changed with a l t e r a t i o n s i n motivational l e v e l , without concomitant variations i n the perception of tone quantities, then i t can be concluded that the stimulus attribute(s) of food attended to by rats when performing quantity discriminations is/are related to the hedonic properties of food. On the other hand, i f changes i n the perceptions of quantities of both stimulus modalities occurs, then one must conclude that perceptual systems are influenced by motivational processes, regardless of the hedonic or a f f e c t i v e s i g n i f i c a n c e of the s t i m u l i , providing that the tones are neutral s t i m u l i . Thus, t h i s experiment addresses the issue of the nature of the stimulus attributes of food 89 attended to by rats when trained to discriminate quantity, assessing whether an operation that a l t e r s the hedonic properties of food also s e l e c t i v e l y a l t e r s the perception of quantity. Further predictions of the form of the generalization gradients can be made, depending on the type of cues the animals attend to. As the incentive value of food i s a p o s i t i v e l y accelerated function of quantity, the food quantity generalization gradients should take t h i s form i f animals attend to the absolute hedonic value of food. Furthermore, increasing food deprivation w i l l s h i f t the curve to the l e f t , producing a smaller PSE, as a r e s u l t of increased hedonic value. Decreasing food deprivation should have the opposite e f f e c t . However, i t rats discriminate food st i m u l i by attending to some property other than hedonic value, such as number, then the curve should be a straight l i n e , providing that animals discriminate d i f f e r e n t quantities by the absolute difference i n number. Furthermore, food deprivation should have no e f f e c t on the PSE, r e f l e c t i n g a lack of s h i f t i n the generalization gradient along the X-axis. Procedure The rats from Experiment 3 were used i n th i s experiment. The rats were assigned to one of two equally sized groups (n=9), such that each group was composed of about half of the rats (4-5) from each group discussed i n Experiment 12. A l l rats were given ad libitum access to food 90 for one week. One group was then given r e s t r i c t e d feeding to bring t h e i r weights to 95% of the free-feeding body weights, and the weights of the second group were brought to 75% of th e i r ad libitum weights. After t r a i n i n g and generalization t e s t i n g (discussed below), the rats of both groups were brought to 85% of t h e i r ad libitum weights over a period of 1 week, and given another phase of t r a i n i n g and generalization. F i n a l l y , the rats from the group that were i n i t i a l l y food deprived to 95% of t h e i r ad libitum body weights were deprived to 75% over a period of 1 week. Those rats i n i t i a l l y deprived to 75% of t h e i r ad libitum weights were brought to a 95% deprivation l e v e l over the same period. A l l rats were then given a week of t r a i n i n g followed by a f i n a l week of generalization t e s t s . The rats were given 5 t r a i n i n g sessions on consecutive days of t r a i n i n g (discrimination t r a i n i n g with 75% reinforcement - see Experiment 3) and then 5 sessions of generalization testing, also on consecutive days. The general protocol of the generalization testing was sim i l a r to that used i n Experiment 3 i n regards to ITI, st i m u l i presentations, and a 5 min delay after placing the animals i n the tes t i n g apparatus. The major differences were that t r i a l s were chosen sequentially out of a l i s t of 60 items (rather than 20) for a t o t a l of 60 t r i a l s (rather than 80), with the t r i a l s being i n a d i f f e r e n t pseudo-random order, and with the addition of 5 presentations of each of 4 novel 91 s t i m u l i : 2 and 3 p e l l e t s and tones, responses following which were never reinforced. The order of t r i a l s was as follows, where F = food p e l l e t s , T = tones, and the subscript refers to the number delivered: Fi_ T1 T]_ F]_ T 4 F 4 F 4 F]_ F 2 F 2 Fi_ Ti_ T1 F 4 T 2 F1 T 2 T 3 T 4 F 2 T 4 T X F 4 F 3 F 4 T 3 T 4 T 3 F 3 T 4 F 4 T 4 T 2 F 4 T 3 T 3 T 2 F 4 F 3 F 2 F 4 T 4 F X T 4 F 3 11 F 3 T± F 2 T 2 F1 T 4 F± T 4 T]_ F X T l F l T l F4* T ^ e n u m h e r of t r i a l s was chosen so that rats would not receive t h e i r t o t a l d a i l y food requirement during a session, even when they were under the most r e s t r i c t i v e deprivation condition. The number of t r i a l s involving intermediate quantities of st i m u l i was kept to half of that of the regular discrimination t r i a l s (10 t r i a l s of each discrimination t r i a l - F]_, F 4, T]_ and T 4 ) , 75% of which were reinforced, so that the o v e r a l l reinforcement rate was 50%. The number of "probe" t r i a l s (5 of each of F 2, F 3, T 2 and T 3) of intermediate quantities was kept lower than those of the discrimination t r a i n i n g t r i a l types i n order to decrease the p r o b a b i l i t y that the rats would learn that probe t r i a l s were not reinforced. Furthermore, the f i r s t and l a s t 8 t r i a l s did not include probe t r i a l s of intermediate quantities. Results The point of subjective equality (PSE) and the slope from the generalization gradient was calculated indepen-dently for each rat under each deprivation condition for 92 both food and tone t r i a l s . The percent responses directed to the levers associated with 4 sti m u l i for each t r i a l type were calculated for each rat over each set of 5 sessions. These data were used to calculate the slope and PSE for every rat for both food and quantity t r i a l s , using the following formulae: PSE = 50 - (Y - (S * 2.5)) S where Y = the mean of the percent responses to the 4 stimulus lever (the lever cued by the F4 or T4 s t i m u l i ) , and S = the slope of the l i n e , calculated by the following formula: S = ((R 1+2R 2+3R 3+4R 4)-((10*(R 1+R 2+R 3+R 4))/4)) / 5 where R = % of the responses directed to the 4 stimulus lever a f t e r presentation of the stimulus i n the quantity designated by the subscript. These algebraic formulae are the regression equations for raw scores (Roscoe, 1975), derived from the least squares solution. Some of the results of these calculations were checked by comparison to those derived from a computer c a l c u l a t i o n of the PSE and slope from the l i n e of best f i t as determined by the methods of least squares, and were found to match + 0.002 (PSE) or 0.2 (slope) i n each case, which i s within rounding error. The r e s u l t s of the manipulations of food deprivation l e v e l on the generalization gradients (data collapsed across both groups) are depicted i n Figure 7 (food p e l l e t t r i a l s ) and i n Figure 8 (tone t r i a l s ) . S t a t i s t i c a l analyses were 93 100n Food Pellet Cues FIGURE 7. Food p e l l e t generalization gradients determined while rats were under each of three deprivation conditions (body weights 75%, 85% or 95% of ad libitum body weights) during consecutive periods. 94 cn 100-1 i 1 1 r 1 2 3 4 NUMBER OF TONES FIGURE 8. Tone generalization gradients determined while rats were under each of three deprivation conditions (body weights 75%, 85% or 95% of ad libitum body weights) during consecutive periods. 95 Legend o FOOD PSE • TONE PSE 2.50-, 2.25-1.75-1-50-1 1 1 , 75 85 95 FOOD DEPRIVATION CONDITION FIGURE 9. Points of subjective equality (PSE) derived from generalization gradients determined while rats were under each of three deprivation conditions (body weights 75%, 85% or 95% of ad libitum body weights) during consecutive periods. * S i g n i f i c a n t l y d i f f e r e n t from 75% condition, p < 0.02. 96 conducted on the two measures derived from the generalization gradients from i n d i v i d u a l animals: the PSE and slope of the gradients. Figure 9 displays the e f f e c t s of d i f f e r e n t food deprivation l e v e l s on the point of subjective equality derived from food t r i a l s (PSE F) and tone t r i a l s (PSE T). Analysis of variance revealed that there was a s i g n i f i c a n t e f f e c t of deprivation l e v e l on the point of subjective equality (PSE F) derived from food p e l l e t generalization gradients (R(2,16) = 4.94, p < 0.025). Comparisons between the d i f f e r e n t deprivation l e v e l s indicated that while the mean PSE from the 75% deprivation phase did not d i f f e r s i g n i f i c a n t l y from the 85% deprivation phase (F(l,17) = 1.41, p > 0.1), i t was s i g n i f i c a n t l y d i f f e r e n t from that observed during the 95% deprivation phase (F(l,17) = 7.67, p < 0.025). Furthermore, the mean PSE of the 95% phase was s i g n i f i c a n t l y lower than that exhibited during the intermediate deprivation condition (F(l,17) = 7.24, p < 0.025) . Analysis of the results of the tone PSE's were conducted aft e r exclusion of the results of one rat, which f a i l e d to perform the discrimination (the slopes of i t s generalization gradients for tones ranged from 2.6 to 15, with a mean of 9.5; PSE's ranged from -5.40 to +1.04, with a mean of -1.23), although i t s generalization gradient for food p e l l e t s was indistinguishable from those obtained from other r a t s . No s i g n i f i c a n t e f f e c t of food deprivation l e v e l on tone PSE's was apparent i n the groups of rats when thi s 97 rat was excluded (F(2,32) = 2.64, p > 0.05; R(2,15) = 2.27, p > 0.1). Analysis of the PSE T including the one rat that did not discriminate between d i f f e r e n t tone quantities did not a l t e r these r e s u l t s . Although food deprivation l e v e l influenced the perception of quantity of food p e l l e t s , as indicated by changes i n the PSE F, no s i g n i f i c a n t e f f e c t was observed on the slopes (SLOPEp) of the food generalization gradients (R(2,16) = 0.20, p > 0.1; see Figure 10). This indicates that deprivation l e v e l had no influence on variables related to the accuracy of performance of the food discrimination. Contrary to the r e s u l t s observed with the generalization gradients for the food cues, where deprivation affected the PSE F, but not the SLOPE F, deprivation l e v e l affected the slopes of the tone generalization gradients (SLOPE T), but not the PSE T. The e f f e c t of deprivation l e v e l on performance of the generalization and discrimination tasks i s indicated by a s i g n i f i c a n t decrease i n slope produced by changing the l e v e l of food deprivation (R(2,15) = 3.79, p < 0.05). This action was apparent only when maintaining body weights at 95% of ad libitum body weights (95% X 85%: F(l,16) = 4.56, p < 0.05; 95% X 75%: F(l,16) = 7.08, p < 0.02); there was no s i g n i f i c a n t difference between the slopes of the generalization gradients observed during the 85% and 75% feeding regimens (F(l,16) = 0.06, p > 0.1), as can be seen i n Figure 10. 98 50-. 40-30-20-O o--o -o 10-Legend o FOOD SLOPE • TONE SLOPE 1— 95 75 85 FOOD DEPRIVATION CONDITION FIGURE 10. Slopes derived from tone and p e l l e t generalization gradients determined while rats were under each of three deprivation conditions (body weights 75%, 85% or 95% of ad libitum body weights) during consecutive periods. * S i g n i f i c a n t l y d i f f e r e n t from 85% condition, p < 0.05. 99 The latencies to respond aft e r presentation of food cues were altered, depending upon the quantity of s t i m u l i presented, as indicated by a s i g n i f i c a n t deprivation condition X stimulus quantity i n t e r a c t i o n (R(6,12) = 6.22, p < 0.005). As can be observed i n Figure 11, the latencies following the F4 cue were most sensitive to changes i n motivational state. Response latencies to the F4 cue increased s i g n i f i c a n t l y with each 10% increase i n body weight (75% x 85%: F(l,17) = 5.70, p < 0.05; 85% x 95%: F(l,17) = 22.5, p < 0.0005). The latencies following the F x cue were the next most sensitive to motivational conditions. The apparent increase i n response latencies observed between the 75% and 85% deprivation l e v e l s was near significance (F(l,17) = 4.04, p = 0.058), while the increase i n response latencies observed between the 85% and 95% deprivation conditions was s i g n i f i c a n t (F(l,17) = 11.52, p < 0.005). The t h i r d most sensitive food cue to the latency e f f e c t s of motivational l e v e l was the F3 stimulus. These response latencies increased s i g n i f i c a n t l y only from the 85% to the 95% food deprivation condition (F(l,17) = 5.40, p < 0.05). The response latencies following presentations of the F 2 cue were i n s e n s i t i v e to variations i n body weight; no s i g n i f i c a n t changes following changes i n deprivation l e v e l were observed. Although the response latencies to food cues were d i f f e r e n t i a l l y affected by the a l t e r i n g deprivation l e v e l , 100 10-! c? Ul OT, 8 -(/) UJ o Ul _l Ul CO z o Q_ CO Ul or 6-4-2-Legend O FOOD TRIALS 75% • FOOD TRIALS 85% V FOOD TRIALS 95% • TONE TRIALS 75% • TONE TRIALS 85% A TONE TRIALS 95% , , j f 1 2 3 4 NUMBER OF STIMULI PRESENTED FIGURE 11. Response latencies i n seconds (SEC) on t r i a l s i n which d i f f e r e n t quantities of food p e l l e t s or tones were presented. T r i a l s were conducted while rats were maintained on one of three d i f f e r e n t food deprivation conditions (75%, 85% or 95% of ad libitum body weights). 101 the relationships among d i f f e r e n t quantities of stimuli remained constant throughout the 3 phases of body weight changes. Response latencies a f t e r the F]_ cue were shorter than aft e r the other three cues throughout the t e s t i n g phases (Fi_xF 2: F = 129, p < 0.0001; Fi_xF 3: F = 144, p < 0.0001; F]_xF4: F = 50, p < 0.0001, df = 1,17 for a l l comparisons). Latencies after the F 4 cue were the next shortest, being s i g n i f i c a n t l y shorter than the F 3 cue (F(l,17) = 7.62, p < 0.02). Although latencies after the F 2 cue appeared to be the longest across a l l phases, v a r i a b i l i t y i n these latencies resulted i n a lack of s i g n i f i c a n t differences between them and those following the F 3 or F 4 cues. Thus, the order of the latencies was: F 2 = F 3 > F 4 >>> F]_, while the order of the s e n s i t i v i t y of the response latencies to the changes i n deprivation conditions was d i f f e r e n t : F 4 >>> F1 => F 3 > F 2. Latencies to respond after the presentation of d i f f e r e n t quantities of tones were also affected by the changes i n the body weights of the rats (see Figure 11). However, the i n t e r a c t i o n of food deprivation and stimulus quantity was not s i g n i f i c a n t (R(6,ll) = 1.60, p > 0.1). The main e f f e c t of deprivation l e v e l was s i g n i f i c a n t (R(2,15) = 11.98, p = 0.001), as was the main e f f e c t of tone quantities (R(3,14) = 3.60, p < 0.05). The main e f f e c t of deprivation r e f l e c t s the fact that each 10% increase i n body weight was associated with an increase i n response latencies to tone 102 cues (75% x 85%: F(l,16) = 5.10, p < 0.05; 85% x 95%: F(l,16) = 24.89, p < 0.0005). Latencies to respond did not d i f f e r between the and the T 4 cues (F(l,16) = 0.02, p > 0.1), nor were there s i g n i f i c a n t differences between latencies associated with the T 2 and T3 cues (F(l,16) = 1.96, p > 0.1). However, both intermediate quantity cues were associated with response latencies s i g n i f i c a n t l y longer than those following discrimination cues (T]_+T4 x T2+T3: F(l,16) = 8.92, p < 0.01). Responses during the i n t e r - t r i a l i n t e r v a l (ITI) decreased i n proportion to changes i n food deprivation (R(2,16) = 5.40, p < 0.02), dropping close to 50% for each 10% increase i n body weight. However, only the drop i n responding during the 95% food deprivation schedule was s i g n i f i c a n t l y d i f f e r e n t from the others (95% X 85%: F = 8.95, p < 0.01; 95% X 75%: F = 4.26, p = 0.05); 75% X 85%: F = 2.64, p > 0.1, df = 1,17 i n a l l cases), even though the drop i n ITI responding from 75% to 85% was of a similar proportion as that observed from 85% to 95%. This appears to be due to a large v a r i a t i o n i n responding during the ITI under the 75% deprivation l e v e l (Figure 12). The mean response bias index was not s i g n i f i c a n t l y altered by the deprivation conditions (75% = 0.38, 85% = 0.41, 95% = 0.46; R(2,16) = 2.14, p > 0.1) . 103 150-. oH 1 1 1 1 75 85 95 FOOD DEPRIVATION CONDITION FIGURE 12. Mean responses during i n t e r - t r i a l i n t e r v a l s (ITT) between stimulus quantity generalization t r i a l s . T r i a l s were conducted while rats were maintained on one of three d i f f e r e n t food deprivation conditions (75%, 85% or 95% of ad libitum body weights). 104 Discussion The r e s u l t s of t h i s experiment revealed several i n t e r e s t i n g features of the perceptions of food quantities by rats. The generalization gradients for varying quantities of tone cues exhibited negative acceleration as a function of quantity (Figure 8). The difference between 1 and 2 tones was apparently greater than the difference between 2 and 3 tones, and t h i s l a t t e r difference was greater than the difference between 3 and 4 tones. This r e s u l t suggests that rats perceived varying numbers of tones as d i f f e r i n g i n proportion, not i n absolute number. This finding agrees with previous research which demonstrated that rats count numbers of s t i m u l i by proportion (Church and Meek, 1983). The generalization gradients for d i f f e r e n t food p e l l e t quantities were not s u b s t a n t i a l l y d i f f e r e n t i n form from the tone gradients (Figure 8). They indicate that food p e l l e t quantities were also perceived i n terms of r e l a t i v e proportions. The form of the food generalization gradients suggests that rats were not attending to the incentive value of the food p e l l e t s when discriminating quantity, i f the incentive value of food p e l l e t s i s a p o s i t i v e l y accelerated function of p e l l e t number (Leventhal et a l . , 1959). However, i n the l i g h t of the present findings, the evidence for t h i s conclusion deserves some consideration. Leventhal et a l . (1959) trained rats i n a 2-ally maze. Entry into the goal box of one a l l y was consistently 105 rewarded with 1 food p e l l e t , while entries into the alternate goal box were rewarded with a variable amount of reward, 0 or 2 food p e l l e t s randomly and equally d i s t r i b u t e d across t r i a l s . They observed that rats r e l i a b l y chose to enter the goal box which contained the variable amount of reinforcement, even though average amount of reward was equivalent for the two goal boxes. This r e s u l t has been interpreted as i n d i c a t i n g that 2 food p e l l e t s are more than twice as r e i n f o r c i n g as 1 food p e l l e t , since the animals treat the mean of 0 + 2 as greater than 1. It i s t h i s , and si m i l a r , data that has led theorists to suggest that incentive value i s a p o s i t i v e l y accelerated function of reward magnitude (vide Mackintosh, 1974). Inspection of the generalization gradients obtained from food p e l l e t t r i a l s (Figure 7) reveals some properties of rats' perceptions of food quantity that are pertinent to the i n t e r p r e t a t i o n of Leventhal et a l . ' s (1959) r e s u l t s . Assuming that rats perceive food quantity according to simple arithmetic rules, then 2 food p e l l e t s are 1 unit d i f f e r e n t from 1 food p e l l e t and 2 units d i f f e r e n t from 4 food p e l l e t s . Therefore, rats should judge 2 p e l l e t s to be more sim i l a r to 1 p e l l e t 66.7% of the time, and judge 2 p e l l e t s to be l i k e 4 p e l l e t s 33.3% of the time. In fact, 2 p e l l e t s are judged to be more sim i l a r to 4 food p e l l e t s 38 -50% of time, depending upon l e v e l of food deprivation. This r e l a t i o n i s r e f l e c t e d i n the PSE F: i f rats perceived food p e l l e t s on the basis of simple arithmetic r e l a t i o n s , then 106 the mean PSEp would be 2.5. In r e a l i t y , the PSEp of i n d i v i d u a l rats i s consistently lower than 2.5 (mean PSEp ranges from 2.18 - 2.35, SEM ranges from 0.03 - 0.05, depending on deprivation condition). Thus, rats require presentation of less than the mean of 1 and 4 food p e l l e t s before they perceive a given quantity of food p e l l e t s as indistinguishable from 1 and 4 p e l l e t s , rather than being more sim i l a r to 1 p e l l e t . This observation i s s u f f i c i e n t to account for the rats' preference for the goal box containing 2 or 0 p e l l e t s , over that consistently containing 1 p e l l e t . Furthermore, the results of the food quantity generalization te s t i n g indicate that the r e l a t i o n between d i f f e r e n t quantities of food depends upon the absolute quantities. While 2 food p e l l e t s are equally d i f f e r e n t from both 1 and 3 p e l l e t s , the rats perceive 3 p e l l e t s as being more sim i l a r to 4 p e l l e t s than to 2 p e l l e t s . This r e l a t i o n i s the opposite of a p o s i t i v e l y accelerated function of magnitude. On the basis of t h i s observation, one would predict that given the choice between alternating 4 or 2 p e l l e t s i n one goal box versus a consistent 3 p e l l e t s i n the other, rats w i l l choose to enter the goal box containing 3 p e l l e t s . Indeed, t h i s e f f e c t was observed (Leventhal et a l . , 1959): increasing the absolute magnitude of reward, but maintaining arithmetic r e l a t i o n s i d e n t i c a l to the previously discussed experiment, resulted i n rats consistently choosing to enter the goal box containing the intermediate quantity of p e l l e t s . Therefore, the data do not support the view that 107 incentive value of food i s a p o s i t i v e l y accelerated function of number. Indeed, the evidence i s consistent with the present observations: incentive value i s a negatively accelerated function of number. It i s of i n t e r e s t that a given intermediate quantity of food p e l l e t s was perceived by rats as less than an equivalent quantity of tones ( i e . the PSE F was greater than the PSE T). According to Holder and Roberts (1985) and Roberts and Holder (1985), the perception of stimulus duration r e f l e c t s the signal value of the stimulus. As many investigators (eg. Church and Meek, 1984) believe that duration and number perceptions are based on a common mechanism, i t may have been expected that perceptions of food p e l l e t quantities would be biased towards larger numbers than tone quantities. Food p e l l e t s and tones had equivalent signal value i n r e l a t i o n to prediction of reward. This would suggest that perceptions of quantities of food and tones would be equivalent. However, food p e l l e t s have value over and above that associated with reward prediction. Furthermore, asymptotic performance of food p e l l e t number discriminations was higher than that of tone quantity discriminations. Thus, both the inherent value and the signal value was higher for food p e l l e t s than for tones. In addition, a l t e r i n g the signal value of tones by a l t e r i n g the value of the reward tones signal ( i e . food deprivation) did not change the perceptions of tone quantity. It can therefore be concluded that perceptions of quantity are not 108 always correlated with either signal value or inherent value of s t i m u l i . An additional point can be made on the basis of the differences i n food p e l l e t and tone quantity PSE's. Many investigators have argued that counting and timing are amodal i n nature. Discriminations between d i f f e r e n t numbers of s t i m u l i i n one modality may transfer e a s i l y to d i f f e r e n t modalities. However, the present results indicate that the numerical attributes of a b i o l o g i c a l l y s i g n i f i c a n t stimulus (food) are s u b s t a n t i a l l y d i f f e r e n t from the numerical attributes of a stimulus which has only significance with regard to an a r t i f i c i a l l y imposed predictive relationship with reward. Furthermore, i t appears that the a b i l i t y to transfer duration discriminations from one modality to another does not occur i n a l l species (Don Wilkie, personal communication). The putative amodal nature of the numerical attributes of s t i m u l i i s therefore limited. Changes i n food deprivation l e v e l s i g n i f i c a n t l y altered the perception of food quantity, but had no e f f e c t on perceptions of tone quantity, as r e f l e c t e d by the PSE (Figure 9). However, increasing deprivation, which presumably increases the incentive value of food, did not cause the rats to judge a p a r t i c u l a r quantity of food as being more si m i l a r to the food cue with the greatest hedonic value. Rather, increasing deprivation caused rats to judge a given quantity of food as being more similar to one food p e l l e t than to four p e l l e t s . This r e s u l t makes sense from an 109 i n t u i t i v e viewpoint: the hungrier one i s , the smaller a meal seems to be. It suggests that rats do not discriminate between d i f f e r e n t quantities of food on the basis of absolute hedonic value. Rather, i t appears that the cues that rats use to discriminate food quantity are related to some extent to the s a t i a t i n g properties of the food. Thus, a pa r t i c u l a r quantity of food i s perceived p a r t l y i n terms of the proportion i t i s of the t o t a l amount of food a rat requires to become satiated. The hungrier a rat i s , the more food i t requires to be satiated. Therefore, making a rat hungrier decreases the proportion a given quantity of food i s r e l a t i v e to the t o t a l amount that i t requires for s a t i a t i o n . However, i f food p e l l e t s were perceived only i n proportion to the t o t a l amount required for s a t i a t i o n , then the food p e l l e t generalization gradients would be lin e a r i n form. The fact that the generalization gradients f i t a logarithmic function to a better degree than a l i n e a r function indicates that rats perceive two d i f f e r e n t food quantities i n terms of two r e l a t i o n s : the r e l a t i v e proportion of one quantity to the other, and value of each quantity i n r e l a t i o n to the amount of food the rat "wants" to consume. The observation that the response latencies to the various quantities of food p e l l e t s are d i f f e r e n t i a l l y affected by the changes i n body weights (Figure 11) indicates the p o s s i b i l i t y that responding to the various food p e l l e t quantities may be d i f f e r e n t i a l l y affected by the 110 motivational states of the ra t s . In th i s case, the assumption of independence of the PSE from performance ef f e c t s cannot be confidently adopted, because t h i s assumption requires that the treatments do not have d i f f e r e n t i a l e f f e c t s on the performance variables of d i f f e r e n t cue t r i a l s . The int e r p r e t a t i o n that s e l e c t i v e response impairments can account for the change i n food PSE's observed concomitant with alterations i n motivational l e v e l can be ruled out on two counts. F i r s t l y , there was no change i n the slope of the generalization gradients from one food deprivation condition to the other; the slope remained close to the i d e a l slope of 33.3 (assuming that the generalization gradients were lin e a r functions) under a l l 3 deprivation levels (see Figure 10). Secondly, the response latency results indicate that the stimulus which was most affected by the deprivation conditions was the presentation of 4 food p e l l e t s . The change i n PSE indicate that the s h i f t i n responding was biased towards the 4 p e l l e t lever. Thus, the change i n the PSE was i n the opposite d i r e c t i o n of that predicted by a response d e f i c i t s p e c i f i c to the 4 p e l l e t stimulus. I t i s evident from the re s u l t s of t h i s experiment that the range of body weight changes imposed were e f f e c t i v e at a l t e r i n g motivational l e v e l . The latencies to respond to both food and tone . cues increased with body weight, suggesting that the incentive to respond was reduced (Figure I l l 11). Furthermore, the ITI responding decreased i n proportion to the increase i n body weights (Figure 12). Thus, i t i s reasonable to conclude that the e f f e c t s of body weight changes on the PSE of food cues were indeed associated with alterations i n motivational l e v e l . The r e s u l t s of t h i s experiment est a b l i s h that both tone and food quantities are perceived i n terms of r e l a t i v e proportions. Furthermore, the perceptions of food quantities by rats appear to be influenced by t h e i r value, r e l a t i v e to the t o t a l amount required to produce s a t i a t i o n . The l a t t e r observation i s consistent with a behavior-regulation concept of reinforcement as proposed by Hanson and Timberlake (1983). In t h i s formulation, an event i s rewarding to the extent that environmental contingencies produce a deviation from a set-point which determines the preferred l e v e l of the event. Thus, changes i n the value of a reward can occur either through changes i n environmental constraints, or by al t e r a t i o n s of the set-point. As motivational manipulations do not a l t e r environmental constraints, they produce e f f e c t s by changing the set-point. The present r e s u l t s are consistent with the view that rats perceive food quantity i n r e l a t i o n to a set-point for the preferred amount of food consumption, and a l t e r i n g food deprivation changes t h i s set-point. 112 EXPERIMENT 5 A Free-Operant Choice Procedure Assessing Preferences For Food D i f f e r i n g i n Taste but not Quantity The results of the previous experiment demonstrated that rats do not discriminate food quantities on the basis of "pleasantness", i n a manner analogous to human subjects rating sucrose solutions on a pleasantness scale. However, they do suggest that rats assess the dimension of food quantity by attending to both r e l a t i v e proportions of d i f f e r e n t numbers of food p e l l e t s and food value i n r e l a t i o n to the amount the rats would l i k e to consume. This i n t e r p r e t a t i o n of the res u l t s of Experiment 4 can be tested. S p e c i f i c predictions can be made i n the case of the introduction of generalization probes consisting of food p e l l e t s that are equivalent i n quantity, but d i f f e r e n t i n hedonic value. I t i s possible that food quantity discriminations are made purely on the basis of r e l a t i v e proportions. The e f f e c t of changes i n deprivation on quantity discriminations may r e s u l t by some unknown action on some undetermined perceptual process s p e c i f i c for food. If t h i s i s true, then rats w i l l respond to d i f f e r e n t quantities of novel food p e l l e t s which d i f f e r only i n hedonic value from those used i n Experiment 4 i n the same manner to which they respond to the regular food p e l l e t s . A l t e r n a t i v e l y , rats may s e l e c t i v e l y attend to the value of food p e l l e t s i n r e l a t i o n to the amount of food required to 113 produce s a t i a t i o n when discriminating between p e l l e t quantities. The fact that rats appear to perceive d i f f e r e n t numbers of food p e l l e t s i n proportion to each other may r e f l e c t some unknown property of how rats assign value to food. If t h i s i s the case, then introducing various quantities of novel food p e l l e t s that have less hedonic value than those used i n Experiment 4, but do not d i f f e r i n any other fashion, would be responded to as though they were a greater proportion of the t o t a l required to produce s a t i a t i o n than the regular food p e l l e t s were. That i s a given number of novel, less-preferred food p e l l e t s would be perceived as being a greater number than an equivalent quantity of regular p e l l e t s . A l t e r n a t i v e l y , both r e l a t i v e proportion and r e l a t i v e value may contribute to the judgment of food quantity. If th i s i n t e r p r e t a t i o n i s correct, then d i f f e r e n t quantities of novel, less hedonic food p e l l e t s would be discriminated from each other p a r t l y i n proportion to t h e i r r e l a t i v e quantities, but responding would be biased towards the lever cued by 4 regular p e l l e t s . In order to test these predictions, food that d i f f e r e d from the regular food p e l l e t s only i n degree of sweetness were obtained. Size, c a l o r i c , moisture and n u t r i t i o n a l content were equivalent with the previously used food p e l l e t s . However, rather than containing large proportions of sugars, these p e l l e t s were cereal-based. The present 114 experiment examined whether these p e l l e t s d i f f e r e d i n hedonic value from those used i n Experiment 4. Procedure Twelve naive rats were used i n t h i s experiment. The procedure was si m i l a r to that used i n Experiment 1, with 2 differences. During the preference te s t i n g phases,.instead of rewarding lever responses with 1 or 4 food p e l l e t s , reinforcement consisted of 4 sweetened (S) or unsweetened (U) food p e l l e t s . Both types of "dustless precision p e l l e t s " were manufactured by Bioserve Inc.. Sweetened p e l l e t s contained sucrose and dextrose as carbohydrate sources with a t o t a l energy equivalent of 3.8 kcal/g, while unsweetened p e l l e t s contained a mixture of grains as a carbohydrate source with a t o t a l energy equivalent of 3.9 kcal/g. The two foods were si m i l a r i n amounts of proteins (18.0% and 21.2%, respe c t i v e l y ) , carbohydrates (63.0% and 56.6%, respectively) and fats (6.0% and 3.9%, r e s p e c t i v e l y ) . The second difference i s that preference te s t i n g was conducted for 5 sessions on consecutive days before the contingency-reversal. Testing aft e r the contingency-reversal was also conducted on 5 consecutive days. Results As can be observed i n Figure 13, rats displayed a preference for responding on the lever associated with sweet p e l l e t rewards, and t h i s preference resulted i n a s h i f t i n 115 REVERSAL CONSECUTIVE DAILY 15 MIN TEST SESSIONS FIGURE 13. Percent responding on a lever, responses to which were rewarded with 4 sweet food p e l l e t s . Responses on an alternate lever were rewarded with 4 unsweetened p e l l e t s . After 5 days of tes t i n g for preferences, the reinforcement contingency was reversed (REVERSAL) such that responses previously rewarded with sweet p e l l e t s were rewarded by unsweetened p e l l e t s , and vice versa. This reversal was maintained for the subsequent 5 days of tes t i n g . 116 responding to the alternate lever when the reinforcement contingency was reversed. T-tests comparing the % of responding to the lever associated with sweet p e l l e t s to a neutral proportion of 50% indicated that on each of the 5 days of preference te s t i n g p r i o r to the contingency reversal, responding was s i g n i f i c a n t l y greater than 50% (t = 2.33, p < 0.025; t = 2.71, p < 0.02; t = 3.80, p < 0.002; t = 3.58, p < 0.0025; t = 3.78, p < 0.002, respectively, n = 12). Eleven of the twelve rats exhibited more responses to the sweet p e l l e t lever on each of the l a s t 3 days of tes t i n g p r i o r to contingency reversal. However, after the contingency reversal, 4 rats f a i l e d to track the change i n reinforcement (including the one rat that consistently responded to the lever associated with unsweetened p e l l e t s ) . Thus, 3 of the 11 rats that exhibited an i n i t i a l preference for sweet p e l l e t s did not d i f f e r e n t i a t e between sweetened and unsweetened food p e l l e t s a f t e r the contingency reversal. In spite of t h i s , mean choices of the lever associated with S p e l l e t s increased with each subsequent day of tes t i n g post-reversal, u n t i l a s i g n i f i c a n t preference for responding to the S p e l l e t lever was exhibited on the f i f t h day (t = 2.25, p < 0.025). On the f i r s t day of reversal testing, there 'was s i g n i f i c a n t l y more responding to the lever associated with U p e l l e t s (previously associated with S pe l l e t s ) than was expected by chance (t = 2.17, p < 0.05). On the second and t h i r d day, responding to the S p e l l e t lever was not s i g n i f i c a n t l y d i f f e r e n t than that expected by 117 chance (t = 0.25, p > 0.1; t = 1.48, p > 0.05, re s p e c t i v e l y ) . Preference for responding to the S p e l l e t lever on the fourth day was s t a t i s t i c a l l y marginal (t = 1.72, p = 0.057). Thus, the majority of rats exhibited an i n i t i a l preference for S p e l l e t s , and th i s preference was strong enough to produce tracking 5 days after the contingency reversal. Discussion These results e s t a b l i s h that the majority of rats prefer sweetened to unsweetened food p e l l e t s (see Figure 13). This preference i s strong enough i n most cases to prompt animals to track a reinforcement contingency reversal. Insofar as d i f f e r e n t i a l rates of responding to levers associated with d i f f e r e n t i a l reinforcement are i n d i c a t i v e of the hedonic value of re i n f o r c e r s , the present experiment establishes that sweetened food p e l l e t s have a greater hedonic value than unsweetened food p e l l e t s . In th i s regard, i t should be stressed than i n no instance did a rat exhibit a preference for unsweetened p e l l e t s to the extent of 'tracking' a contingency reversal. Furthermore, the rats that did not track the contingency reversal confined responding almost exclusively to one lever, i r r e s p e c t i v e of the type of reward. This indicates that f a i l u r e to track was due to a response bias, and not to a hedonic equivalence between sweetened and unsweetened p e l l e t s . 118 EXPERIMENT 6 E f f e c t of Body Weight on Preferences for Sweetened versus Unsweetened Food P e l l e t s Although Experiment 2 indicated that decreasing motivation by pre-loading animals with various amounts of food had no e f f e c t on preferences of rats for 4 food p e l l e t s over 1 food p e l l e t , there are cannot be concluded that the hedonic values were not altered by changes i n motivation, as discussed previously (see Experiment 2: Discussion). Of p a r t i c u l a r i n t e r e s t i s the p o s s i b i l i t y that, because rats responded almost exclusively on the lever which was reinforced by 4 food p e l l e t s , continued responding to t h i s lever, even under conditions of food s a t i a t i o n , may have been the r e s u l t of habit or the fact that the food retained i t s r e l a t i v e hedonic properties. The previous experiment established that most rats prefer food sweetened with sucrose and dextrose over unsweetened food, when presented i n i d e n t i c a l quantities. I t appears l i k e l y that the preferences for the two kinds of food can be matched by adjusting the quantities delivered per reinforcement. In t h i s way, the e f f e c t of motivational manipulations on the hedonic values associated with quantity and q u a l i t y (taste) can be investigated without the problem of d i f f e r e n t i a l rates of responding to the d i f f e r e n t levers. 119 There have been reports that food deprivation enhances preferences for saccharin, a non-nutritious sweetener (Young and Greene, 1953; Smith and Capretta, 1956; Smith and Duffy, 1957a, 1957b; Capretta, 1962). Thus, increasing food deprivation may s e l e c t i v e l y increase the hedonic properties of sweetened p e l l e t s . On the other hand, i t appears more l i k e l y that hungrier rats would prefer greater quantities, and more satiated rats would prefer the sweeter tas t i n g food. Procedure The 8 rats used i n Experiment 5 which displayed unequivocal preferences for sweetened food p e l l e t s were given additional days of 10 min test sessions i n which the number of sweetened p e l l e t s delivered per reinforcement were decreased by one each day, u n t i l only 1 p e l l e t per reinforcement was delivered. The session was shortened to 10 min from 15 min to minimize within-session s a t i a t i o n e f f e c t s . Four te s t i n g sessions were conducted with the 4:1 (U:S) d i f f e r e n t i a l quantity reinforcement contingencies. The rats were maintained on 85% food deprivation during t h i s phase of the experiment. The rats were then fed such that t h e i r weights were maintained at 75%, 85% or 95% of t h e i r ad libitum body weights, as described i n Experiment 4. One 10 min preference test was conducted at each weight with each S reinforcement 120 consisting of 1 p e l l e t , and each U reinforcement consisting of 4 p e l l e t s . Results Figure 14 displays the results of the progressive decrease i n the number of sweetened p e l l e t s delivered per reinforcement. The asterisks indicate those days (and reinforcement ratios) on which responding to the lever reinforced by sweetened p e l l e t s was s i g n i f i c a n t l y greater than 50% (4/4: t = 12.81, p < 0.0001; 4/3: t = 7.65, p < 0.0001; 4/2: t = 5.39, p < 0.0001; 1 s t 4/1: t = 2.74, p < 0.025; 2 n d 4/1: t = 0.76, p > 0.1; 3 r d 4/1: t = 0.72, p > 0.1; 4 t h 4/1: t = 0.38, p > 0.1, n = 8, where #/# refers to the respective quantities of U/S delivered per reinforce-ment ) . The e f f e c t s of changing body weights on preferences are presented i n Figure 15. The e f f e c t of changing body weights was s i g n i f i c a n t (F(2,14) = 5.21, p < 0.025; R(2,6) = 5.50, p < 0.05). Individual comparisons indicated that reducing rats' body weights to 75% of ad libitum weights s i g n i f i -cantly reduced preferences for the sweetened p e l l e t s , s h i f t i n g preferences from near the iso-hedonic point to a preference for quantity (85% x 75%: F(l,7) = 11.04, p < 0.02). Increasing body weights to 95% did not have a s i g n i f i c a n t e f f e c t on preferences (85% x 95%: F(l,6) = 0.61, p > 0 .1) . 121 to CO C £ O L_ 00 LLI C O z o D_ — " -i I 4:4 4:3 4:2 4:1 4:1 4:1 4:1 CONSECUTIVE DAILY 10 MIN TEST SESSIONS FIGURE 14. Percent responding on a lever, responses to which were rewarded with sweet food p e l l e t s . Responses on an alternate lever were rewarded with 4 unsweetened food p e l l e t s . The number of sweet p e l l e t s delivered per response was decreased on consecutive days, as indicated by the unsweetened:sweet p e l l e t r a t i o s on the abscissa. 122 FIGURE 15. Percent responses to a lever rewarded with 1 sweet food p e l l e t . Responses on an alternate lever were rewarded with 4 unsweetened food p e l l e t s . Preferences were tested during 3 consecutive periods. During each period animals were maintained on one of three food deprivation conditions (75%, 85% or 95% of body weights while animals were given ad libitum access to food). * S i g n i f i c a n t l y d i f f e r e n t from 85% deprivation condition. 123 Discussion Decreasing the quantity of sweetened p e l l e t s delivered per each reinforcement was found to be an e f f e c t i v e method of assessing the isohedonic r a t i o of sweetened and unsweetened p e l l e t s . I t was determined that 4 unsweetened p e l l e t s were considered by rats to be equivalent i n hedonic value (isohedonic) to 1 sweetened p e l l e t (see Figure 14). Increasing hunger increased rats' preferences for 4 unsweetened p e l l e t s , as assessed by the r e l a t i v e responding to either of 2 levers: responses to one being reinforced with 4 unsweetened p e l l e t s , and responses to the other being reinforced with 1 sweetened p e l l e t s (see Figure 15). Consistent with the results of Experiment 2, decreasing hunger had no e f f e c t on preferences, although a s l i g h t ( s t a t i s t i c a l l y i n s i g n i f i c a n t ) tendency to s h i f t i n g responding to the lever reinforced with 1 sweet p e l l e t was evident. I t remains possible that s a t i a t i n g the rats further may have produced a s h i f t i n preference towards the sweetened p e l l e t s . The r e s u l t s of t h i s experiment e s t a b l i s h that motivational determinants can influence hedonic properties of food, and that t h i s factor i s more i n f l u e n t i a l with respect to quantity than with q u a l i t y ( i e . taste) of food. It i s of further i n t e r e s t to note that very l i t t l e e f f e c t was observed aft e r decreasing food deprivation, s i m i l a r to 124 the lack of e f f e c t found i n Experiment 2. Increasing food deprivation was more e f f e c t i v e at a l t e r i n g choice. 125 EXPERIMENT 7 Generalization of Food Quantity Discriminations to Food D i f f e r i n g Only i n Taste and Hedonic Value The purpose of t h i s experiment was to determine to what degree number and hedonic value served as discriminative cues i n the performance of the discrimination and generalization tasks of Experiments 3, 4 and 12. To accomplish t h i s , generalization gradients were obtained for 1, 2, 3 and 4 p e l l e t s of a food that d i f f e r e d i n taste and hedonic value, but not i n energy value. These generalization gradients were obtained during the same session as generalization gradients for sweetened p e l l e t s for comparison, and to avoid e x t i n c t i o n of the discrimination. As described i n the introduction to Experiment 5, a number of predictions can be made regarding the results of presenting generalization probes consisting of novel food p e l l e t s that d i f f e r from regular p e l l e t s only i n hedonic value, depending upon the nature of the stimulus attributes that rats attend to when trained to discriminate between quantities of food. Rats may attend exclusively to the proportional differences between d i f f e r e n t quantities of food when performing a food quantity discrimination. If t h i s i s so, then the generalization gradients obtained from equivalent quantities of unsweetened p e l l e t s should be indistinguishable from the sweetened food p e l l e t 126 generalization gradients. No change i n either PSE or slope would be predicted. It i s also possible that rats attend exclusively to the value of food p e l l e t s r e l a t i v e to the t o t a l amount required to induce s a t i a t i o n . Assuming that the preferences for sweetened p e l l e t s exhibited i n a free-operant choice procedure (Experiment 5) r e f l e c t differences i n the amount required for s a t i a t i o n , then t h i s amount i s lower for unsweetened p e l l e t s then for sweetened p e l l e t s . In the case that food quantity perceptions are determined larg e l y by value, responding to unsweetened food p e l l e t cues would be directed predominantly to the lever cued by 4 sweetened p e l l e t s . This i s because a given quantity of unsweetened p e l l e t s represents a larger proportion of the amount required for s a t i a t i o n than an equivalent amount of sweetened p e l l e t s . Such a bias i n responding would be evident as a reduction i n the PSE for unsweetened p e l l e t s , r e l a t i v e to sweetened p e l l e t s . Furthermore, value i n r e l a t i o n to some constant amount of food would be r e f l e c t e d by l i n e a r , not proportional, unsweetened p e l l e t generalization gradients. The r e s u l t s of Experiment 3 indicated that rats attend to both r e l a t i v e proportion and r e l a t i v e value when discriminating between d i f f e r e n t quantities of sweetened p e l l e t s . If t h i s i n t e r p r e t a t i o n i s correct, then the generalization gradients obtained from unsweetened p e l l e t generalization probes would display proportional differences 127 between quantities and would tend to e l i c i t more responses to the lever cued by the 4 sweetened p e l l e t s . This case would be r e f l e c t e d by a decrease i n the unsweetened p e l l e t PSE, and the unsweetened p e l l e t generalization gradients would be proportional functions. An additional factor must be considered. Rats increase t h e i r l i c k i n g rate for sucrose i n proportion to the logarithmic increase i n concentration (Davis, 1973). Therefore, rats may judge differences between various quantities of sweetened p e l l e t s by attending to proportional differences i n sweetness, rather than quantity. To the extent t h i s i s true, rats w i l l f a i l to transfer the sweet p e l l e t discrimination learning to unsweetened p e l l e t s . A f a i l u r e to transfer the discrimination learning would be r e f l e c t e d i n a decrease i n the slopes of the unsweetened p e l l e t generalization gradients, r e l a t i v e to sweetened p e l l e t generalization gradients. This experiment provides a method of determining to what extent various stimulus properties are attended to by rats when discriminating between d i f f e r e n t quantities of food. Furthermore, the hypothesis proposed to account for the r e s u l t s i n Experiment 4 makes e x p l i c i t predictions concerning the outcome of the present experiment. 128 Procedure Sixteen of the rats used i n Experiment 3 and 4 were used i n the present experiment. The rats were given 2 weeks of discrimination t r a i n i n g , i d e n t i c a l with that used i n Experiment 3 and 4 with the exception that tone t r i a l s were omitted. Each session consisted of t r i a l s i n the order of F4 F 4 F]_ F 4 F1 F1 F1 F 4 Fi_ F]_ F 4 F 4 F1 F 4 F^L F 4 F 4 F1 F 4 Fi_, repeated for a t o t a l of 40 t r i a l s . The rats were then given 5 test sessions on consecutive days i n which they were presented with 1, 2, 3, and 4 sweetened and unsweetened food p e l l e t t r i a l s . Correct responses after presentations of 1 and 4 sweet p e l l e t s were reinforced on a 75% reinforcement schedule. Responses during t r i a l s i n which other quantities or unsweetened p e l l e t s were presented were not reinforced. Each session consisted of si x t y t r i a l s , 15 of each of the 1 and 4 sweetened p e l l e t t r i a l s , and 5 each of a l l generalization t r i a l s . The types of cues were as evenly d i s t r i b u t e d between the f i r s t 30 and second 30 t r i a l s as possible. Thus, each session could be divided into two f a i r l y equivalent halves for the purposes of examining e f f e c t s of food consumption during the sessions on the generalization gradients (that i s , comparing generalization gradients derived from the f i r s t half of the sessions to the those determined from second). The order of cues was as follows, where S = sweetened p e l l e t s , U = unsweetened p e l l e t s and the subscript refers to the quantity of p e l l e t s delivered: S 4 S 2 Si_ U 2 S 4 S3 S 4 Ui_ S]_ U 4 S 4 S 2 S 4 129 S3 S 1 U 4 S 1 U 2 S 4 S3 S]_ u3 s4 s2 s4 u1 sx U 4 Si_ u4 s4 s3 s1 u l s l u3 s l u4 s4 u3 s l u l s4 u2 s l u3 s4 s2 s l u3 s4 s2 s4 U 2 S i S 3 S 4 U 2 S i Ui_. Results The results were analyzed separately for the f i r s t and second halves of the sessions, combining the data across the 5 sessions. In one case only the results from the f i r s t 2 complete sessions were analyzed to determine whether or not cer t a i n differences between sweetened (S) and unsweetened (U) p e l l e t s were evident before rats had an opportunity to learn that U p e l l e t s did not sig n a l reward a v a i l a b i l i t y . The sessions were divided into halves because as a l l t r i a l types involved the presentation of food cues (72-73 p e l l e t s during the f i r s t half of the session), i t was considered possible that within-session s a t i a t i n g e f f e c t s might have been evident. The generalization gradients from a l l sweetened and unsweetened p e l l e t t r i a l s are displayed i n Figure 16. The PSE's and slopes of the generalization gradients were calculated as i n Experiment 4 (Results). Analysis of variance applied to the PSE data revealed that the main e f f e c t of p e l l e t sweetness was s i g n i f i c a n t , while the main ef f e c t s of session half and the sweetness X session half i n t e r a c t i o n were not. The PSE from the unsweetened p e l l e t t r i a l s (PSErj) was s i g n i f i c a n t l y lower than the PSE (PSE S) from the sweetened p e l l e t t r i a l s (mean PSE S = 2.10, mean PSEy = 1.20; F (1,15) =4.65, p < 0.05,). Planned comparisons revealed that there was a s i g n i f i c a n t 130 , j 1 2 3 NUMBER OF FOOD PELLETS FIGURE 16. Generalization gradients obtained from presentation of various quantities of sweet (S) and unsweetened (U) food p e l l e t s . Unsweetened p e l l e t probes occurred interspersed among sweet p e l l e t probes. Seventy-five percent of the correct responses afte r 1 or 4 sweet p e l l e t s were rewarded. 131 drop i n the PSE S from the f i r s t to the second half (1st half mean = 2.17, 2nd half mean = 2.04; F(l,15) = 29.6, p < 0.0005), but not i n the PSEg (1st half mean = 1.28, 2nd half mean = 1.12, F(l,15) = 0.07, p > 0.1). There was a s i g n i f i c a n t main e f f e c t of p e l l e t type on the slopes of the generalization gradients (F(l,15) = 58.8, p < 0.0001), with the slope from the unsweetened p e l l e t generalization gradient (SLOPEy) s i g n i f i c a n t l y lower than the slope (SLOPEs) of the sweetened p e l l e t generalization gradient (mean SLOPE s = 27.7, mean SLOPEy = 13.08). Slopes did not vary for either the sweetened or the unsweetened generalization gradients from the f i r s t to the second halves of the sessions (F(l,15) = 3.82, p > 0.05). The latencies to respond after presentation of the various cues, expressed as the mean for the f i r s t and second halves of the combined sessions are displayed i n Figure 17. The i n t e r a c t i o n among the session half, p e l l e t type and p e l l e t quantity was not s i g n i f i c a n t . However, a l l three main eff e c t s were s i g n i f i c a n t ( f i r s t halves of sessions X second halves: F(l,15) = 10.89, p = 0.005; sweetened p e l l e t s X unsweetened p e l l e t s : F(l,15) = 41.73, p < 0.0002; p e l l e t quantity: F(l,15) = 13.98, p < 0.005, R(3,13) = 12.14, p < 0.001). Response latencies a f t e r presentation of unsweetened p e l l e t s were longer than after sweetened p e l l e t s , and latencies a f t e r both kinds of p e l l e t cues were longer i n the second halves of the sessions than during the f i r s t . Individual comparisons of the p e l l e t quantities revealed 132 O LJJ CO, CO Ld O z St CO z o Q_ CO Ld or 40 30-20-10 Legend o s PELLETS • U PELLETS I I I 1 1 1 2 3 4 NUMBER OF FOOD PELLETS FIGURE 17. Response latencies i n seconds (SEC) during t r i a l s i n which various quantities of sweet (S) and unsweetened (U) food p e l l e t s were presented. Unsweetened p e l l e t probes occurred interspersed among sweet p e l l e t probes. Latencies after a l l numbers of U p e l l e t s were s i g n i f i c a n t l y longer than after S p e l l e t s . 133 that there were no s i g n i f i c a n t differences i n latencies to respond among the 2, 3 and 4 p e l l e t cues (2x3: F(l,15) = 1.14, p > 0.1; 2x4: F(l,15) = 1.92, p > 0.1; 3x4: F(l,15) = 0.93, p > 0.1), but the latencies were s i g n i f i c a n t l y shorter aft e r the 1 p e l l e t cues than after a l l the others (F(l,15)=37.06, p < 0.0002). The response bias index was calculated as the ITI responses directed to the 4-pellet cued lever divided by the t o t a l ITI responses. Total ITI responses over the 5 sessions ranged between 30 and 1231. The response bias index mean was 0.42 + 0.07. This was not s i g n i f i c a n t l y d i f f e r e n t from chance leve l s of responding (t = 1.16, p > 0.1, n = 16). Figure 18 depicts the generalization gradients obtained from the f i r s t 2 sessions. The data from one rat was excluded from analysis as i t s U generalization gradients exhibited a negative slope [-8.0] during the f i r s t 2 sessions. The PSEy was s i g n i f i c a n t l y lower than the P S E S (mean P S E S = 2.02, mean PSEy = 1.23; F(l,14) = 5.89, p < 0.05). The SLOPEy from the generalization gradients derived from the f i r s t two sessions was also s i g n i f i c a n t l y lower than the S L O P E s (mean S L O P E s = 27.0, mean SLOPErj = 18.27; F(l,14) = 14.16, p < 0.0025), although the magnitude of the difference was much smaller than when a l l sessions were included i n the analysis (see above). Response latencies to U p e l l e t s were s i g n i f i c a n t l y longer than those to S p e l l e t s on the f i r s t 2 sessions (mean S response latencies = 8.02 134 a: 1 0 0 i u i 1 , , , , 1 2 3 4 NUMBER OF FOOD PELLETS FIGURE 18. Generalization gradients obtained from the presentation of various quantities of sweet (S) and unsweetened (U) food p e l l e t s . Unsweetened p e l l e t probes occurred interspersed among sweet p e l l e t probes. Gradients were derived only from t r i a l s conducted during the f i r s t 2 sessions. 135 sec, mean U response latencies = 22.03 sec; F(l,14) = 16.3, p < 0.005). As response latencies for unsweetened p e l l e t t r i a l s were s i g n i f i c a n t l y longer than for sweetened p e l l e t t r i a l s i t was considered possible that the differences i n the PSE's and slopes from the two types of generalization gradients may have been related to the long response latencies exhibited during unsweetened p e l l e t . t r i a l s . To examine th i s p o s s i b i l i t y , the P S E ' s and slopes were calculated only for t r i a l s over a l l sessions with response latencies of less than 30 seconds. The mean P S E " s (PSE S = 2.15, PSEu = 1.38) exhibited the same r e l a t i o n as observed when long response latency t r i a l s are excluded: the P S E J J was s i g n i f i c a n t l y lower than the PSE S (F(l,14) = 11.3, p < 0.01). A decrease i n the SLOPEy r e l a t i v e to SLOPE s was also observed when t r i a l s of response latencies longer than 30 sec (SLOPE s = 28.08, SLOPEy = 15.16) were excluded from analysis (F(l,14) = 40.4, p < 0.0002). Discussion The generalization gradients derived from unsweetened p e l l e t t r i a l s (U) d i f f e r e d from those obtained from the sweetened p e l l e t t r i a l s (S) i n a number of c h a r a c t e r i s t i c s (see Figure 18). A given quantity of unsweetened p e l l e t s was perceived by rats to be greater than an equivalent quantity of sweetened p e l l e t s , as indicated by the decrease i n the PSEy r e l a t i v e to the PSE S . This i s the re s u l t predicted by 136 the view that rats' perceptions of food quantity are determined to some degree by the hedonic value of food, r e l a t i v e to the amount of food required for s a t i a t i o n . I t i s possible that rats learned that the unsweetened p e l l e t t r i a l s signaled non-reinforcement. The observed res u l t s may r e f l e c t processes related to conditioned i n h i b i t i o n , and imply l i t t l e concerning discrimination and generalization of sweetened p e l l e t s . If t h i s were so, then the difference i n PSE should not be present, or be minimal, during the f i r s t two sessions, before the rats had an opportunity to learn that U p e l l e t s were not correlated with reinforcement. However, a magnitude of decrease i n the PSEy equivalent to that observed over a l l the sessions was apparent during the f i r s t 2 sessions. Thus, the difference i n PSE between the two p e l l e t types cannot be attributed to learning of d i f f e r e n t i a l reinforcement. The SLOPEy was also decreased, r e l a t i v e to the SLOPE s. Furthermore, t h i s decrease i n performance was apparent, although to a much smaller degree, during the f i r s t two sessions. One rat was an exception to t h i s trend: i t responded to the 4 p e l l e t lever more frequently af t e r presentation of 1 and 2 U p e l l e t s than after 3 U p e l l e t s during the f i r s t 2 sessions, r e s u l t i n g i n a negative slope. The reasons for t h i s apparent aberration are not clear. The decrease i n performance accuracy of the other rats suggests that taste contributed to a small degree to the discrimination. However, i t should be noted that "accuracy" 137 of responses to the U p e l l e t s during the f i r s t 2 sessions were above 70% for the 1 U p e l l e t stimulus and 85% for the 4 U p e l l e t stimulus. It can therefore be concluded that quantity was the main discriminative cue. The decrease i n performance accuracy and the i n PSE cannot be attributed to rats responding to a lever to which they are biased towards, i n the absence of e x p l i c i t reinforcement contingencies. No bias towards the lever associated with the 4 p e l l e t cue was observed. The observation that response latencies to U cues were s i g n i f i c a n t l y longer than those to S cues indicates that the response a c t i v a t i n g (incentive) e f f e c t s of S cues did not generalize to the U cues. This suggests that the acquired incentive e f f e c t s of S cues were associated with taste, as t h i s was the only i d e n t i f i a b l e sensory difference between the two s t i m u l i . I t also appeared possible that the difference i n PSE may be attributable to differences i n response latencies. Indeed, increasing the duration between a stimulus presentation and the relevant response a l t e r s the perception of the stimulus duration (Spetch and Wilkie, 1983). However, t h i s change i n perception i s i n the opposite d i r e c t i o n of the one observed i n the present experiment. That i s , perception of stimulus duration decreases with increased delay ("subjective shortening", Spetch and Wilkie, 1983). Furthermore, i t i s unclear whether the increase i n U p e l l e t t r i a l response latencies are the r e s u l t of an increase i n the time between p e l l e t consumption and the 138 response, an increase i n the time of p e l l e t d e l i v e r y and consumption, or both. However, i t was found that the PSE e f f e c t was s t i l l evident when t r i a l s of response latencies greater than 30 seconds were excluded from analysis. Indeed, analysis of the PSE data excluding t r i a l s with response latencies of greater than 15 seconds revealed that the PSE ef f e c t remained (data not shown), although t h i s finding i s less trustworthy due to the difference i n number of S and U t r i a l s available for analysis when t h i s c r i t e r i o n i s applied. Another finding of the present experiment was that the PSE S exhibited s a t i a t i o n e f f e c t s ( i e . decreased from the f i r s t to the second halves of the sessions). This confirms the observation of a decrease i n PSE S with an increase i n body weight (Experiment 4). A decrease i n the mean PSEy of a similar magnitude was observed, but was not s i g n i f i c a n t . The res u l t s of t h i s experiment support the view that the perception of food quantity i s modified by the value of food r e l a t i v e to the amount of food required for s a t i a t i o n . The fact that a prediction of t h i s i nterpretation, developed to explain the results of Experiment 4, was confirmed, and that t h i s r e s u l t was not predictable by other interpretations i s convincing evidence. However, there i s a p o s s i b i l i t y that the difference i n the PSEy and PSE S i s related to a novelty e f f e c t , although why th i s should be i s not c l e a r . 139 EXPERIMENT 8 Generalization of Tone Quantity Discrimination to a Novel Tone Experiment 7 established that proportional quantity, value and taste (to some degree) contribute to the stimulus complex that rats attend to when discriminating between d i f f e r e n t quantities of food. However, i t i s possible that the generalization of discrimination t r a i n i n g to a novel stimulus a l t e r s the perception of quantity because of some unknown general property of perceptual processes. In order to investigate t h i s p o s s i b i l i t y , generalization gradients were obtained by introducing probes of various quantities of novel tones during the tone quantity discrimination described i n Experiment 3. Stimuli of the same modality, but of a d i f f e r e n t frequency, were chosen to most c l o s e l y resemble the presentation of unsweetened p e l l e t s i n the previous experiment. That i s , unsweetened p e l l e t s are of the same modality as sweet p e l l e t s , but d i f f e r i n i n t e n s i t y , having a less intense sweet taste. Thus, tones of less i n t e n s i t y (lower frequency) were used as probes of a novel stimulus within the same modality of the "regular" tones. Probe t r i a l s of intermediate quantities of the "regular" tone were also conducted so that the generalization gradients from both regular and novel tones could be compared. 140 Procedure The rats from Experiment 7 were given seven t r a i n i n g sessions, each consisting of 40 t r i a l s . Half of the t r i a l s were of the 1-tone type (T]_), and the remaining were of the 4-tone kind (T4), presented i n the same sequence as the food p e l l e t t r i a l s during t r a i n i n g sessions described i n Experiment 7. Correct responses were reinforced with a 75% schedule. Generalization gradients were then obtained by presenting 5 of each of the following s t i m u l i , N 2 N3 N4 R 2 R3, interspersed among 15 presentations of each of R^  and R4, where N = novel tones, R = regular tones, and the subscript refers to the number of tones presented per t r i a l , on each of two sessions. Order of t r i a l s was the same as i n Experiment 7, substituting regular tones for sweet p e l l e t s and novel tones for unsweetened p e l l e t s . Novel tones were produced by a tone generator that emitted a wide frequency of sounds, r e s u l t i n g i n a "buzzing" sound, i n contrast to the pure regular (3200 Hz) tone. The procedure for the determination of generalization gradients was i d e n t i c a l as the one followed i n Experiment 7, with exceptions of number of sessions and kinds of s t i m u l i . Results Seven of the sixteen rats retrained on the tone discrimination f a i l e d to regain accuracy levels of 70% or greater. These rats were rejected from further study. The generalization gradients for the remaining 9 rats are 141 presented i n Figure 19. PSE's and slopes were calculated as described i n Experiment 4. There was no s i g n i f i c a n t difference between the PSE obtained from the regular tone generalization gradient (mean PSE R = 1.70) and the PSE (mean PSE N = 1.95) derived from the novel tone generalization gradient (F(l,8) = 2.92, p > 0.1). Furthermore, there was no s i g n i f i c a n t difference between the slopes of the two generalization gradients (mean SLOPE R = 19.33, mean SLOPE N = 22.89; F(l,8) = 3.09, p > 0.1). Discussion As can be observed i n Figure 19, the rats transferred the tone quantity discrimination to a novel tone completely. Therefore, i t can be concluded that there i s no general perceptual process that can account for the results obtained i n Experiment 7. In addition, t h i s evidence supports the view that the discrimination of tone quantities i s performed by a process d i f f e r e n t from that of food quantity discriminations, as the introduction of stimuli d i f f e r i n g only i n i n t e n s i t y resulted i n very d i f f e r e n t response patterns depending on whether or not they were food or tones. 142 FIGURE 19. Generalization gradients obtained from the presentation of various quantities of regular (R) 3200 Hz and novel (N) mixed frequency tones. N probes occurred interspersed among R probes. Seventy-five percent of the correct responses after 1 and 4 R tones were rewarded. 143 EXPERIMENT 9 Conditioned Reinforcement Properties of Tone Stimuli An assumption of the forgoing experiments i s that the tone st i m u l i used i n the discrimination and generalization procedures are devoid of r e i n f o r c i n g or incentive value. However, thus far no evidence has been presented to j u s t i f y t h i s assumption. If the tone stimuli have incentive value, either because of some innate property, or because i t has been acquired as a re s u l t of the association of tones with reinforcement, then the conclusion of some of the previous experiments that rats attend to the incentive value of food st i m u l i when discriminating quantities may be i n v a l i d . There are a variety of methods for testing for rewarding properties of a stimulus. Acquisition of a new response on which presentation of the relevant stimulus i s contingent, an increase i n responding for another reward upon presentation of the stimulus, resistance to extinction i n the presence of the stimulus, or the maintenance of responding during extinction i f the stimulus i s made contingent upon the response, would each indicate that the stimulus has rewarding properties. The l a t t e r method was used i n the present case. To investigate the incentive or conditioned r e i n -forcement properties of tones, half of the previously trained rats were given e x t i n c t i o n t r a i n i n g , and the remaining rats were given e x t i n c t i o n t r a i n i n g during which 144 responses to one lever resulted i n the production of one tone, and responses to the alternate lever produced 4 tones. If tones are e f f e c t i v e r e i n f o r c e r s , either because of inherent or secondary reinforcement properties, then the rats that were tested with a response-tone contingency should exhibit a resistance to extinction, r e l a t i v e to those rats without the contingency. Procedure The 9 rats from Experiment 8 were given two 30 min sessions on consecutive days i n which no discriminative s t i m u l i were delivered, but responses were recorded. Presentation of 1 or 4 tones was made contingent upon responses directed towards the lever associated with reinforcement aft e r presentation of 1 or 4 tones (respectively) i n Experiments 3, 4 and 8, for four of the nine rats. The responses of the remaining 5 rats were recorded, but had no programmed consequences. The houselight, which had served as a t r i a l cue i n previous discriminations, was l e f t on during the f u l l 30 min of each session. This was to maximize the p o s s i b i l i t y of observing a pos i t i v e r e s u l t by reducing the differences i n cues between the e x t i n c t i o n tests and the t r a i n i n g conditions. As discussed i n the Introduction, extinction i s f a c i l i t a t e d by increasing the number of cues that are d i f f e r e n t between tr a i n i n g and extinc t i o n . 145 Results The mean number of responses directed to each of the two levers for both the group which experienced response-tone contingencies and that which experienced no response-related programmed consequence are displayed i n Figure 20. No s i g n i f i c a n t differences i n the number of responses emitted between the tone and no-tone groups were apparent (F(l,8) = 0.63, p > 0.1). There was also no s i g n i f i c a n t difference between the number of responses directed towards the lever associated with 1 tone and the lever previously associated with 4 tones i n either group ( F J N I J ; E R A C T I O N ( ^ ' ^ ^ = 0.27, p > 0.1). There was no s i g n i f i c a n t days e f f e c t (F(l,7) = 4.77, p > 0.05), and no i n t e r a c t i o n term involving days was s i g n i f i c a n t . Discussion The finding that the response rate was not s i g n i f i c a n t l y increased i n rats that were "reinforced" with tones r e l a t i v e to those which had no programmed consequences contingent on responses indicate that the tones did not possess s i g n i f i c a n t r e i n f o r c i n g properties, nor did they acquire properties of secondary reinforcement as a r e s u l t of extensive t r a i n i n g as discriminative s t i m u l i . Furthermore, that the number of responses were low, and did not decrease from the f i r s t to the second session (that i s , did not exhibit e x t i n c t i o n - l i k e properties) supports t h i s conclusion. Thus, i t appears that the presumed n e u t r a l i t y of tones with regards to hedonic or incentive value was 146 100 75 50-Legend EZ2 T l L E V E R + TONE CD T 1 L E V E R - TONE [ S l T4 L E V E R + TONE W U L E V E R —TONE DAY 1 DAY 2 FIGURE 2 0 . Responses to levers during 2 30 min free-operant tests for reward properties of tones. Responses to levers which were previously signalled by 1 (Tl) or 4 (T4) tones were rewarded with either 1 (Tl LEVER+TONE) or 4 tones (T4 LEVER+TONE). Responses by control groups were not rewarded (Tl-TONE, T4-TONE). No s i g n i f i c a n t differences were observed between experimental and control groups. 147 j u s t i f i e d . This establishes the appropriateness of using tones as a control stimulus to which behavioral responses to food s t i m u l i may be compared. 148 EXPERIMENT 10 Role of P e l l e t Dispenser " C l i c k s " In P e l l e t Discriminations The evidence accumulated i n Experiments 3, 4, 7, 8 and 9 are indicate that the discrimination of d i f f e r e n t quantities of food p e l l e t s i s both qu a n t i t a t i v e l y and q u a l i t a t i v e l y d i f f e r e n t from the discrimination of quantities of tones. Nevertheless, there remains the p o s s i b i l i t y , however s l i g h t , that rats attended to the sounds of the p e l l e t dispenser mechanism during stimulus presentations, and t h i s sound may have served as the discriminative stimulus. In fact, responses during extinction of an operant task can be increased by making the sound of the feeder mechanism contingent upon the response, without a c t u a l l y d e l i v e r i n g food (Bugelski, 1938). It i s conceivable that the sounds of the feeder mechanism may have acquired incentive value, which may also serve as a discriminative cue. Thus, i t appeared possible, a l b e i t u n l i k e l y , that rats discriminated d i f f e r e n t p e l l e t quantities exclusively on the basis of the sounds of the p e l l e t dispenser. If t h i s were so, then the interpretations of the results of the food p e l l e t generalization experiment are questionable. There-fore, the degree to which accurate responding could be e l i c i t e d by the feeder c l i c k s , without the presence of the food p e l l e t s t i m u l i was assessed. 149 Procedure The sixteen rats used i n Experiment 8 were given one day (40 t r i a l s ) of re t r a i n i n g to the food p e l l e t quantity discrimination task (see Experiment 3), with a 75% reinforcement schedule for correct responses. The next day, they were given 20 t r i a l s i n every way sim i l a r to a t r a i n i n g session with the exception that p e l l e t s were delivered into a cup outside of the inner operant box, inaccessible to the rats, and that no responses were rewarded. Thus, a t r i a l consisted of the t r i a l cue l i g h t illuminating, followed by the sound of the feeder mechanism providing either one or four c l i c k s . Each t r i a l ended aft e r a response was made, and the response was recorded as being either correct or incorrect, depending on whether i t was made to the lever which i n the normal discrimination t r a i n i n g was associated with the number of feeder c l i c k s presented. The t r i a l s occurred i n a repeated series of ten; thi s enabled the response accuracy during the f i r s t 10 t r i a l s to be compared with that during the l a s t 10 t r i a l s , to determine i f extinction of the discrimination occurred. Results Five of the sixteen rats f a i l e d to exhibit response accuracy of 70% or better on both one p e l l e t and 4 p e l l e t t r i a l s during the re t r a i n i n g session; data from these rats were excluded from further analysis. The mean percent correct on the test day was 52.5 (+ 3.3, SEM), which i s not s i g n i f i c a n t l y d i f f e r e n t from chance l e v e l (t = 0.78, p > 150 0.1, n = 11). However, the percent correct responses to the single c l i c k stimulus was 64.2% (+ 4.7) which was s i g n i f i c a n t l y above chance l e v e l (t = 3.01, p < 0.01). In comparison, the accuracy of responding to the 4-click stimulus was not s i g n i f i c a n t l y d i f f e r e n t from chance (mean = 42.7% + 5.2, t = 1.39, p > 0.1). This e f f e c t i s not attributable to a response bias; on the t r a i n i n g session, the accuracy of responding to the 4-pellet stimulus was s i g n i f i c a n t l y higher than to the 1-pellet stimulus (mean % correct: 86.4 + 2.6 (1-pellet t r i a l s ) , 98.2% + 1.0 (4-pellet t r i a l s ) ; t D = 4.22, p < 0.005). Analysis of the percent responding to the lever which was previously associated with reinforcement predicted by the 1-pellet stimulus (Li_), regardless of the number of c l i c k s presented, revealed that the majority of the responses during the f i r s t 10 t r i a l s were directed to L]_ (mean % L]_ responses: 75.5 + 3.7 ( f i r s t 10 t r i a l s ) , 45.5 + 5.3 (second 10 t r i a l s ) , F(l,10) = 30.9, p < 0.0005), r e l a t i v e to the second 10 t r i a l s . The mean % L]_ responses were s i g n i f i c a n t l y above chance l e v e l (t = 6.96, p < 0.0001) during the f i r s t , but not the second (t = 0.86, p > 0.1) ten t r i a l s . Discussion This experiment established that the sound of the feeder c l i c k s did not contribute to the stimulus complex attended to by rat s trained to discriminate between two quantities of food p e l l e t s . Response accuracy was at chance 151 l e v e l when only the sound of the food p e l l e t dispenser mechanism was supplied as a cue. However, i t was observed that responses aft e r presentation of no food p e l l e t s (sound of p e l l e t dispenser only) were directed towards the lever previously associated with the 1-pellet discriminative stimulus. Therefore, i t i s apparent that rats perceive no p e l l e t s as being more sim i l a r to 1 p e l l e t than to 4 p e l l e t s . In fact, t h i s experiment provided an extension of the generalization gradient determinations, i n d i c a t i n g that the discrimination operated on a continuum: 0 p e l l e t s were perceived by rats to be more simi l a r to 1 p e l l e t than to 4 p e l l e t s , 3 p e l l e t s were generally perceived to be more sim i l a r to 4 p e l l e t s , and 2 p e l l e t s were perceived as either more similar to 1 p e l l e t than 4, or close to being as s i m i l a r to 1 as to 4. It i s important to note that responses were directed to the "1-p e l l e t lever" during the f i r s t 10 t r i a l s regardless of the number of c l i c k s of the p e l l e t dispenser mechanism produced on a p a r t i c u l a r t r i a l ; the sound of the p e l l e t dispenser was i r r e l e v a n t to the choice of which lever to press. This f i n d i n g appears to contradict the view of Church and Meek (1984) that the numerical attributes of s t i m u l i are "amodal". However, i t i s possible that the dispenser c l i c k s became associated with the presentation of food. Although not acquiring r e i n f o r c i n g properties, the presence of c l i c k s may have provided contextual cues i n d i c a t i n g the relevance of food p e l l e t discriminations. Thus, the rats may have 152 responded to c l i c k s i n a manner relevant to the food discrimination, responding with a bias to the lever associated with the quantity of p e l l e t s most sim i l a r to zero. It may also be relevant that animals respond with a "short duration" bias when tested under extinction (Roberts and Holder, 1985). However, the bias i n responding was towards the small quantity only for food s t i m u l i . If considered with regard to tone quantity discriminations, the bias i n responding was i n the d i r e c t i o n of large quantities. I t therefore appears l i k e l y that the c l i c k s did indeed serve as contextual cues, i n d i c a t i n g the relevance of the food p e l l e t discrimination rather than the tone discrimination. 153 EXPERIMENT 11 The E f f e c t of a Dopamine Agonist and a Dopamine Antagonist On the Generalization Gradients of Food P e l l e t s and Tones The previous experiments established that the generalization gradients of animals trained to discriminate between d i f f e r e n t quantities of food p e l l e t s r e f l e c t the perception by rats of the hedonic value of that food. A l t e r i n g the value of food p e l l e t s by decreasing the l e v e l of food deprivation or by reducing the sweetness produced systematic decreases i n the PSEp. Treatments that enhance the incentive value of food should therefore produce an increase i n the PSE derived from food p e l l e t generalization gradients (PSEp); those that decrease the incentive value of food should decrease the PSEp. Neither of these treatments should a f f e c t the PSE's derived from generalization gradients of tone st i m u l i (PSE T) i n the same d i r e c t i o n . I t has been demonstrated that drugs a l t e r i n g dopamine transmission a f f e c t discriminations of both stimulus quantities (Church and Meek, 1984) and durations (Maricq and Church, 1983; Spetch and T r e i t , 1984). Agonists increase the perception of number and duration (decrease the PSE) and anatagonists decrease the quantity or time perceived by the animals (increase the PSE). Therefore, i t may be expected that the perception of tone st i m u l i w i l l be affected by dopaminergic drugs i n t h i s fashion, i n the opposite 154 d i r e c t i o n predicted for ef f e c t s on the perception of quantities of food p e l l e t s . The food quantity generalization gradient paradigm has the additional advantage of being able to separate motor from perceptual e f f e c t s . Perceptual influences are r e f l e c t e d i n the PSE, while performance factors are r e f l e c t e d i n the slopes of the generalization gradients. The e f f e c t s of a DA antagonist (haloperidol) and a DA agonist (d-amphetamine) on the generalization gradients of food and tone quantities were investigated. If neuroleptics, such as haloperidol, attenuate the value of food; as suggested by Wise (1981), and dopamine agonists increase the incentive action of food, then these e f f e c t s w i l l be ref l e c t e d i n the PSE F. PROCEDURE The eighteen rats from Experiment 3 were used i n t h i s experiment. The rats were assigned to one of two groups, such that each group was counterbalanced for operant box and for associations between st i m u l i and the "correct" lever. The basic design was as follows. Rats were given one week (5 days) of discrimination t r a i n i n g ( i d e n t i c a l to that given during the 75% reinforcement schedule phase i n Experiment 3), followed by one week (5 sessions on 5 days) of generalization gradient determination (similar to that obtained from Experiment 4); each session was conducted afte r an i n j e c t i o n (i.p.) of vehicle ( d i s t i l l e d H2O) 155 immediately or 20 min pri o r to placement i n the operant box (5 or 250 min pri o r to commencement of t r i a l s ) . The following week, animals were placed on the discrimination t r a i n i n g , followed by 3-5 days of generalization gradient determination sessions, conducted aft e r injections of either haloperidol (McNeil) or d-amphetamine sulphate (Smith Kline and French). One week of discrimination t r a i n i n g sessions followed each drug treatment session. Thus, the series was: trai n i n g , vehicle generalization, t r a i n i n g , drug genera-l i z a t i o n . This series was repeated u n t i l each group had experienced 2 doses of each drug. Blocks of drug treatments were separated by at least 3 weeks. Group 1 was given the drugs i n the following sequence: AMPH (0.5), HAL (0.05), AMPH (0.25), HAL (0.083), where AMPH i s d-amphetamine, injected immediately p r i o r to placement i n the testing box (5 min before the f i r s t t r i a l ) , HAL i s haloperidol, injected 20 min pr i o r to box entry (25 min before the f i r s t t r i a l ) and the numbers i n brackets refer to the dosage i n mg/kg. Group 2 was given drugs i n a d i f f e r e n t sequence: HAL (0.03), AMPH (1.0), HAL (0.083), AMPH (0.25). The doses of haloperidol were chosen so that motor ef f e c t s would be minimal during the f i r s t 3 days of testing, and so that the proportional difference was i d e n t i c a l . Amphetamine doses were chosen to be at the l e v e l of motor ac t i v a t i o n , but to be sub-threshold for anorectic e f f e c t s . When one group received drug pretreated generalization gradient sessions, the other was given vehicle pre-treatments. 156 During each generalization session, 120 t r i a l s were presented (20 F l f 20 F 4, 20 T1, 20 T 4, 10 F 2, 10 F 3, 10 T 2 and 10 T 3 ) , i n the following sequence: Fi_ T^ T]_ Fj_ T 4 F 4 F 4 F l F 2 F 2 F1 T X Ti_ F 4 T 2 Fi_ T 2 T 3 T 4 F 2 T 4 Ti_ F 4 F 3 F 4 T 3 T 4 T 3 F 3 T 4 F 4 T 4 T 2 F 4 T 3 T 3 T 2 F 4 F 3 F 2 F 4 T 4 F x T 4 F 3 Ti_ F 3 T l F2 T2 F l T4 T4 T l F l T l F l T l F4 F l ' repeated once. An i n t e r - t r i a l i n t e r v a l (ITT) of 20 sec separated the end of one t r i a l from the beginning of the next. The f i r s t t r i a l began 5 min after a rat was placed within the operant chamber. Results Response generalization gradients were derived from the data compiled from the f i r s t 60 and second 60 t r i a l s of every session of each treatment; the gradients for the vehicle treatments were obtained from the test sessions p r i o r to the relevant drug treatment. The e f f e c t of vehicle and 3 doses of haloperidol (HAL) on response generalization gradients obtained from the l a s t 60 t r i a l s per session are i l l u s t r a t e d i n Figure 21 (food p e l l e t t r i a l s ) and Figure 22 (tone t r i a l s ) . Three factor analysis of variance was conducted on two measures (point of subjective equality [PSE] and slope) derived from the generalization gradient from each animal, as described i n Experiment 4: Results. The f i r s t (between) factor was group (2 l e v e l s ) ; the second (within) factor was drug dosage (3 levels) and the t h i r d (within factor) was block ( f i r s t 60 t r i a l s X second 60 157 FIGURE 21. E f f e c t of vehicle (0.000) and three doses of haloperidol (HAL) on food p e l l e t quantity generalization gradients. The PSE (the point on the abscissa that intersects the 50% point on the ordinate) was not s i g n i f i c a n t l y altered by any dose of haloperidol. The slope of the gradient was s i g n i f i c a n t l y decreased by the highest dose. 158 i 1 1 r 1 2 3 4 N U M B E R O F T O N E S P R E S E N T E D FIGURE 2 2 . E f f e c t of vehicle (0.000) and three doses of haloperidol (HAL) on tone quantity generalization gradients. The PSE (the point on the abscissa that intersects the 50% point on the ordinate) was s i g n i f i c a n t l y increased i n a dose-dependent manner by haloperidol. The slope of the gradient was s i g n i f i c a n t l y decreased, also i n a dose-dependent manner. 159 t r i a l s ) . As the drug factor consisted of 3 levels for each group (vehicle [VEH], 0.05 mg/kg HAL, and 0.083 mg/kg HAL for one group, and VEH, 0.03 mg/kg HAL and 0.083 mg/kg HAL for the second group), a multivariate test of Wilks Lambda, which does not assume homogeneity of covariances, was used on a l l main ef f e c t s involving the drug factor (Rao's R, d i s t r i b u t e d as the exact F). The following comparisons were planned: VEH x low HAL dose for each group, and for each block ( i e . f i r s t and second halves of sessions), block 1 X block 2 for each group and each drug dose, and group 1 X group 2 for each dose and each block. In cases where a p a r t i c u l a r factor did not contribute s i g n i f i c a n t l y to the main e f f e c t s , comparisons were sometimes made combining across that factor, except for the dose l e v e l i n which the two groups received d i f f e r e n t dosages of HAL ( i e . 0.05 and 0.03 mg/kg). Analysis of the e f f e c t of HAL on the PSE's derived from generalization gradients of responding to food p e l l e t cues (PSE F) revealed that there was no s i g n i f i c a n t main e f f e c t of group (F(l,16) = 2.81, p > 0.1). Nor was a drug e f f e c t apparent (R(2,15) = .3.46, p > 0.05). The group X drug e f f e c t was not s i g n i f i c a n t (R(2,15) = 0.57, p > 0.1). However, PSEp's derived from the second block were s i g n i f i c a n t l y lower than those from the f i r s t block, as revealed by a s i g n i f i c a n t main e f f e c t of blocks (F(l,16) = 5.92, p < 0.05). The group X block, HAL dosage X block and the group X HAL dosage X block i n t e r a c t i o n terms were not s i g n i f i c a n t 160 (F(l,16) = 0.03, p > 0.1; R(2,15) = 1.47, p > 0.1; R(2,15) = 0.14, p > 0.1, resp e c t i v e l y ) . Thus, further comparisons were not conducted. The PSEp's are displayed i n Table 1. Analysis of the slopes (SLOPEp) of the food p e l l e t generalization gradients (Table 2) revealed a s i g n i f i c a n t HAL dose X block e f f e c t (R(2,15) = 6.69, p < 0.01). No other in t e r a c t i o n term was s i g n i f i c a n t (group X HAL dose: R(2,15) = 0.27, p > 0.1; group X block: F(l,16) = 0.68, p > 0.1; group X HAL dose X block: R(2,15) = 0.66 p > 0.1). There was also no in d i c a t i o n of a s i g n i f i c a n t group e f f e c t (F(l,16) = 0.78, p > 0.1). Comparisons between doses within blocks and between blocks within doses were conducted. The lowest dose of HAL (0.03 mg/kg) used did not a f f e c t the SLOPEp, r e l a t i v e to the vehicle gradients on either block (F(l,16) = 1.32, p > 0.1; F(l,16) = 1.99, p > 0.1, respe c t i v e l y ) . The middle dose (0.05 mg/kg) did not a f f e c t the SLOPEp obtained during the f i r s t block (F(l,16) = 1.5, p > 0.1) but s i g n i f i c a n t l y lowered them r e l a t i v e to the SLOPEp of vehicle gradients during the second block (F(l,16) = 5.60, p < 0.05). The highest dose of HAL (0.083 mg/kg) s i g n i f i c a n t l y flattened the generalization gradients obtained from both the f i r s t and second block (F(l,16) = 9.69, p < 0.01; F(l,16) = 16.9, p < 0.002, resp e c t i v e l y ) , but the e f f e c t was stronger during the second block (block 1 X block 2: F(l,16) = 10.9, p < 0.005). The block dependency of the degree of the decrease i n the SLOPEp appears related to drug e f f e c t s ; no differences are apparent i n the SLOPEp 161 T A B L E I. E f f e c t of haloperidol (HAL) and d-amphetamine (AMP) on the points of subjective equality (PSE) for 1 and 4 food p e l l e t s and tones. The PSE was calculated from the resu l t s of the f i r s t and l a s t 60 t r i a l s of the test sessions, and refers to the number of p e l l e t s at which rats w i l l respond equally to two levers, one cued by 1 stimulus, and the other signaled by 4 s t i m u l i . * S i g n i f i c a n t l y d i f f e r e n t from 0.0 dose, p < 0.05. PELLETS TONES HAL DOSE (mg/kg) FIRST LAST FIRST LAST 0.000 2.22 2.20 1.81 1.85 0.030 2.32 2.26 1.90* 1.93* 0.050 2.25 . 2.21 2.15 2.18 0.083 2.17 2.05 2.14* 2.73* AMP DOSE (mg/kg) FIRST LAST FIRST LAST 0.00 2.25 2.21 1.90 1.92 0.25 2.32 2.28 1.79 1.81 0.50 2.40* 2.34* 1.74 1.77 1.00 2.42 2.44* -0.91* 0.48* 162 T A B L E II E f f e c t of haloperidol (HAL) and d-amphetamine (AMP) on the slopes of generalization gradients of d i f f e r e n t quantities of p e l l e t s and tones. The gradients were derived from the results of the f i r s t and l a s t 60 t r i a l s of the test sessions. * S i g n i f i c a n t l y d i f f e r e n t from 0.0 dose, p < 0.05. PELLETS TONES HAL DOSE (mg/kg) FIRST LAST FIRST LAST 0.000 33.3 33.1 24.9 25.7 0.030 32.7 31.8 22.8 23.7 0.050 31.1 * 29.9 23.6* 24.1 0.083 28.6* * 26.5 22.3 20.6* AMP DOSE (mg/kg) FIRST LAST FIRST LAST 0.00 32.8 32.7 25.2 26.4 0.25 33.7 33.2 26.5 26.7 0.50 33.4 34.1 26.3 27.3 1.00 32.4 33.2 13.3* 15.2* 163 obtained from d i f f e r e n t blocks obtained during vehicle pre-treatment sessions (F(l,16) = 2.47, p > 0.1). Differences i n the SLOPESF between the f i r s t and second block aft e r treatment with either the lowest or the middle HAL dose were not s i g n i f i c a n t (F(l,16) = 1.56, p > 0.1; F(l,16) = 3.08, p > 0.05). HAL produced s i g n i f i c a n t e f f e c t s on PSE's derived from tone generalization gradients (PSE T). There was a s i g n i f i c a n t group X HAL dose i n t e r a c t i o n (R(2,15) = 3.72, p < 0.05). Furthermore, there was a s i g n i f i c a n t main e f f e c t observed with the block factor (F(l,16) = 8.00, p < 0.02), ind i c a t i n g that the PSE T was s i g n i f i c a n t l y increased during the l a s t block, regardless of treatment. None of the other in t e r a c t i o n terms were s i g n i f i c a n t (group X block: F(l,16) = 1.99, p > 0.1; dose X block: R(2,15) = 3.23, p > 0.05; group X dose X block: R(2,15) = 1.27, p.> 0.1). The lowest dose of HAL s i g n i f i c a n t l y increased the PSE T over that observed after VEH treatment (F(l,16) = 8.06, p < 0.02). However, the apparent increase aft e r treatment with the middle dose of HAL was not s i g n i f i c a n t (F(l,16) = 3.02, p > 0.05). The highest dose produced a s i g n i f i c a n t increase i n the PSE T (F(l,16) = 11.75, p < 0.005). Analysis of the slopes derived from the tone generalization gradients (SLOPET) revealed a s i g n i f i c a n t group X HAL dose X block i n t e r a c t i o n (R(2,15) = 3.99, p < 0.05). In contrast to the findings with the PSE T, the low dose e f f e c t on the SLOPE T was not s i g n i f i c a n t l y d i f f e r e n t from the SLOPE T observed af t e r vehicle treatment, during either the f i r s t or second block (F(l,16) = 0.75, p > 0.1; F(l,16) = 0.76, p > 0.1, respe c t i v e l y ) , while the middle dose did s i g n i f i c a n t l y reduce the SLOPE T r e l a t i v e to VEH on both blocks (F(l,16) = 4.52, p < 0.05; F(l,16) = 5.49, p < 0.05, re s p e c t i v e l y ) . The highest dose also s i g n i f i c a n t l y depressed the SLOPET, but only during the second block (block 1: F(l,16) = 2.77, p > 0.1; block 2: F(l,16) = 7.04, p < 0.02). The SLOPE T was s i g n i f i c a n t l y higher i n the second block than i n the f i r s t during VEH treatment (F(l,16) = 7.82, p < 0.01), but was not s i g n i f i c a n t l y d i f f e r e n t between the two blocks following any of the HAL treatments (0.03 mg/kg: F(l,16) = 2.44, p > 0.1; 0.05 mg/kg: F(l,16) = 0.65, p > 0.1; 0.083 mg/kg: F(l,16) = 3.05, p > 0.05). The SLOPE T of the two groups did not d i f f e r from each other during any of the HAL treatments (including VEH) or any of the blocks (VEH: block 1: F(l,16) = 1.94, p > 0.1; block 2: F(l,16) = 2.09, p > 0.1; 0.03 X 0.05 mg/kg: block 1: F(l,16) = 0.16, p > 0.1; block 2: F(l,16) = 0.04, p > 0.1; 0.083 mg/kg: block 1: F(l,16) = 0.47, p > 0.1; block 2: F(l,16) = 1.00, p > 0.1) . Figure 23 displays the e f f e c t of HAL on response latencies observed during the l a s t 60 t r i a l s ( a l l sessions combined). The latencies of responses following the presentation of food p e l l e t cues were analyzed separately for each p e l l e t quantity. Latencies to respond aft e r presentation of 1 food p e l l e t (Fi_) were affected by HAL, as 165 FIGURE 23. Ef f e c t of vehicle (0.000) and three doses of haloperidol (HAL) on response latencies on food p e l l e t generalization t r i a l s . Response latencies were increased i n a dose-dependent manner by haloperidol. 166 indicated by a s i g n i f i c a n t i n t e r a c t i o n between HAL dose and blocks (R(2,15) = 8.05, p < 0.005). No other i n t e r a c t i o n term was s i g n i f i c a n t (group X dose: R(2,15) = 0.52, p > 0.1; group X block: F(l,16) = 1.18, p > 0.1; group X dose X block: R(2,15) = 0.53, p > 0.1). The main e f f e c t of group was not s i g n i f i c a n t (F(l,16) = 1.52, p > 0.1). The lowest HAL dose did not a f f e c t the response latencies s i g n i f i c a n t l y on either block (F(l,16) = 0.02, p > 0.1; F(l,16) = 3.45, p > 0.05). The middle dose s i g n i f i c a n t l y increased response latencies on F]_ t r i a l s during the second block of t r i a l s (F(l,16) = 4.97, p < 0.05), but not during the f i r s t (F(l,16) = 2.58, p > 0.1), although the difference i n response latencies during the two blocks of drug treatment was not s i g n i f i c a n t (F(l,16) = 1.72, p > 0.1). The highest dose increased response latencies during both blocks (F(l,16) = 11.19, p < 0.005; F(l,16) = 19.19, p < 0.001, respe c t i v e l y ) , and the latencies were s i g n i f i c a n t l y longer during the second block than during the f i r s t (F(l,16) = 21.2, p = 0.0005). There was no s i g n i f i c a n t difference between the two blocks during vehicle sessions (F(l,16) = 3.69, p > 0.05). Examination of the response latencies after the presentation of the F 2 cue revealed a s i g n i f i c a n t main HAL dose e f f e c t (R(2,15) = 8.53, p < 0.005). Response latencies were s i g n i f i c a n t l y longer during the second block than those from the f i r s t , as indicated by a s i g n i f i c a n t main e f f e c t of blocks (F(l,16) = 8.80, p < 0.01). None of the i n t e r a c t i o n 167 terms were s i g n i f i c a n t (group X dose: R(2,15) = 0.18, p > 0.1; group X block: F(l,16) = 3.02, p > 0.05; dose X block: R(2,15) = 0.29, p > 0.1; group X dose X block: R(2,15) = 0.76, p > 0.1). There was no s i g n i f i c a n t main e f f e c t of group (F(l,16) = 0.35, p > 0.1). The lowest and highest HAL dose s i g n i f i c a n t l y increased response latencies (F(l,16) = 4.40, p < 0.05; F(l,16) = 17.46, p = 0.001, respectively), but the e f f e c t of the middle dose was not s i g n i f i c a n t (F(l,16) = 2.08, p > 0.1) . Analysis of response latencies after the presentation of the F3 cue indicated a s i g n i f i c a n t group X HAL dose X block e f f e c t (R(2,15) = 3.78, p < 0.05). Neither the lowest nor the middle HAL dose s i g n i f i c a n t l y influenced response latencies during either block (VEH x HAL (0.03): block 1: F(l,16) = 1.08, p > 0.1; block 2: F(l,16) = 3.6, p > 0.05; VEH x HAL (0.05): block 1: F(l,16) = 0.42, p > 0.1; block 2: F(l,16) = 0.03; p > 0.1). On the other hand, the highest dose of HAL s i g n i f i c a n t l y increased response latencies i n both groups during both blocks (group 1: block 1: F(l,16) = 9.10, p < 0.01; block 2: F(l,16) = 9.74, p < 0.01; group 2: block 1: F(l,16) = 6.63, p < 0.02; block 2: F(l,16) = 11.5, p < 0.005). The response latencies of group 2 d i f f e r e d from group 1 i n that the latencies of rats i n group 2 during the second block were s i g n i f i c a n t l y longer than those during the f i r s t , regardless of the drug condition (VEH: F(l,16) = 6.18, p < 0.025; HAL (0.03): F(l,16) = 14.4, p < 0.002; HAL (0.083): F(l,16) = 15.3, p < 0.002), while those of the 168 f i r s t group were s i g n i f i c a n t l y d i f f e r e n t only after the high dose of HAL (VEH: F(l,16) = 3.98, p > 0.05; HAL (0.05): F(l,16) = 0.0002, p > 0.1; HAL (0.083): F(l,16) = 5.03, p < 0.05). However, there were no s i g n i f i c a n t differences between the 2 groups after VEH treatments, regardless of the block (block 1: F(l,16) = 1.26, p > 0.1; block 2: F(l,16) = 0.93, p > 0.1). Response latencies following presentation of the F 4 cue were also s i g n i f i c a n t l y altered by HAL treatment (main e f f e c t of dose: R(2,15) = 0.005). Analysis revealed that response latencies during the second block were s i g n i f i c a n t l y longer than during the f i r s t block (main e f f e c t of block: F(l,16) = 37.9, p = 0.0001). However, none of the i n t e r a c t i o n terms were s i g n i f i c a n t (group X dose: R(2,15) = 1.10, p > 0.1; group X block: F(l,16) = 0.69, p > 0.1; dose X block: R(2,15) = 2.49, p > 0.05; group X dose X block: R(2,15) = 0.65, p > 0.1); nor was there a s i g n i f i c a n t main e f f e c t of group (F(l,16) = 1.97, p > 0.1). Neither the lowest nor the middle HAL dose s i g n i f i c a n t l y influenced response latencies r e l a t i v e to vehicle (F(l,16) = 0.90, p > 0.1; F(l,16) = 0.17, p > 0.1, respectively. However, the highest dose s i g n i f i c a n t l y increased response latencies r e l a t i v e to vehicle treatment (F(l,16) = 17.93, p < 0.001). The e f f e c t of HAL on response latencies af t e r presentations of the d i f f e r e n t quantities of tone cues during the l a s t 60 t r i a l s of a l l sessions are presented i n Figure 24. Response latencies during T]_ t r i a l s were 169 15-i UJ Z O QL to cm 12.5-to y 10 H 7.5 5-2.5-HAL (MG/KG) A 0.000 • 0.030 EI 0.050 LI 0.083 A - -A- -A-T 2 3 - A i 4 NUMBER OF TONES PRESENTED FIGURE 24. E f f e c t of vehicle (0.000) and three doses of haloperidol (HAL) on response latencies on tone generalization t r i a l s . Response latencies were increased i n a dose-dependent manner by haloperidol. 170 s i g n i f i c a n t l y affected by HAL (main e f f e c t of dose: R(2,15) = 24.5, p = 0.0001), and response latencies were s i g n i f i c a n t l y increased during the second block, r e l a t i v e to the f i r s t (main e f f e c t of blocks: F(l,16) = 8.06, p < 0.02). None of the int e r a c t i o n terms were s i g n i f i c a n t (group X dose: R(2,15) = 1.62, p > 0.1; group X block: F(l,16) = 0.20, p > 0.1; dose X block: R(2,15) = 1.97, p > 0.1; group X dose X block: R(2,15) = 2.41, p > 0.1), nor was there a s i g n i f i c a n t main e f f e c t of groups (F(l,16) = 0.10, p > 0.1). A l l three HAL doses s i g n i f i c a n t l y increased response latencies, r e l a t i v e to VEH (0.03 mg/kg: F(l,16) = 8.96, p < 0.01); 0.05 mg/kg: F(l,16) = 26.78, p < 0.0005; 0.083 mg/kg: F(l,16) = 22.60, p < 0.0005). Response latencies during T 2 t r i a l s were s i g n i f i c a n t l y affected by the HAL treatments (main e f f e c t of dose: R(2,15) = 8.18, p < 0.005), but were not influenced by the other factors (main e f f e c t of group: F(l,16) = 0.04, p > 0.1; main e f f e c t of block: F(l,16) = 0.08, p > 0.1; group X dose: R(2,15) = 0.21, p > 0.1; group X block: F(l,16) = 0.43, p > 0.1; dose X block: R(2,15) = 3.31, p > 0.05; group X dose X block: R(2,15) = 0.12, p > 0.1). A l l three doses of HAL s i g n i f i c a n t l y increased the response latencies r e l a t i v e to the VEH treatment (0.03 mg/kg: F(l,16) = 7.33, p < 0.02; 0.05 mg/kg: F(l,16) = 5.55, p < 0.05; 0.083 mg/kg: F(l,16) = 7.67, p < 0.tf2). The e f f e c t of HAL on response latencies during T 4 t r i a l s was si m i l a r to that observed during T3 t r i a l s , except that there was a dose X block i n t e r a c t i o n (R(2,15) = 7.63, p < 0.01). No other i n t e r a c t i o n term was s i g n i f i c a n t (group X dose: R(2,15) = 0.50, p > 0.1; group X block: F(l,16) = 0.01, p > 0.1; group X dose X block: R(2,15) = 0.03, p > 0.1), and the main e f f e c t of group was not s i g n i f i c a n t (F(l,16) = 0.45, p > 0.1). A l l three HAL doses significantly-increased response latencies during the second block (0.03 mg/kg: F(l,16) = 8.95, 0.05 mg/kg: F(l,16) = 16.5, p < 0.002; 0.083: F(l,16) = 11.2, p < 0.005), but only the two higher doses s i g n i f i c a n t l y affected latencies during the f i r s t block (0.03 mg/kg: F(l,16) = 1.79, p > 0.1; 0.05 mg/kg: F(l,16) = 6.93, p < 0.02; 0.083 mg/kg: F(l,16) = 5.57, p < 0.05). The response latencies were s i g n i f i c a n t l y longer i n the second block only aft e r the highest dose of HAL (VEH: F(l,16) = 0.21, p > 0.1; 0.03 mg/kg: F(l,16) = 2.07, p > 0.1; 0.05 mg/kg: F(l,16) = 1.26, p > 0.1; 0.083 mg/kg:. F(l,16) = 11.08, p < 0.005). The e f f e c t s of the three doses of HAL on i n t e r - t r i a l i n t e r v a l (ITI) responding are depicted i n Figure 25. There was a s i g n i f i c a n t dose X block i n t e r a c t i o n (R(2,15) = 6.05, p < 0.02). The main e f f e c t of group was not s i g n i f i c a n t (F(l,16) = 0.17, p > 0.1), and the in t e r a c t i o n terms involving the group factor were not s i g n i f i c a n t (group X dose: R(2,15) = 2.43, p > 0.1; group X block: F(l,16) = 0.31, p > 0.1; group X dose X block: R(2,15) = 3.20, p > 0.05). The lowest dose of HAL s i g n i f i c a n t l y decreased ITI responses on both blocks r e l a t i v e to VEH (block 1: F(l,16) = 172 300 n P 250-O 0.000 0.030 0.050 0.083 DOSE OF HALOPERIDOL (MG/KG) FIGURE 25. E f f e c t of vehicle (0.000) and three doses of haloperidol (HAL) on i n t e r - t r i a l i n t e r v a l responses during generalization sessions. Responses were s i g n i f i c a n t l y decreased by the lowest and highest doses of haloperidol^ -p < C\01, 8.8, p < 0.01; block 2: F(l,16) = 12.41, p < 0.005). However, the middle dose did not s i g n i f i c a n t l y decrease ITI responses (block 1: F(l,16) = 0.12, p > 0.1; block 2: F(l,16) = 0.16, p > 0.1), and the highest dose s i g n i f i c a n t l y decreased ITI responses only on the second block (block 1: F(l,16) = 4.05, p > 0.05; block 2: F(l,16) = 8.97, p < 0.05). The e f f e c t s of vehicle (VEH) and three d i f f e r e n t doses of d-amphetamine (AMP) on food p e l l e t generalization gradients (obtained from the l a s t 60 t r i a l s of a l l sessions) are depicted i n Figure 26, and Figure 27 displays the ef f e c t s on tone generalization gradients obtained from the l a s t 60 t r i a l s over a l l sessions. The PSE's are l i s t e d i n Table 1. Analysis of the e f f e c t of three d i f f e r e n t doses of d-amphetamine (AMP) on the PSE F revealed a s i g n i f i c a n t main e f f e c t of dose (R(2,15) = 5.12, p < 0.02). No other factors were s i g n i f i c a n t (main e f f e c t of group: F(l,16) = 1.29, p > 0.1; main e f f e c t of block: F(l,16) = 3.56, p > 0.05). Furthermore, none of the i n t e r a c t i o n terms were s i g n i f i c a n t (group X dose: R(2,15) = 0.306, p > 0.1; group X block: F(l,16) = 0.24, p > 0.1; block X dose: R(2,15) = 1.73, p > 0.1; group X dose X block: R(2,15) = 1.73, p > 0.1). The apparent increase i n the PSE F observed after the lowest AMP dose (0.25 mg/kg), r e l a t i v e to the PSE F observed during VEH treatment, was not s i g n i f i c a n t (F(l,16) = 1.93, p > 0.1). However, the increases produced by the middle dose (0.5 174 FIGURE 26. Ef f e c t of vehicle (0.00) and three doses of d-amphetamine (AMP) on food p e l l e t generalization gradients. The PSE (the point on the abscissa which intersects the 50% point on the ordinate) was increased i n a dose-dependent manner. No s i g n i f i c a n t e f f e c t of d-amphetamine on the slope of the gradients was observed. 175 1 1—: 1 1 1 2 3 4 NUMBER OF TONES PRESENTED FIGURE 27. Ef f e c t of vehicle (0.00) and three doses of d-amphetamine (AMP) on tone generalization gradients. The PSE (the point on the abscissa which intersects the 50% point on the ordinate) and the slope of the gradients were s i g n i f i c a n t l y decreased by the highest dose of AMP. 176 mg/kg: F(l,16) = 6-12, p < 0.025) and by the highest dose (1.0 mg/kg: F(l,16) = 4.74, p < 0.05) were s i g n i f i c a n t . AMP did not have a s i g n i f i c a n t influence on the SLOPEp (main e f f e c t of dose: R(2,15) = 0.87, p > 0.1), as can be observed i n Table 2. In f a c t , none of the factors were associated with s i g n i f i c a n t e f f e c t s (main e f f e c t of group: F(l,16) = 1.17, p > 0.1; main e f f e c t of block: F(l,16) = 0.05, p > 0.1; group X dose: R(2,15) = 0.59, p > 0.1; group X block: F(l,16) = 0.082, p > 0.1; block X dose: R(2,15) = 1.94, p > 0.1; group X block X dose: R(2,15) = 0.62, p > 0.1). Thus, while HAL affected the SLOPEp but not the PSEp, AMP influenced the PSEp but not the SLOPEp. AMP also influenced the PSE T, as can be observed i n Table 1. In t h i s case, AMP decreased the PSE, rather than increasing i t . There was a s i g n i f i c a n t group X dose main e f f e c t (R(2,15) = 7.63, p < 0.01). The main e f f e c t of blocks was not s i g n i f i c a n t (F(l,16) = 1.27, p > 0.1), and none of the i n t e r a c t i o n terms involving t h i s factor were s i g n i f i c a n t (group X block: F(l,16) = 1,18, p > 0.1; dose X block: R(2,15) = 0.50, p > 0.1; group X dose X block: R(2,15) = 0.50, p > 0.1). The lowest dose did not s i g n i f i c a n t l y decrease the PSE T, and the middle dose was also without e f f e c t (F(l,16) = 0.51, p > 0.1). However, the decrease i n the P S E I J J produced by the highest dose was s i g n i f i c a n t (F(l,16) = 39.2, p = 0.0001). While the s t a t i s t i c a l p r o f i l e of the e f f e c t of AMP on the SLOPE T (see Table 2), was si m i l a r to the e f f e c t of AMP 177 on the PSE T, there was a s i g n i f i c a n t increase i n slope from block 1 to 2 (main e f f e c t of blocks: F(l,16) = 5.48, p < 0.05). None of the in t e r a c t i o n terms involving blocks were s i g n i f i c a n t (group X block: F(l,16) = 2.15, p > 0.1; dose X block: R(2,15) = 1.21, p > 0.1; group X dose X block: R(2,15) = 0.94, p > 0.1). The group X dose in t e r a c t i o n was s i g n i f i c a n t (R(2,15) = 11.0, p < 0.005). Neither the lowest nor the middle AMP dose s i g n i f i c a n t l y affected the SLOPE T (F(l,16) = 0.56, p > 0.1; F(l,16) = 0.19, p > 0.1, respe c t i v e l y ) , but the highest dose s i g n i f i c a n t l y depressed i t (F(l,16) = 35.8, p = 0.0001). The e f f e c t of the three doses of AMP on response latencies a f t e r presentations of food p e l l e t cues are d i s -played i n Figure 28. AMP had no s i g n i f i c a n t e f f e c t on the response latencies after presentation of the F]_ cue (main e f f e c t of dose: F(2,15) = 0.23, p > 0.1; group X dose: R(2,15) = 2.4, p > 0.1; dose X block: R(2,15) = 0.48, p > 0.1; group X dose X block: R(2,15) = 1.45, p > 0.1). The main e f f e c t s of group and block were also i n s i g n i f i c a n t (group: F(l,16) = 0.27, p > 0.1; block: F(l,16) = 2.18, p > 0.1; group X block: F(l,16) = 0.05, p > 0.1). There was a s i g n i f i c a n t group X dose X block i n t e r -action on the response latencies during F 2 t r i a l s (R(2,15) = 5.70, p < 0.02). An decrease i n response latencies produced by AMP were consistent, regardless of block or group (VEH X 0.25 mg/kg: block 1: group 1: F(l,16) = 0.09, p > 0.1; group 2: F(l,16) = 0.64, p > 0.1; block 2: group 1: F(l,16) = 178 15 O # 12.5-1 GO L J o y 10 H t< 7.5-j Ld to 5-O Q_ to 2.5H Ld AMP (MG/KG) A CL00 • CL25 M O50 El 1.00 1 T 2 3 4 NUMBER OF PELLETS PRESENTED FIGURE 28. E f f e c t of vehicle (0.00) and three doses of d-amphetamine (AMP) on response latencies during food p e l l e t generalization t r i a l s . The two highest doses s i g n i f i c a n t l y decreased response latencies after intermediate p e l l e t quantities were presented, and the intermediate dose s i g n i f i c a n t l y decreased response latencies to the 4-pellet stimulus. The lowest dose had no s i g n i f i c a n t e f f e c t on latencies. 1.17, p > 0.1; group 2: F(l,16) = 1.14, p > 0.1; VEH "X 0.5 mg/kg: block 1: F(l,16) = 5.50, p < 0.05; block 2: F(l,16) = 5.32, p < 0.05; VEH X 1.0 mg/kg: block 1: F(l,16) = 12.52, p < 0.005; block 2: F(l,16) = 25.3, p < 0.0005). There was a peculiar difference between the two groups: while there were no s i g n i f i c a n t differences between the f i r s t and second blocks regardless of treatment i n group 1 (VEH: F(l,16) = 0.04, p > 0.1; 0.25 mg/kg: F(l,16) = 1.32, p > 0.1; 0.5 mg/kg: F(l,16) = 0.0005, p > 0.1), group 2 exhibited s i g n i f i c a n t l y longer latencies during the second block after VEH (F(l,16) = 18.05, p < 0.001) or 0.25 mg/kg AMP (F(l,16) = 12.27, p < 0.005), but not after 1.0 mg/kg (F(l,16) = 3.83, p > 0.05) . There was a s i g n i f i c a n t increase i n response latencies from block 1 to block 2 during F3 t r i a l s (main e f f e c t of block: F(l,16) = 12.91, p < 0.005), as well as a s i g n i f i c a n t decrease produced by AMP (main e f f e c t of dose: R(2,15) = 5.25, p < 0.02). There was no s i g n i f i c a n t main e f f e c t of group (F(l,16) = 1.33, p > 0.1), and no s i g n i f i c a n t interactions (group X dose: R(2,15) = 0.59, p > 0.1; group X block: F(l,16) = 1.69, p > 0.1; dose X block: R(2,15) = 2.18, p > 0.1; group X dose X block: R(2,15) = 0.12, p > 0.1). Neither the lowest nor the middle dose of AMP s i g n i f i c a n t l y decreased response latencies r e l a t i v e to VEH (F(l,16) = 1.27, p > 0.1; F(l,16) = 2.89, p > 0.1). However, the highest dose s i g n i f i c a n t l y decreased response latencies (F(l,16) = 8.16, p < 0.02). 180 A s i g n i f i c a n t i n t e r a c t i o n between AMP dose and block was observed i n response latencies during F4 t r i a l s (R(2,15) = 7.30, p < 0.01). The main e f f e c t group was not s i g n i f i c a n t (F(l,16) = 0.02, p > 0.1), and none of the in t e r a c t i o n terms involving the group term were s i g n i f i c a n t (group X dose: R(2,15) = 0.51, p > 0.1; group X block: F(l,16) = 0.07, p > 0.1; group X dose X block: R(2,15) = 1.14, p > O.i). The lowest dose of AMP was i n e f f e c t i v e at a l t e r i n g response latencies (block 1: F(l,16) = 1.17, p > 0.1; block 2: F(l,16) = 0.76, p > 0.1). The middle dose significantly-decreased response latencies (block 1: F(l,16) = 6.76, p < 0.02; block 2: F(l,16) = 8.43, p < 0.02), but the decrease observed af t e r treatment with the highest dose was not s i g n i f i c a n t (block 1: F(l,16) = 1.26, p > 0.1; block 2: F(l,16) = 3.58, p > 0.05). Response latencies were s i g n i f i c a n t l y longer during the second block than during the f i r s t 60 t r i a l s a f t e r VEH (F(l,16) = 59.9, p = 0.0001), 0.25 mg/kg AMP (F(l,16) = 23.3, p < 0.0005), and 0.50 mg/kg AMP (F(l,16) = 5.46, p < 0.05), but not after 1.0 mg/kg AMP (F(l,16) = 1.68, p > 0.1). The lack of e f f e c t of AMP on latencies to respond to tone cues are presented i n Figure 29. As analysis of variance revealed no s i g n i f i c a n t e f f e c t s on the response latencies of any of the tone cues, the results w i l l not be discussed further. AMP also had no s i g n i f i c a n t actions on ITI responses, as can be observed i n Figure 30. The e f f e c t of the highest dose of AMP on the response bias index was 181 15-, 12.5-^ 10-7.5-AMP (MG/KG) A 0.00 • 0.25 EI 0.50 1.00 2.5--A 1" 2 T~ 3 1 T 4 NUMBER OF TONES PRESENTED FIGURE 29. Eff e c t of vehicle (0.00) and three doses of d-amphetamine (AMP) on response latencies during tone generalization t r i a l s . No s i g n i f i c a n t e f f e c t on tone cue response latencies were observed at any AMP dose. 182 € ( f^i € i I E £ C j ' f f i £ / ,r /• / - J t ^ ' | ^ 0.00 0.25 0.50 1.00 DOSE OF AMPHETAMINE (MG/KG) FIGURE 30. Ef f e c t of vehicle (0.00) and three doses of d-amphetamine (AMP) on i n t e r - t r i a l i n t e r v a l (ITI) responses latencies between generalization t r i a l s . The intermediate dose (0.50 mg/kg) apparently increased ITI responses; t h i s e f f e c t was not s i g n i f i c a n t . 183 analyzed (VEH: 0.32; AMP (1.0 mg/kg): 0.28). No s i g n i f i c a n t e f f e c t of AMP on response bias was observed (F(l,8) = 0.24, p > 0.1). Furthermore, while AMP e f f e c t s of PSEp indicated that rats responded to a given quantity of food p e l l e t s with a bias towards the 4-pellet cued lever, the response bias during the ITI was towards the 1-pellet cued lever. Discussion The two major observations i n t h i s experiment were that haloperidol was i n e f f e c t i v e at a l t e r i n g the PSE derived from food p e l l e t generalization gradients while d-amphetamine produced an increase i n the PSE F. The lack of e f f e c t of haloperidol could not be attributed to the fact that low doses were employed. Haloperidol was e f f e c t i v e at decreasing performance, as both the 0.05 (mg/kg) and the 0.083 doses induced some degree of reduction i n the SLOPEp. Response latencies to cues of both types were increased by haloperidol i n a dose-dependent manner. In addition, responding during the i n t e r - t r i a l i n t e r v a l was attenuated by haloperidol, even when the lowest dose (0.03 mg/kg) was administered. Therefore, the doses of haloperidol were s u f f i c i e n t to produce "motor" e f f e c t s . This i s an important point, as Wise (1982) has suggested that the e f f e c t s of neuroleptics on reinforcement processes should occur before (i e . at lower doses) motor e f f e c t s . Furthermore, the PSE T was s i g n i f i c a n t l y increased by haloperidol, an i n t e r e s t i n g extension of a s i m i l a r finding 184 by Maricq and Church (1983) i n regards to the e f f e c t of haloperidol at decreasing time estimation. A similar finding with a d i f f e r e n t behavioral procedure was reported e a r l i e r (Szostak and Tombaugh, 1981) for pimozide. That haloperidol also decreases the perception of number, as r e f l e c t e d by an increase i n the PSE T, supports Church and Meck's (1984) suggestion that counting and timing are based on an i d e n t i c a l mechanism. As the doses of haloperidol were e f f e c t i v e at influencing perceptions of tone number, i t appears that the lack of e f f e c t on food p e l l e t quantity i s not due to i n s u f f i c i e n t dosage. Amphetamine was observed to decrease the PSE T, consistent with previous observations that methamphetamine (Maricq, Roberts and Church, 1981; Maricq and Church, 1983) and d-amphetamine (Spetch and T r e i t , 1984) increases the perception of duration. Methamphetamine has also been reported to increase the perception of quantity (Church and Meek, 1983). Thus, the present experiment indicates that a DA agonist and a DA antagonist have opposite e f f e c t s on the perception of quantity of a neutral stimulus, providing strong evidence that DA-releasing neurons are involved i n the numerical attributes of s t i m u l i . Amphetamine affected the perception of food p e l l e t quantities i n the predicted d i r e c t i o n , apparently increasing the value of food. This e f f e c t was not associated with any performance impairment, as r e f l e c t e d by the slopes of the food p e l l e t generalization gradients. The highest two doses 185 also decreased the response latencies to some of the food cues. No s i g n i f i c a n t e f f e c t on response latencies to tones were observed. The performance of tone discriminations was impaired to some degree by amphetamine, as evidenced by a s i g n i f i c a n t decrease i n slope of the tone generalization gradient. However, performance of tone discriminations was also impaired by haloperidol, which produced an opposite e f f e c t on the PSE T. Thus, the changes i n PSE cannot be attributed to performance impairments. Furthermore, amphetamine did not produce a s i g n i f i c a n t a l t e r a t i o n i n response bias, as measured by d i f f e r e n t i a l responding to the two levers during i n t e r - t r i a l i n t e r v a l s . Indeed, there was a s l i g h t trend to increase responses to the 1-pellet cued lever, the opposite lever to which animals increased responses to during t r i a l s . Changes i n the PSE's are therefore u n l i k e l y to be related to a non-specific e f f e c t of amphetamine on increasing responding to a p a r t i c u l a r lever. I t i s of i n t e r e s t that tone discriminations were more vulnerable to disrupting e f f e c t s of both drugs. Clody and Carlton (1980) proposed that neuroleptics reduce stimulus e f f i c a c y on the basis of evidence that behavior proximal to reinforcement or maintained by large magnitudes of reward were less s e n s i t i v e to disruption by neuroleptics than behaviors less proximal or maintained with small reward magnitudes. If t h i s e f f e c t i s mediated by DA systems, then i t would be expected that amphetamine would increase stimulus e f f i c a c y . However, both drugs disrupted the least 186 e f f e c t i v e s t i m u l i (tones) to a greater extent than the more e f f e c t i v e s t i m u l i (food). Therefore, i t can be concluded that stimulus e f f i c a c y i s not s e l e c t i v e l y impaired by neuroleptics. Rather, i t appears that behavior sustained by more e f f e c t i v e stimuli i s simply more res i s t a n t to disruption by any treatment. The observation that amphetamine increased the hedonic value of food while haloperidol did not a f f e c t i t suggests that the action of amphetamine on the perceived value of food i s not related to an increase i n DA at receptor s i t e s . Amphetamine also i s a potent i n d i r e c t noradrenaline agonist. To what extent e f f e c t s of amphetamine on neurotransmitters other than DA contribute to i t s apparent reward enhancing actions are not presently known.• Alt e r n a t i v e l y , i t i s possible that the release of DA augments the hedonic value of food, but that DA release i s not normally concomitant with food consumption. Haloperidol (a DA receptor antagonist) would not have an e f f e c t as DA receptors are not occupied. This explanation appears u n l i k e l y i n the l i g h t of a report by Heffner, Hartman and Seiden (1980) that feeding by hungry rats i s associated with an increase i n mesolimbic DA metabolism. The p o s s i b i l i t y that haloperidol does not antagonize a l l types of DA receptors must also be considered. Some evidence for t h i s has been presented, but i s controversial at present (Seeman, 1980). Whatever the mechanism of action of amphetamine's e f f e c t at increasing the hedonic value of food, i t i s clear 187 from the present observations that the response suppressant e f f e c t s of haloperidol i n operant situations cannot be attributed to a decrease i n the hedonic value of food. 188 GENERAL DISCUSSION Rationale The evidence reviewed i n the Introduction indicated that the role of DA transmission i n mediating the hedonic properties of events i s not c l e a r . A p a r t i c u l a r problem has been that the vast majority of behavioral procedures used i n the investigation of the physiology of rewards have not been e f f e c t i v e at d i f f e r e n t i a t i n g between the sensory attributes of rewards and t h e i r behavioral consequences. The intent of the present series of experiments was to develop a procedure that would measure the hedonic value of a reward independently of i t s response consequences, and to use t h i s procedure to determine whether DA systems are involved i n the evaluation of the sensory attributes of rewards. The f i e l d of psychophysics has developed a number of procedures designed to d i f f e r e n t i a t e perceptual processes from other factors contributing to performance. Church and Meek (1984) have found that animals can discriminate s t i m u l i on the basis of quantity. Furthermore, quantity generalization gradients can be obtained from rats trained to discriminate between d i f f e r e n t numbers of s t i m u l i . The advantage of such gradients i s that a measure (the point of subjective equality - the PSE) can be derived from them that r e f l e c t s the perception of quantity with l i t t l e confounding by performance variables. In addition, the presence or absence of performance impairments may be i n f e r r e d from the slope of the gradients. 189 It appeared reasonable that rats may be trained to discriminate between two d i f f e r e n t reward magnitudes. Generalization gradients obtained by introducing "probes" of intermediate reward magnitudes could provide a PSE that r e f l e c t s rats' perceptions of the hedonic value of food. The f i r s t ten experiments i n the present series were designed to assess whether rats can discriminate between two d i f f e r e n t reward magnitudes, and whether the PSE obtained from reward magnitude generalization gradients r e f l e c t the hedonic value of rewards. The dimension of reward magnitude chosen for the present experiments was quantity, primarily for methodological reasons. Experiments were conducted to assess the r e l a t i v e hedonic value of st i m u l i to be used i n other experiments, to v e r i f y that they d i f f e r i n hedonic value. Additional experiments examined the e f f e c t of motivational manipulations on r e l a t i v e hedonic value of d i f f e r e n t quantities or flavors of food p e l l e t s . The primary experiments began with the t r a i n i n g of animals to discriminate between d i f f e r e n t reward magnitudes (quantity), and between d i f f e r e n t numbers of a hedonically neutral stimulus, as a control. This was undertaken to enable the determination of quantity generalization gradients that might r e f l e c t rats' perceptions of hedonic value of rewards. Included i n these experiments were two that assessed whether or not the generalization gradients would r e f l e c t a l t e r a t i o n s i n hedonic value produced by two independent 190 methods. These experiments were necessary to determine the v a l i d i t y of the generalization gradient procedure as a measure of hedonic value of rewards. Further experiments were conducted to determine whether the control stimuli used were hedonically neutral, whether the effects of reward "devaluation" by changing the stimulus attributes of food could be explained by stimulus novelty, and whether rats may have discriminated reward quantity with cues provided by the reward d e l i v e r y system, rather than those provided by the reward i t s e l f . The l a s t experiment examined the e f f e c t s of a DA antagonist and an agonist on generalization gradients for reward quantity and for numbers of a neutral stimulus. Summary of Experiments  Experiment 1 Experiment 1 established that two d i f f e r e n t quantities of sweetened food p e l l e t s (4 and 1) d i f f e r e d i n hedonic value, with 4 p e l l e t s preferred over 1. The use of a free-operant choice procedure i s complicated by possible response biases. Therefore, the ef f e c t s of a contingency reversal were investigated. Rats r e l i a b l y and quickly tracked the reversal of reinforcement contingencies, i n d i c a t i n g that the preference for a greater quantity of food p e l l e t s i s strong and r e l i a b l e , exhibited by a l l rats tested. Experiment 2 The p o s s i b i l i t y that al t e r a t i o n s i n motivation may a l t e r the r e l a t i v e hedonic values of d i f f e r e n t quantities of 191 food p e l l e t s was investigated i n Experiment 2. Decreasing hunger by either pre-loading animals with food p r i o r to testing, or maintaining the rats on ad libitum access to food, had no e f f e c t on preferences. While t h i s r e s u l t appears to support the view that 'drive' does not interact with the hedonic values of food quantity, the preponderance of the l i t e r a t u r e suggests that i t does (vide Mackintosh, 1974; Beck, 1978). Furthermore, there i s an extensive l i t e r a t u r e establishing that l e v e l of hunger influences the hedonic value of sweet food (Young and Greene, 1953; Smith and Capretta, 1956; Smith and Duffy, 1957a, 1957b; Capretta, 1962). F i n a l l y , Young (1977) has provided evidence that changes i n hedonic value sometimes have l i t t l e control over behavior, once that behavior has become 'habitual'. Thus, i t appears that the lack of e f f e c t of decreasing hunger on food quantity preferences observed i n Experiment 2 may not r e f l e c t the absence of change i n hedonic value. Experiment 3 Experiment 3 established that rats can learn to discriminate between d i f f e r e n t quantities of food p e l l e t s . Furthermore, the discrimination of p e l l e t quantities was found to be: (1) q u a n t i t a t i v e l y d i f f e r e n t from the discrimination of quantities of a presumed neutral stimulus (tones) i n terms of a c q u i s i t i o n rate and l e v e l of asymptotic accuracy, and (2) q u a l i t a t i v e l y d i f f e r e n t i n r e l a t i o n to response patterns (response l a t e n c i e s ) . 192 Experiment 4 If rats attend to the hedonic value of food when discriminating between food quantities, then changing the l e v e l of hunger would s e l e c t i v e l y a l t e r t h e i r perceptions of food quantity. To determine i f t h i s i s so, generalization gradients of stimulus quantities for both food and tone quantities were obtained by presenting rats discriminating between 1 and 4 food and tone cues with probe t r i a l s consisting of intermediate quantities of cues (Experiment 4). Sessions including intermediate stimulus quantity probes were conducted while levels of food deprivation were altered. Interpretation of Generalization Gradients. Quantity generalization gradients are useful because two r e l a t i v e l y orthogonal measures may be derived from them, one representing perception and the second r e l a t i n g to performance. The f i r s t measure i s the point of subjective equality (PSE), which i s the number of stimuli judged to be as s i m i l a r to a 1-stimulus cue as i t i s to a cue composed of 4 s t i m u l i ( i n the present case). A downward s h i f t i n the PSE r e f l e c t s that fewer stimuli are required for rats to s h i f t responding from the lever cued by 1 stimulus to the 4 stimulus lever; a given quantity i s perceived as being greater i n number than the standard. An upward s h i f t i n the PSE r e s u l t s from the converse: a given quantity i s perceived as less than the standard (more st i m u l i are required before 193 responding s h i f t s from the 1-stimulus to the 4-stimuli l e v e r ) . The second measure derived from a generalization gradient i s the slope of the function. This i s primarily a performance measure. As performance (accuracy) of the discrimination decreases, so does the slope. Theoretically, alterations i n perception s h i f t the generalization gradient to the ri g h t or l e f t , with the PSE value s h i f t i n g respectively down or up, with no change i n slope. A reduction i n performance, but not i n perception, produces a fl a t t e n i n g of the generalization gradient, r e f l e c t e d by a decrease i n the slope. In the present case of food p e l l e t generalization, the th e o r e t i c a l s i t u a t i o n i s not completely applicable. The one and four p e l l e t cues are associated with near maximal accuracy; s h i f t s of the gradient to the ri g h t would f l a t t e n the curve between the one and two p e l l e t cues, while s h i f t s to the l e f t would f l a t t e n the curve between the 3 and 4 p e l l e t cues. This would reduce the slope of the l i n e to some extent, but these cases are e a s i l y d i s c e r n i b l e because the f l a t t e n i n g occurs only at one of the extremes of the curve. Another p o t e n t i a l complication may occur i n the case of a response a r t i f a c t . If responding to one lever i s d i f f e r e n t i a l l y affected by some treatment, then a s h i f t i n the PSE could r e f l e c t t h i s confound and be misinterpreted as a change i n perception. As the response requirements of each cued t r i a l were equivalent, and as no major systematic 194 response bias was exhibited during i n t e r - t r i a l i n t e r v a l s i n Experiment 3, t h i s does not appear a l i k e l y p o s s i b i l i t y . Furthermore, the response bias was monitored during a l l treatment conditions i n subsequent experiments. In no case was a change i n response bias observed that could account for any of the r e s u l t s . I n i t i a l Predictions. Generalization gradients were obtained during three d i f f e r e n t periods i n which rats were maintained at one of 3 d i f f e r e n t food deprivation conditions: body weights at 75%, 85% or 95% of body weights during ad libitum access to food. I t was predicted that enhancing food.value by increasing deprivation would be r e f l e c t e d by a given quantity of food being perceived as more similar to a larger food quantity with higher hedonic value. It was further predicted, on the basis of evidence suggesting that incentive value was a p o s i t i v e l y accelerated function of reward quantity (eg. Leventhal et a l . , 1959), that the food generalization gradients would take the form of a p o s i t i v e l y accelerated function. Obtained Forms of Generalization Gradients. Neither of the two predictions were verified.. The forms of both food and tone quantity generalization gradients were negatively accelerated functions, s i m i l a r to the forms taken by generalization gradients of animals trained to discriminate between d i f f e r e n t numbers or durations of neutral stimuli (Church and Meek, 1983). Closer examination of the evidence supporting the accepted view that incentive i s a p o s i t i v e l y accelerated function of reward magnitude (Mackintosh, 1974) revealed that t h i s appears to be true only when reward magnitudes are varied within small magnitudes (zero and two food p e l l e t s , for example). Thus, the difference between zero and one food p e l l e t s was found to be smaller than the difference between one and two p e l l e t s (Leventhal et a l . , 1959). However, increasing the size of the units of reward magnitude produces re s u l t s i n d i c a t i n g that the difference between zero and an intermediate unit of reward i s greater than the difference between the intermediate unit and the large unit (Leventhal et a l . , 1959; Pubols, 1962). Therefore, the r e l a t i o n between hedonic value and reward magnitude appears to be a sigmoid function ( p o s i t i v e l y accelerated at small values and negatively accelerated at larger values). That the relationship between value and reward magnitude i s a sigmoid function i s consistent with the food generalization gradients obtained i n Experiment 4. In t h i s case, a negative acceleration was less apparent between 1 and 2 food p e l l e t s than between greater numbers of p e l l e t s . This was i n contrast to the tone quantity generalization gradients, which exhibited clear negative acceleration throughout the range of tones tested. Differences i n the absolute perceptions of food quantities r e l a t i v e to tone quantities were observed. The existence of these differences argue against the "amodal" 196 nature of numerical attributes of s t i m u l i , suggested by some investigators (Roberts, 1983; Church and Meek, 1984; Holder and Roberts, 1985). However, i t i s possible that differences i n perceptions of the quantity of d i f f e r e n t s t i m u l i are not related to modality differences, but to differences i n inherent b i o l o g i c a l s i g n i f i c a n c e . Effects of Food Deprivation on Quantity Perceptions. A l t e r i n g the l e v e l of food deprivation of rats altered the perception of food quantity, as assessed by the point of subjective equality (PSE). No such e f f e c t of deprivation was observed on the perception of tone quantities. These observations established that the perception of food quantity by rats i s dependent, at least to some degree, on the value of that food. It has been proposed that rats' perceptions of stimulus duration, which some investigators believe to be based on a mechanism i n common with perceptions of stimulus quantity (eg. Church and Meek, 1984), i s correlated with the signal value of the stimulus. Insofar as a l t e r i n g food deprivation changes the signal value of tones, the present r e s u l t suggests that quantity discriminations do not r e f l e c t signal value of inherently neutral s t i m u l i . Contrary to expectations, a given quantity of food appeared more similar to food of lower value and less similar to food of greater value with increasing food deprivation. This r e s u l t suggested that s p e c i f i c quantities of food appear to be assigned value i n r e l a t i o n to the 197 amount required for s a t i a t i o n . Increasing deprivation increases the amount required for s a t i a t i o n , and therefore a given quantity of food i s perceived as smaller. However, i f the rats' perceptions of food quantities were based exclusively on t h i s relationship, then the generalization gradients would have taken a l i n e a r form: the difference between the proportion of one p e l l e t ( r e l a t i v e to the s a t i a t i n g quantity) and the proportion of two p e l l e t s (in r e l a t i o n to the s a t i a t i n g quantity) would be the same as the difference between the proportions of two and three p e l l e t s ( r e l a t i v e to the s a t i a t i n g quantity). For example, 1/100 -2/100 = 3/100 - 4/100. However, t h i s l i n e a r r e l a t i o n s h i p was not observed: the differences between two quantities of food p e l l e t s lessened as the numbers of p e l l e t s increased. Therefore, i t appears that while the perception of absolute quantity of food i s dependent upon the motivational state of the rats, the discrimination between two quantities of food i s based on proportional differences i n number or taste. Model of Food Quantity Perceptions. The above reasoning explains both the forms of the food generalization gradients and the e f f e c t s of food deprivation on those gradients. However, such a post hoc account i s not completely convincing and therefore further experimentation was warranted to e x p l i c i t l y test predictions of t h i s hypothesis. In summary, the postulates of the present thesis are: 198 1) Rats discriminate between d i f f e r e n t quantities of food p e l l e t s on the basis of proportional differences i n number or taste i n t e n s i t y . 2) Quantity of a re i n f o r c i n g stimulus i s perceived i n r e l a t i o n to the t o t a l amount required to satiate the animal. "Satiation" refers to the maximum exposure to a stimulus an animal i n a given state w i l l choose i n the r e l a t i v e absence of environmental constraints; i t does not refer to drive reduction, but rather to stimulus s a t i a t i o n . 3) The amount of a stimulus required to satiate an animal i s a d i r e c t r e f l e c t i o n of the value the animal assigns to that stimulus. 4) Value i s determined on the basis of stimulus c h a r a c t e r i s t i c s (such as taste) and b i o l o g i c a l need (such as produced by food deprivation). In regards to the l a t t e r point, i t should be emphasized that factors such as pr i o r experience and post-ingestive e f f e c t s are not excluded, but are not relevant to the present s i t u a t i o n . This hypothesis i s c l o s e l y related to the concepts of reinforcement proposed by Hanson and Timberlake (1983), i n which behavior-regulation systems operate to diminish deviations i n behavioral set-points produced by environmental constraints. Within t h i s framework, i t i s proposed that rats perceive food quantity i n r e l a t i o n to the i r "set-point" for consumption of p a r t i c u l a r foods. S p e c i f i c predictions can be derived from t h i s account for the case when animals trained to discriminate between two quantities of food p e l l e t s are presented with probes consisting of equivalent quantities of p e l l e t s with lower hedonic value. The quantity of a s p e c i f i c food i s perceived i n r e l a t i o n to the set-point for that food. The set-point of less-valued food i s lower than the set-point of preferred food. Therefore, a given quantity of food would be perceived as a greater amount than an equal quantity of food with higher hedonic value. The o r i g i n a l hypothesis that perceived food quantity i s p o s i t i v e l y correlated with food value would predict the opposite: a given quantity of non-preferred food would be perceived as less than an equal quantity of preferred food. A hypothesis that the hedonic value of food i s not related to the motivational e f f e c t s observed i n Experiment 4 would predict that a given quantity of less valued food would be perceived as being equivalent to an equal amount of preferred food. Experiment 5 Experiment 5 investigated whether unsweetened (U) food p e l l e t s had less hedonic value than the regular sweetened food p e l l e t s (S). A free-operant two-lever choice procedure with contingency reversal tests was employed. Rats were i n i t i a l l y found to prefer S over U, when presented i n i d e n t i c a l quantities, and the contingency reversal was tracked. Therefore, t h i s food appeared suitable for use i n testing the predictions of the thesis presented above. 200 Experiment 6 Before t e s t i n g the prediction that non-preferred food would be perceived to occur i n greater quantities than preferred food, an additional experiment was conducted. The observation i n Experiment 5 that equal quantities of S and U d i f f e r i n hedonic value provided an opportunity to remedy a possible confound i n the previous experiment (Experiment 2) on the role of motivation i n preferences. In Experiment 6, the quantity of S presented per lever-press was systematically decreased u n t i l rats exhibited equivalent preferences for S and U. Any lack of e f f e c t of changes i n hunger on r e l a t i v e responding for the two kinds of p e l l e t s could not be attributed to habitual responding for one type of p e l l e t . I t was found that increasing food deprivation s h i f t e d preferences s l i g h t l y but s i g n i f i c a n t l y for the U p e l l e t s , available i n greater quantities than the S p e l l e t s . While t h i s may seem i n t u i t i v e l y predictable, Valenstein (1967) found that i n rats given a choice between a c a l o r i e -containing sucrose solution and a c a l o r i e - f r e e saccharin solution, preferences for the saccharin solution increased with food deprivation, even to the point where some rats died of starvation. The r e s u l t s of Experiment 6 established that motivational factors can a l t e r r e l a t i v e preferences between foods d i f f e r i n g i n both quantity and q u a l i t y (taste). Furthermore, the value of quantity appears to be affected to a larger degree by increasing food deprivation than the 201 value of taste. That deprivation l e v e l can influence hedonic value of food quantities provides further support that the deprivation manipulations i n Experiment 4 altered the hedonic value of the food cues. Experiment 7 The previous two experiments established that U p e l l e t s were suitable to test the prediction that quantities of non-preferred p e l l e t s would be perceived to be greater than equal amounts of preferred p e l l e t s . In Experiment 7, rats were given generalization probes of U p e l l e t s , interspersed among S p e l l e t probes. Two important features of the U p e l l e t generalization gradients are of i n t e r e s t . F i r s t l y , the prediction that a given quantity of U would be perceived as greater than an equivalent quantity of S was v e r i f i e d . Secondly, the transfer of the discrimination from S to U was not complete. It was evident that the accuracy of responding to the U stimuli decreased with prolonged testing, l i k e l y due to learning that the U p e l l e t s were not correlated with reinforcement. However, the difference i n the perception of quantity was apparent during the f i r s t two sessions, and did not become greater with repeated t e s t i n g . The slope of the U generalization gradient was also lower than that of the S gradient on the i n i t i a l t r i a l s . This finding indicates that there was some property of S not shared by U that contributed to the discrimination. Since the p e l l e t s d i f f e r e d only i n taste, i t appears that taste of S p e l l e t s 202 was involved to some degree i n the quantity discrimination. However, rats did perform the 1 versus 4 U discrimination at acceptable le v e l s of accuracy (71 and 85% "correct"). Therefore, i t can be concluded that proportional differences i n number were the major cues attended to by rats when discriminating food quantities. Experiment 8 I t appeared possible that the novelty of U p e l l e t s may have been a determining factor i n the f a i l u r e to transfer the S p e l l e t discrimination completely to the U probes during the i n i t i a l sessions; In f a c t , t h i s difference between S and U p e l l e t s may have produced the PSE s h i f t , a l b e i t for unknown reasons. Therefore, an experiment equivalent to Experiment 7 was conducted i n which generalization probes of quantities of novel tones were introduced, interspersed among generalization probes of the "regular" tones (Experiment 8). Transfer to the novel tones was complete, throughout the t e s t i n g sessions, and no s h i f t i n the PSE was observed. Thus, the difference i n quantity perception between S and U p e l l e t s l i k e l y cannot be attributed to differences i n f a m i l i a r i t y with the food p e l l e t s . Experiment 9 A central assumption of the previous experiments i s that the tones are d i f f e r e n t from the food p e l l e t s i n that the former are hedonically neutral s t i m u l i , while food p e l l e t s are hedonically p o s i t i v e . I f , i n fact, tones have hedonic value then the use of tones as a control i s questionable. Experiment 9 investigated the hedonic value of the tones i n rats trained i n the discrimination task. The a b i l i t y of response-contingent tones to maintain lever-pressing i n the absence of any other reinforcer was determined. Rats "rewarded" for lever-pressing by the presentation of tones did not exhibit s i g n i f i c a n t l y higher rates of responding than rats not receiving reward. Thus, the assumption that tones are hedonically neutral was supported. Experiment 10 It i s t h e o r e t i c a l l y possible that rats discriminated between d i f f e r e n t quantities of food p e l l e t s by attending to the number of c l i c k s produced by the p e l l e t dispenser. Considering the difference i n performance of p e l l e t and tone discriminations, and the difference between food and tone generalization gradients following the introduction of novel st i m u l i d i f f e r i n g i n i n t e n s i t y , t h i s p o s s i b i l i t y appears unl i k e l y . Nevertheless, Experiment 10 tested t h i s hypothesis by observing the ef f e c t s of providing dispenser c l i c k cues during e x t i n c t i o n . If dispenser c l i c k s provided discrimin-ative cues during the previous experiments, then some degree of discriminative control should have been apparent during the i n i t i a l e x t i n c t i o n t r i a l s . This was not found to be the case. Indeed, during the i n i t i a l t r i a l s , rats p r e f e r e n t i a l l y responded to the 1-pellet cued lever, regardless of the number of dispenser c l i c k s . This indicates that rats 204 perceived the absence of. food to be more sim i l a r to 1 food p e l l e t than to 4 food p e l l e t s . Experiment 11 The f i r s t 10 experiments established that the food quantity generalization procedure provides a method of assessing rats' perceptions of the hedonic value of food. Experiment 11 investigated the e f f e c t s of haloperidol and amphetamine on food and tone generalization gradients. The major observations of t h i s experiment were that while haloperidol did not a l t e r rats' perceptions of food quantity, amphetamine increased them. Furthermore, amphetamine decreased the number of tones perceived by rats, while haloperidol increased tone quantity perceptions. Haloperidol produced additional behavioral e f f e c t s , increasing response latencies and decreasing i n t e r - t r i a l i n t e r v a l responses. Therefore, the lack of haloperidol action on the perception of food quantity cannot be attributed to the use of behaviorally i n e f f e c t i v e dosages. It may be argued that haloperidol did not appear to have an e f f e c t on the perception of food quantities because i t s action on quantity perceptions i s opposite to i t s e f f e c t on reward perception. That i s , the two ef f e c t s cancel each other. Amphetamine did have opposite actions, depending on the hedonic value of the s t i m u l i . However, i t s e f f e c t s on rats' perceptions of food quantity i s apparent to some degree even at the lowest dose. The question of why the eff e c t s would counter each other i n the case of haloperidol, 205 but not i n the case of amphetamine poses serious d i f f i c u l t i e s for t h i s hypothesis. Conclusions The major observations of the present series of experiments are: 1) Rats' perceive a given quantity of food p e l l e t s as less than an equivalent quantity of tones. 2) Two independent methods of "devaluating" a reward produce si m i l a r s h i f t s i n perceptions of reward quantity. Decreases i n either l e v e l of food deprivation or food sweetness increase the perceived quantity of a given number of food p e l l e t s . 3) A l t e r i n g food deprivation or decreasing tone i n t e n s i t y have no influence on the numerical attributes of tones. 4) Haloperidol does not a l t e r rats' perceptions of food quantity, but decreases the perceived quantity of tones. 5) Amphetamine decreases the perceived quantity of food p e l l e t s , and increases the perceived quantity of tones. The f i r s t observation contradicts the view proposed by Holder and Roberts (1985) and Roberts and Holder (1985) that perceptions of stimulus duration are p o s i t i v e l y correlated with si g n a l value, providing one accepts that stimulus duration and stimulus quantity are perceived v i a a common mechanism (Church and Meek, 1984). Food p e l l e t s may be considered to have equivalent signal value as tones, as the c o r r e l a t i o n with reinforcement of the two types of s t i m u l i 206 were equivalent. A l t e r n a t i v e l y , the higher asymptote l e v e l of responding maintained by food p e l l e t cues may be interpreted as r e f l e c t i n g a greater signal value of food p e l l e t s . Whatever interpetation i s adopted, the f i r s t observation indicates that signal value i s not always correlated with perceived stimulus quantity. Furthermore, the inherent value of s t i m u l i i s not correlated with perceived quantity; food p e l l e t s have much greater inherent value than tones, but perceived p e l l e t quantities were less than for equivalent quantities of tones. The f i r s t observation indicates that the perceptions of the numerical attributes of b i o l o g i c a l l y s i g n i f i c a n t s t i m u l i d i f f e r from the numerical attributes of inherently neutral s t i m u l i . Whether th i s contradicts the amodal hypothesis of quantity perceptions (Roberts, 1983; Church and Meek, 1984) i s a moot point. It may be that modality i s not relevant when rats discriminate between d i f f e r e n t quantities of inherently neutral s t i m u l i , but quantity perceptions of inherently s i g n i f i c a n t s timuli are dependent upon mechanisms separate from those underlying the perceptions of stimuli without inherent s i g n i f i c a n c e . The differences i n the perceived quantities of food p e l l e t s and tones are understandable i f i t i s assumed that quantity perceptions are influenced by stimulus value. That i s , the more of a stimulus that i s required by an animal, the smaller a given quantity of that stimulus appears to be. Tones have lower value than food, at least to rats under 207 food deprivation that are not deprived of auditory stimulation. A given quantity of tones i s a larger proportion of a "required" amount than an equivalent quantity of food p e l l e t s . Thus, a ce r t a i n number of tones i s perceived to be greater than an equivalent number of food p e l l e t s . I t i s equally possible that perceptions of tone quantities occurs v i a mechanisms quite d i f f e r e n t from those mediating perceptions of food p e l l e t quantities. Differences i n the perceived quantities of equal numbers of tones and p e l l e t s may be due to differences i n these as yet unidentified mechanisms. Decreasing the amount of food required to minimize deviations from a regulatory set-point, either by increasing consumption (providing the animals with more food while i n th e i r home cages), or by decreasing the set-point (decreasing the sweetness of the food) increases the perceived quantity of food p e l l e t s . The e f f e c t of manipulations of l e v e l of food deprivation cannot be attributed to a general influence of motivational processes on perceptual processes, because tone quantity perceptions are not affected. The e f f e c t of a reduction i n taste i n t e n s i t y i s not a re s u l t of stimulus novelty or a general, hitherto unknown, feature of i n t e n s i t y reductions because tone quantity perceptions are not altered by reducing the in t e n s i t y of the tone s t i m u l i . Therefore, i t can be concluded that increases i n the perceived quantity of food 208 p e l l e t s r e s u l t from reductions i n the value of food. Conversely, increasing the value of food decreases the perceived quantity of food. The observation that haloperidol does not influence the perceived quantity of food p e l l e t s indicates that haloperidol does not a l t e r the hedonic value of food p e l l e t s . The interpretation of the e f f e c t of haloperidol on perceived quantities of tones i s less clear. If the perceptions of tone quantities occur v i a a mechanism sim i l a r to perceptions of food p e l l e t quantities, i t can be concluded that haloperidol increased the value of tones. While counter-intuitive, Irwin et a l . (1983) reported that pimozide apparently enhanced the incentive value of contextual cues associated with food a v a i l a b i l i t y . A l t e r n a t i v e l y , i t may be that haloperidol decreased the signal value of the tones. Holder and Roberts (1985) and Roberts and Church (1985) observed that reducing the signal value of a stimulus resulted i n a bias towards responding to various stimulus durations as though they were perceived to be shorter than usual. Considering the evidence reviewed i n the introduction i n d i c a t i n g that neuroleptics do not a f f e c t stimulus control, t h i s explanation appears implausible. I t appears most l i k e l y that the e f f e c t of haloperidol on perceptions of tone quantity i s related to actions on the neural mechanism underlying "counting" and "timing" behaviors (Szostak and Tombaugh, 1981; Maricq and Church, 1983; Church and Meek, 1984). The most acceptable 209 inte r p r e t a t i o n i s that the discrimination of tone quantities occurs v i a a process that has been referred to as an "i n t e r n a l clock", while the discrimination between d i f f e r e n t numbers of food p e l l e t s i s based on a separate mechanism, which involves an e x i s t i n g behavior-regulatory system. The observation that amphetamine decreases the perceived quantity of food indicates that amphetamine enhances the value of food. This might occur by ef f e c t s similar to reductions i n food consumption (increasing "hunger"). Indeed, amphetamine increases glucose u t i l i z a t i o n (Cole, 1967). However, t h i s e f f e c t occurs with doses much larger than those used i n the present study. Furthermore, amphetamine has an anorectic action, i n doses s l i g h t l y larger than those used i n the present study, but lower than those required to increase metabolic a c t i v i t y (Cole, 1967). A decrease i n perceived quantity of food might also occur by r a i s i n g the set-point, as occurs by sweetening food. Such an e f f e c t of amphetamine may occur i n d i r e c t l y , by enhancing the a c t i v i t y of neurons relaying taste. A l t e r n a t i v e l y , amphetamine may act d i r e c t l y on the processes that are involved i n the establishment of the set-point. I t i s possible that the set-point i s established by a unitary evaluative process, as suggested by Young (1971). In thi s case, amphetamine can be understood to enhance the hedonic value of rewards i n general. Support for th i s i n t e r p r e t a t i o n comes from work that indicates that DA agonists enhance the conditioning of rewarding properties of 210 food to previously neutral s t i m u l i ( H i l l , 1970; Robbins, 1975; Beninger et a l . , 1980; Robbins et a l . , 1983; Taylor and Robbins, 1984). The amphetamine-induced increase i n the reinforcement e f f i c a c y of milk reported by Heyman (1983) and of rewarding e l e c t r i c a l brain stimulation ( G a l l i s t e l and Karras, 1984) also argue that the e f f e c t of amphetamine i s not r e s t r i c t e d to food p e l l e t s . On the other hand, amphetamine does not a l t e r the perceived i n t e n s i t y of rewarding e l e c t r i c a l brain stimulation (Druhan, Martin-Iverson, Wilkie, Fibiger and P h i l l i p s , 1984). This could be taken to indicate that the ef f e c t of amphetamine i s s p e c i f i c for a feeding regulatory system, although cue properties of e l e c r i c a l brain stimulation may not be related to i t s rewarding actions. The rewarding properties of e l e c t r i c a l brain stimulation of some brain s i t e s may be related to ef f e c t s on a feeding regulatory system (Delius and Pellander, 1982; Carr and Simon, 1984), and i t may be that amphetamine affects reinforcement e f f i c a c y of e l e c t r i c a l brain stimulation of only these s i t e s . I t i s also possible that i n addition to i t s enhancement of the hedonic value of food, amphetamine decreases the perceived e f f o r t required to respond. Such an action may be the basis of amphetamine-induced increases i n reinforcement e f f i c a c y i n ICSS procedures. Thus, the generality of amphetamine's e f f e c t s on hedonic value are not presently c l e a r . The increase i n the perceived quantity of tones produced by amphetamine i s best interpreted i n the same terms as the e f f e c t s of haloperidol. That i s , amphetamine appears to influence the rate of an i n t e r n a l clock (Maricq et a l . , 1981; Maricq and Church, 1983; Church and Meek, 1984; Spetch and T r e i t , 1984). The fact that amphetamine has opposite e f f e c t s on tone and food quantity perceptions further supports the view that d i f f e r e n t mechanisms underlie these two perceptual processes. Theoretical Implications The hypothesis that central DA transmission plays a c r i t i c a l role i n the mediation of the hedonic attributes of rewards has been championed by Wise and his colleagues (see Introduction). I t resembles the more general "stimulus e f f i c a c y hypothesis", a version of which was f i r s t proposed by Dews and Morse (1961) to account for the behavioral e f f e c t s of DA receptor antagonists, and which has been further elaborated by Clody and Carlton (1980). The p o s s i b i l i t y that central dopamine systems may be involved i n reinforcement processes has arisen from research of four d i s t i n c t phenomena: se l f - s t i m u l a t i o n of various brain s i t e s with e l e c t r i c a l current, self-administration of drugs, conditioned place preferences, and operant behavior. The l i t e r a t u r e reviewed i n the introduction indicated that while DA i s involved i n the expression of behavior associated with ICSS supported by electrodes implanted i n at 212 least some brain regions, the exact nature of i t s involvement i s not c l e a r . The strongest evidence i n support of the view that DA-releasing neurons activate brain mechanisms underlying "pleasure" i s that DA agonists are self-administered by animals and r e l i a b l y produce conditioned place preferences. However, there are a number of controversial issues associated with these apparent rewarding e f f e c t s . Do animals " l i k e " DA agonists because they d i r e c t l y produce "pleasure", or do they produce a l e v e l of sensory and/or motor e x c i t a b i l i t y that animals prefer to maintain? I t i s possible that rather than d i r e c t l y a c t i v a t i n g reward c i r c u i t s i n the brain, DA agonists produce other e f f e c t s that are "evaluated" by these c i r c u i t s and are perceived as desirable. Thus, psychostimulants may produce t h e i r reward-l i k e actions i n d i r e c t l y . I t i s not c e r t a i n that the properties of DA agonists that are responsible for t h e i r reward-like e f f e c t s are related to t h e i r actions on DA neurons (Martin-Iverson et a l . , 1985). However, evidence has been provided that the motor stimulant e f f e c t s of DA agonists are not related to t h e i r rewarding actions (Martin-Iverson, et a l . , 1985; DiScala et a l . , 1985; Martin-Iverson, Radke and Vincent, i n preparation; Mithani, Martin-Iverson and Fibiger, i n preparation). Therefore, i f the rewarding actions of DA agonists are i n d i r e c t , they must occur through some process other than motor stimulation. The evidence for a role of DA i n natural reinforcement processes i s also controversial. Much of the experimentation i n t h i s regard has r e l i e d upon the use of DA receptor antagonists (neuroleptics). The fact that at least some DA neurons are important i n motor processes makes i t d i f f i c u l t to dissociate the eff e c t s of neuroleptics on these functions from a putative function i n reward. While the evidence reviewed i n the Introduction provides confidence that many of the actions of neuroleptics on operant behavior are not the r e s u l t of simple motor impairments, the p o s s i b i l i t y that high l e v e l integration of sensory and motor processes are disrupted by neuroleptics cannot be discounted. The DA hypothesis of reward proposes that DA-releasing neurons mediate to some extent a unitary process underlying animals' evaluation of environmental events, l i k e l y i n terms of t h e i r p o t e n t i a l b i o l o g i c a l u t i l i t y (Cabanac, 1971; Young, 1977). The goal of thi s evaluative process i s to a l t e r behavior i n order to maximize the occurrence of useful events (and minimize the occurrence of harmful events). In humans, thi s process sometimes results i n a phenomenological experience of pleasure and i t i s often assumed that the same occurs i n non-human animals. This assumption i s quite l i k e l y correct, given the s i m i l a r i t y i n central nervous systems across mammals. Thus, i t appears reasonable that i f DA systems are involved i n t h i s process of evaluation, then alte r a t i o n s i n the functioning of DA neurons w i l l a l t e r 214 "sensations" of pleasure. Indeed, th i s was suggested by Wise, Spindler, DeWit and Gerber (1978). A l t e r n a t i v e l y , the evaluation of events may not be affected by DA systems. Rather, DA-releasing neurons may be involved i n the a b i l i t y of the outcome of the evaluative process to determine behavior. The separation of a sensory evaluative process from the behavioral consequences of that process has recently been questioned. Sinnamon (1982) proposed that the evaluative process i s concerned not only with the sensory attributes of a stimulus, but also with the e f f o r t required to obtain the stimulus. While the b i o l o g i c a l s i g n i f i c a n c e of a stimulus and the work required to obtain i t may very well be compared at some l e v e l of information processing, the evaluation of each must occur separately at some l e v e l . In fact, evidence has been reported that suggests that the d i f f e r e n t i a l hedonic values of a v a r i e t y of sweet-tasting substances are r e f l e c t e d by d i f f e r e n t i a l a c t i v i t y of primary sensory neurons (Pfaffman, 1982). Changes i n behavior towards a b i o l o g i c a l l y s i g n i f i c a n t stimulus produced by neuroleptics may be due to a l t e r a t i o n s i n the evaluation of the stimulus, i n the evaluation of the e f f o r t required to obtain the stimulus, or at the l e v e l of comparing the results of the two evaluations. Furthermore, there are a number of lev e l s within each evaluative process at which neuroleptics may exert t h e i r e f f e c t s . An animal may perceive a p a r t i c u l a r response as requiring more e f f o r t after neuroleptic 215 treatment because the threshold l e v e l of a c t i v a t i o n of motor systems may have been increased, or because of a l t e r a t i o n s i n the evaluation process i t s e l f . S i m i l a r l y , the hedonic value of rewards may be decreased because of a decrease i n the sensory input to the brain substrate of the evaluative process, or because of changes i n that process. It i s possible that there i s no unitary evaluative process. Each behavior-reward contingency may be "evaluated" by s p e c i f i c f u n c t i o n a l l y relevant systems. If t h i s i s the case, i t would be necessary to examine the involvement of s p e c i f i c neurotransmitter systems i n each functional system independently. Generalizing r e s u l t s from one system to another would be detrimental to the understanding of the physiological processes underlying behavior regulation. The present results suggest that the reduction i n reinforcement e f f i c a c y produced by neuroleptics (Zarevics and Setler, 1979; Heyman, 1983; S t e l l a r et a l . , 1983; Wauquier et a l . , 1983; G a l l i s t e l and Karras, 1984) i s not related to a reduction i n perceived hedonic value of rewards. The food quantity generalization gradient procedure has a p a r t i c u l a r advantage i n that the PSE does not r e f l e c t a l t e r a t i o n s i n either the e f f o r t required to produce responses or the perceived response e f f o r t . The PSE i s a function of the sensory c h a r a c t e r i s t i c s of food s t i m u l i . By default, i t appears that neuroleptic-induced reductions i n reinforcement value are the r e s u l t of evaluations of increased e f f o r t required to obtain reward, or a reduction 216 i n the significance of the hedonic value of rewards when weighed against e f f o r t . The observation that haloperidol produced a dose-dependent increase i n response latencies indicate that neuroleptic-treated rats may be less "eager" to engage i n lever-pressing. According to the stimulus e f f i c a c y hypothesis of neuroleptic action proposed' by Clody and Carlton (1980), neuroleptics disrupt performance controlled by less e f f e c t i v e s t i m u l i to a greater degree than performance controlled by more e f f e c t i v e s t i m u l i . In the present case, food p e l l e t cues were more e f f e c t i v e than tone cues, r e f l e c t e d by both a c q u i s i t i o n rate and l e v e l of asymptotic performance. Therefore, haloperidol should have had a stronger e f f e c t on disrupting performance on tone t r i a l s than food t r i a l s . In fact, the decrease i n performance (slope) on both kinds of t r i a l s produced by the highest dose of haloperidol was 20%. Furthermore, the r e l a t i v e increase i n latencies produced by the highest dose of haloperidol was equivalent between the two t r i a l types (about 100%). In absolute terms, responses to food t r i a l s were affected to a greater extent i n terms of both gradient slopes and response latencies. Thus, the present evidence argues against a stimulus e f f i c a c y hypothesis. As the response requirements for both kinds of t r i a l s were the same, these results are supportive of a response e f f i c a c y hypothesis of neuroleptic action. 217 How the present finding that amphetamine enhances the hedonic value of food relates to the primary r e i n f o r c i n g properties of DA agonists awaits further investigation. It i s possible that animals self-administer DA agonists because the hedonic value of normally rewarding s t i m u l i are enhanced. In most self-administration situations, the number of r e i n f o r c i n g stimuli are limited. Hence th i s does not appear to be a p a r t i c u l a r l y appealing hypothesis. If DA agonists enhance the hedonic value of proprioceptive or kinesthetic s t i m u l i , as well as exteroceptive s t i m u l i , then i t appears reasonable that DA agonists are self-administered because of t h e i r reward-enhancing actions. This view, proposed by Katz (1982), may also account for stereotypy. That i s , the enhanced hedonic value of kinesthetic s t i m u l i produces p o s i t i v e feedback, u n t i l behavioral control i s r e s t r i c t e d l a r g e l y to the active kinesthetic inputs. P a r t i a l support for t h i s view has been provided by an early experiment demonstrating that apomorphine increases responding reinforced by access to a wooden block which the rats could gnaw (Robinson, Daley and Wolff, 1967). It i s possible that neuroleptics decrease the hedonic value of actions, as suggested by Katz, but not the value of s t i m u l i , while stimulants enhance the value of both. The e f f e c t s of haloperidol and amphetamine on the numerical attributes of s t i m u l i deserve further consideration. The present results indicate that DA systems are involved i n "counting". Church and Meek (1983) have 218 proposed that both counting and timing are based on a common mechanism. Indeed, the decrease and increase i n perceived quantity respectively produced by neuroleptics and stimulants i s s i m i l a r to decreases and increases i n duration perceptions. Thus, DA systems appear to be a l i n k i n an " i n t e r n a l clock" mechanism. Such a clock l i k e l y has major implications for behavior. Behavior occurs within a temporal framework; i t i s sequential i n nature. It may well be that a major function of central DA systems i s to coordinate behavior over time. Roberts and Holder (1985) proposed that the i n t e r n a l clock r e f l e c t s the signal value of s t i m u l i . The present results indicate that i f t h i s i s true, i t holds only for stimuli that do not have inherent b i o l o g i c a l s i g n i f i c a n c e . Nevertheless, t h i s may be an important function of an i n t e r n a l clock, e s p e c i a l l y i n human beings, whose behavior i s c ontrolled by a plethora of predictive s t i m u l i with no innate b i o l o g i c a l s i g n i f i c a n c e . The delay i n responding to the cues observed i n the present experiment aft e r treatments with haloperidol may r e f l e c t a slowing of the i n t e r n a l clock. That amphetamine produces the converse e f f e c t on response latencies to food cues supports t h i s i n t e r p r e t a t i o n . However, amphetamine did not decrease response latencies on tone t r i a l s . Thus, i t s e f f e c t may have been s p e c i f i c for food s t i m u l i . An alternative explanation for the lack of e f f e c t of amphetamine on response latencies to tone cues i s that these 219 response latencies were normally very short. There might have been a " f l o o r " e f f e c t . Be that as i t may, the role of DA systems i n the temporal coordination of behavior merits further research. The f a c t that drugs a f f e c t i n g DA systems apparently a l t e r the function of an i n t e r n a l timing mechanism suggests another point for consideration. It i s l i k e l y that central DA systems are involved i n a vari e t y of behavioral and cognitive processes. Changes i n complex behaviors, such as those studied i n operant procedures, l i k e l y r e f l e c t a l t e r a t i o n s i n many of these processes. Interpretations of the mechanisms underlying drug-induced al t e r a t i o n s i n behavior must bear t h i s i n mind. The Progressive Nature of Neuroleptic-Induced  Suppression of Behavior: Some Speculations One of the strongest points i n favor of the anhedonia hypothesis i s that i t accounts for the progressive nature of the response d e f i c i t observed aft e r neuroleptic treatment. Simple motor impairment hypotheses of neuroleptic actions have f a i l e d to provide a s a t i s f a c t o r y explanation for t h i s aspect of the response impairment. Yet the s i m i l a r i t y of neuroleptic-induced decreases i n response rate to the effe c t s of non-reward are limited (see Introduction: S i m i l a r i t y of E f f e c t s of Neuroleptics to Reward Omission). The progressive nature of the response-suppressant ef f e c t s of neuroleptics prompted investigators to examine possible behavioral reasons for the d e f i c i t . Cumulative drug 220 e f f e c t s have generally been ruled out as a possible contributing factor. Investigators have examined the action of neuroleptics given outside of the t e s t i n g s i t u a t i o n , and found that such treatments do not induce a progressive decrease i n responding, as i s seen when neuroleptics are associated with the t e s t i n g s i t u a t i o n (eg. Wise et a l . , 1978). However, i t i s possible that the progressive response-suppressant e f f e c t s of neuroleptics may develop with repeated association with a p a r t i c u l a r state of a set of neurons. Response suppressant e f f e c t s of a neuroleptic may depend upon the development of responses to the drug by s p e c i f i c neurons that occur only when DA receptor blockade i s concomitant with a behaviorally or s i t u a t i o n a l l y s p e c i f i c change i n f i r i n g rate of those neurons. For example, Poulos and Hinson (1982) reported that tolerance to the c a t a l e p t i c action of haloperidol occurs only when animals are tested i n the same environment i n which they received p r i o r haloperidol i n j e c t i o n s ; tolerance was not exhibited when animals were tested i n a d i f f e r e n t environment. The actions of neuroleptics change with repeated treatments, and not a l l of the e f f e c t s of chronic administration of neuroleptics can be related to DA receptor s u p e r s e n s i t i v i t y (Dewey and Fibiger, 1983; Poulos and Hinson, 1982). Furthermore, haloperidol reduces the metabolic rates of c e l l s i n a number of brain areas that do not correspond to the areas containing DA terminals. The time-course of these metabolic alte r a t i o n s does not correspond with that of DA receptor 221 blockade (Pizzolato, Soncrant and Rapoport, 1984). Present knowledge of the pharmacodynamics of neuroleptics i s far too incomplete to rule out a pharmacological basis for the progressive nature of the behavioral e f f e c t s of neuroleptics. Some unpublished data from th i s laboratory (Ortmann, Martin-Iverson and Fibiger) are pertinent to t h i s point. Rats were trained on a variable r a t i o schedule i n which response rates were extremely high, between 1000 and 2250 responses per 15 minute sessions. Repeated treatment with neuroleptics, either on consecutive days, or every t h i r d day with drug-free tests interspersed between drug tests, produced a gradual diminishing response rate over the f i r s t 4 drug days ( a l b e i t response rates did not decrease below level s normal for a CRF reinforcement schedule). However, with continued drug testing, rats began to exhibit 'escape' from the response suppressant e f f e c t s : d i f f e r e n t rats on d i f f e r e n t days would exhibit baseline levels of responding. A p a r t i c u l a r animal may exhibit response suppression on one drug day, baseline le v e l s the next, and response suppression the following day, i n an e n t i r e l y unpredictable fashion. Such drug actions on behavior are not e a s i l y explicable i n behavioral terms, but may r e f l e c t complex pharmacodynamic-neurophysiological r e l a t i o n s . Thus, the progressive nature of the behavioral response to neuroleptics i n the operant paradigm may not be p a r t i c u l a r l y relevant to the nature of the behavioral impairment. * 222 Whatever the cause of the progressive nature of neuroleptic e f f e c t s on behavior, the present study indicates that neuroleptics do not attenuate the hedonic value of st i m u l i . Furthermore, i t i s of in t e r e s t that not a l l the eff e c t s of amphetamine are the mirror-image of those of haloperidol. The underlying mechanisms of amphetamine's enhancement of the hedonic value of rewards i s an in t e r e s t i n g area for future research. 223 BIBLIOGRAPHY Amsel, A., The role of f r u s t r a t i v e nonreward i n noncontinuous reward sit u a t i o n s . Psychol. B u l l . , 55 (1958) 102-119. Anand, B.K. and Brobeck, J.R., Loca l i z a t i o n of a feeding center i n the hypothalamus of the rat. Proc. Soc. Exp. B i o l . Med., 77 (1951) 323-324. Asin, K.E. and Fibiger, H.C., Force requirements i n lever-pressing and responding after haloperidol. Pharmac. Biochem. Behav., 20 (1984) 323-326. Beck, R.C., Motivation: Theories and P r i n c i p l e s , New Jersey, Prentice H a l l , Inc. (1978). Beninger, R.J., A comparison of the eff e c t s of pimozide and non-reinforcement on discriminated operant responding i n rats. Pharmac. Biochem. Behav., 16 (1982) 667-669. Beninger, R.J., Hahn, B.L., Pimozide blocks establishment but not expression of amphetamine-produced environment-s p e c i f i c conditioning. Science, 220 (1983) 1304-1306. Beninger, R.J., Hanson, D.R. and P h i l l i p s , A.G., The ef f e c t s of pipradol on the ac q u i s i t i o n of responding with conditioned reinforcement: A role for sensory preconditioning. Psychopharm. 69 (1980) 235-242. Bindra, D. , A u n i f i e d i n t e r p r e t a t i o n of emotion and motivation. Annals N.Y. Acad. S c i . , 159 (1969) 621-1121. Breese, G.R. and Cooper, B.R., Relationship of dopamine neural systems to the maintenance of sel f - s t i m u l a t i o n . In: Neurotransmitter Balances Regulating Behavior, Eds: E.F. Domino and J.M. Davis, Ann Arbor, (1975). Bugelski, B.R., Extinction with and without sub-goal reinforcement. J. Comp. Psychol., 26 (1938) 121-134. Cabanac, M. , Physiological role of pleasure. Science, 173 (1971) 1103-1107. Cabanac, M., Duclaux, R. and Spector, N.H., Sensory feedback i n regulation of body weight: i s there a ponderostat? Nature, 229 (1971) 125-127. Camp, C.H. and Ettenberg, A., Comparable e f f e c t s of haloperidol and p a r t i a l reinforcement on the resistance to extinctio n of a food-rewarded runway task i n rats. Soc. Neurosci. Abstracts #77.6 (1984) 265. 224 Capaldi, E.J., Stimulus s p e c i f i c i t y : nonreward. J. Exp. Psychol., 72 (1966) 410-414. Capretta, P.J., Saccharin consumption under varied conditions of hunger drive. J. Comp. Physiol. Psychol., 55 (1962) 656-660. Carr, K.D. and Simon, E.J., Potentiation of reward by hunger i s opioid mediated. Brain Res., 297 (1984) 369-373. Church, R.M. and Meek, W.H., The numerical a t t r i b u t e of s t i m u l i . In: Animal Cognition, Eds: H.L. Roitblat, T.G. Bever and H.S. Terrace, New Jersey, Lawrence Erlbaum Associates (1984). Clavier, R.M. and Fibiger, H.C., On the role of ascending catecholaminergic projections i n i n t r a c r a n i a l s e l f -stimulation of the substantia nigra. Brain Res., 131 (1977) 271-286. Clavier, R.M., Fibiger, H.C. and P h i l l i p s , A.G., Evidence that s e l f - s t i m u l a t i o n of the region of the locus coeruleus i n rats does not depend upon noradrenergic projections to the telencephalon. Brain Res., 113 (1976) 71-81. Clavier, R.M. and Routtenberg, A., In search of reinforcement pathways: A neuroanatomical Odyssey. In: Biology of Reinforcement: Facets of Brain Stimulation Reward, Ed: A. Routtenberg, New York, Academic Press, (1980) . Clody, D.E. and Carlton, P.L., Stimulus e f f i c a c y , chlorpromazine, and schizophrenia. Psychopharm. 69 (1980) 127-131. Cole, S.O., Experimental e f f e c t s of amphetamine: A review. Psychol. B u l l . , 68 (1967) 81-90. C o l l i n s , R.J., Weeks, J.R., Cooper, M.M., Good, P.I. and Russel, R.R., Prediction of abuse l i a b i l i t y of drugs using IV self-administration by rats. Psychopharm. 82 (1984) 6-13. Crespi, L.P., Quantitative v a r i a t i o n of incentive and performance i n the white r a t . Amer. J. Psychol., 55 (1942) 467-517. Crespi, L.P., Amount of reinforcement and l e v e l of performance. Psychol. Rev., 51 (1944) 341-357. Crow, T.J., A map of the rat mesencephalon for e l e c t r i c a l brain stimulation. Brain Res., 36 (1972) 265-273. 225 Dahlstrom, A. and Fuxe, K. Evidence for the existence of mono-aminee neurons i n the central nervous system. I. Demonstration of monoamines i n the c e l l bodies of brain stem neurons. Acta. Physiol. Scand. Suppl. 232, 64 (1964) 1-55. Davis, J.D., The effectiveness of some sugars i n stimulating l i c k i n g behavior i n the rat. Physiol. Behav., 11 (1973) 39-45. Delius, J.D. and Pellander, K., Hunger dependence of e l e c t r i c a l brain s e l f - s t i m u l a t i o n i n the pigeon. Physiol. Behav., 28 (1982) 63-66. Denny, M.R. and Adelman, H.M., E l i c i t a t i o n theory: I. An analysis of two t y p i c a l learning s i t u a t i o n s . Psychological Review, 62 (1955) 290-296. Deutsch, J.A. and Howarth, C.I., Some tests of a theory of i n t r a c r a n i a l s e l f - s t i m u l a t i o n , Psychol. Rev., 70 (1963) 444-460. Dewey, K.J. and Fibiger, H.C., The ef f e c t s of dose and duration of chronic pimozide administration on dopamine receptor s u p e r s e n s i t i v i t y . Naunyn-Schmiedeberg's Arch. Pharmacol., 322 (1983) 261-270. DeWit, H. and Wise, R.A., Blockade of cocaine reinforcement i n rats with the dopamine receptor blocker pimozide, but not with the noradrenergic blockers phentolamine or phenoxybenzamine. Canad. J. Psychol., 31 (1977) 195-203. Dews, P.B. and Morse, W.H., Behavioral pharmacology. Annu. Rev. Pharmacol., 1 (1961) 145-174. Di Scala, G., Martin-Iverson, M.T., P h i l l i p s , A.G. and Fibiger, H.C., The ef f e c t s of progabide (SL 76002) on locomotor a c t i v i t y and conditioned place preference induced by d-amphetamine. European J. Pharmacol., 107 (1985) 271-274. Druhan, J.P., Martin-Iverson, M.T., Wilkie, D., Fib i g e r , H.C. and P h i l l i p s , A.G., Pharmacological d i s s o c i a t i o n of the cue properties from the r e i n f o r c i n g e f f e c t s of ventral tegmental brain stimulation. Soc. Neurosci. Abstr. 10 (1984) #337.8. Dunham, J. The nature of r e i n f o r c i n g s t i m u l i . In: Handbook of Operant Behavior, Eds: W.K. Honig and J.E.R. Staddon, New Jersey, Prentice-Hall (1977). 226 Esses, V.M. and Herman, CP., P a l a t a b i l i t y of sucrose before and aft e r glucose ingestion i n dieters and nondieters. Physiol. Behav., 32 (1984) 711-715. Ettenberg, A., Conditioned taste preference and response rate as measures of brain-stimulation reward: A comparison. Physiol. Behav., 24 (1980) 755-758. Ettenberg, A., Koob, G.F. and Bloom, F.E., Response a r t i f a c t i n the measurement of neuroleptic-induced anhedonia. Science, 209 (1981) 357-359. Ettenberg, A., P e t t i t , H.O., Bloom, F.E. and Koob, G.F., Heroin and cocaine intravenous self-administration i n rat s : Mediation by separate neural systems, Psychopharm., 78 (1982) 204-209. Ettenberg, A. and White, N., Pimozide attenuates conditioned taste preferences induced by se l f - s t i m u l a t i o n i n rat s . Pharmac. Biochem. Behav., 15 (1981) 915-919. Evenden, J.L. and Robbins, T.W., Dissociable e f f e c t s of d-amphetamine, chlordiazepoxide and alpha-flupenthixol on choice and rate measures of reinforcement i n the rat. Psychopharm., 79 (1983) 180-196. Faustman, W.O. and Fowler, S.C, Use of response duration to dis t i n g u i s h the effects of haloperidol from nonreward. Pharmac. Biochem. Behav., 15 (1981) 327-329. Faustman, W.O. and F u l l e r , S.C, An examination of methodological refinements, clozapine and fluphenazine i n the anhedonia paradigm. Pharmac. Biochem. Behav., 17 (1982) 987-993. Fibiger, H.C, Drugs and reinforcement mechanisms: A c r i t i c a l review of the catecholamine theory. Ann. Rev. Pharmacol. Tox., 18 (1978) 37-56. Fibiger, H.C, Carter, D.A. and P h i l l i p s , A.G., Decreased i n t r a c r a n i a l s e l f - s t i m u l a t i o n after neuroleptics or 6-hydroxydopamine: Evidence for mediation by motor d e f i c i t s rather than by reduced reward. Psychopharm. 47 (1976) 21-27. Fibiger, H.C. and P h i l l i p s , A.G., Role of dopamine and norepinephrine i n the chemistry of reward. J. psychiat. Res., 11 (1974) 135-143. Fouriezos, G. , Hansson, P. and Wise, R.A., Neuroleptic-induced attenuation of se l f - s t i m u l a t i o n reward. J. Comp. Physiol. Psychol., 92 (1978) 659-669. 227 Fouriezos, G. and Wise, R.A., Pimozide-induced extinct i o n of i n t r a c r a n i a l s e l f - s t i m u l a t i o n : response patterns rule out motor or performance d e f i c i t s . Brain Res., 103 (1976) 377-380. Franklin, K.B.J. and McCoy, S.N., Pimozide-induced extinction i n rat s : stimulus control of responding rules out motor d e f i c i t . Pharmac. Biochem. Behav. 11 (1979) 71-75. G a l l i s t e l , CR. , Self-stimulation: The neurophysiology of reward and motivation, In: The Physiological Basis of Memory, Ed: J.A. Deutsch, New York, Academic Press (1973). G a l l i s t e l , C.R., Boytim, M., Gomita, Y. and Klebanoff, L., Does pimozide block the r e i n f o r c i n g e f f e c t of brain stimulation? Pharmac. Biochem. Behav., 17 (1982) 769-781. G a l l i s t e l , CR. and Karras, D., Pimozide and amphetamine have opposing e f f e c t s on the reward summation function. Pharmac. Biochem. Behav., 20 (1984) 73-77. G a l l i s t e l , CR. , Shizgal, P. and Yeomans, J.S., A p o r t r a i t of the substrate for sel f - s t i m u l a t i o n , Psychol. Rev., 88 (1981) 228-273. Gerber, G.J., Sing, J. and Wise, R.A., Pimozide attenuates lever pressing for water reinforcement i n rats. Pharmac. Biochem. Behav., 14 (1981) 201-205. German, D.C and Bowden, D.M., Catecholamine systems as the neural substrate for i n t r a c r a n i a l s e l f - s t i m u l a t i o n : A hypothesis. Brain Res., 73 (1974) 381-419. Glickman, S.E. and S c h i f f , B.B., A b i o l o g i c a l theory of reinforcement. Psychol. Rev., 74 (1967) 81-109. Gramling, S.E., Fowler, S.C and C o l l i n s K.R., Some ef f e c t s of pimozide on nondeprived rats l i c k i n g sucrose solutions i n an anhedonia paradigm. Pharmac. Biochem. Behav., 21 (1984) 617-624. Gratton, A. and Wise, R.A., Hypothalamic reward mechanism: Two f i r s t - s t a g e f i b e r populations with a cholinergic component. Science, 227 (1985) 545-548. Greenshaw, A.J., Sanger, D.J. and Blackman, D.E., The ef f e c t s of pimozide and of reward omission on f i x e d - i n t e r v a l behavior of rats maintained by food and e l e c t r i c a l brain stimulation. Pharmac. Biochem. Behav., 15 (1981) 227-233. 228 Grossman, S.P., Eating or drinking e l i c i t e d by d i r e c t adrenergic or cholinergic stimulation of the hypothalamus, Science, 132 (1960) 301-302. Grossman, S.P., Direct adrenergic and cholinergic stimulation of hypothalamic mechanisms, Amer. J. Physiol., 202 (1962a) 872-882. Grossman, S.P., Effects of adrenergic and cholinergic blocking agents on hypothalamic mechanisms, Amer. J. Physiol., 202 (1982b) 1230-1236. Hanson, S.J. and Timberlake, W. , Regulation during challenge: A general model of learned performance under schedule constraint. Psychol. Rev., 90 (1983) 261-282. Heffner, T.G., Hartman, J.A. and Seiden, L.S., Feeding increases dopamine metabolism i n the rat brain. Science, 208 (1980) 1168-1170. Heth, CD. and Warren, A.G., Response deprivation and response s a t i a t i o n as determinants of instrumental performance: Some data and theory. Anim. Learn. Behav., 6 (1978) 294-300. Hetherington, A.W. and Ranson S.W., The spontaneous a c t i v i t y and food intake of rats with hypothalamic lesions. Amer. J. Physiol., 136 (1942) 609-617. Heyman, G.M., A parametric evaluation of the hedonic and motoric e f f e c t s of drugs: pimozide and amphetamine. J. Expt. Anal. Behav., 40 (1983) 113-122. H i l l , R.T., F a c i l i t a t i o n of conditioned reinforcement as a mechanism of psychomotor stimulation, In: Amphetamines and Related Compounds, Eds: E. Costa and S. G a r a t t i n i , New York, Raven Press (1970) pp. 781-795. Holder, M.D. and Roberts, S., Comparison of timing and c l a s s i c a l conditioning. J. Expt. Psychol.: Anim. Behav. P r o c , 11 (1985) 172-193. Hu l l , C.L., P r i n c i p l e s of Behavior. New York, Appleton-Century-Crofts (1943). Irwin, J., Tombaugh, T.N., Zacharko, R.M. and Anisman, H. , Al t e r a t i o n of exploration and the response to food associated cues after treatment with pimozide. Pharmac. Biochem. Behav., 18 (1983) 235-246. 229 Johanson, C E . and Schuster, C.R., A choice procedure for drug r e i n f o r c e r s : cocaine and methylphenidated i n the rhesus monkey, J. Pharm. Exp. Ther., 193 (1975) 676-688. Katz, R.J., Dopamine and the l i m i t s of behavioral reduction - or why aren't a l l schizophrenics f a t and happy? Behav. Brain S c i . , 5 (1982) 60-61. K i l l e e n , P.R., Incentive theory. In: Nebraska Symposium on Motivation: Response Structure and Organization, Ed: D.J. Bernstein, Lincoln, University of Nebraska Press (1982), pp. 169-216. Kirk, R.E., Experimental Design: Procedures For The Behavioral Sciences, Belmont, Brooks/Cole Publishing Co. (1968) p. 143-144. Kraeling, D., Analysis of amount of reward as a variable i n learning. J. Comp. Physiol. Psychol., 54- (1961) 560-564. Leith, N.J., Effects of apomorphine on s e l f - s t i m u l a t i o n responding: does the drug mimic the current? Brain Res., 277 (1983) 129-136. Lett, B.T., Taste potentiation i n poison avoidance learning. In: Harvard Symposium on Quantitative Analysis of Behavior, Vol. 4, H i l l s d a l e , Erlbaum (1982). Leventhal, A.M., Morrell, R.F., Morgan, E.F. and Perkins, C C , The r e l a t i o n between mean reward and mean reinforcement. J. Exp. Psychol., 57 (1959) 284-287. Liebman, J.M., Discriminating between reward and performance: A c r i t i c a l review of i n t r a c r a n i a l s e l f -stimulation methodology. Neurosci. Biobehav. Rev., 7 (1983) 1-28. Liebowitz, S.F., Reciprocal hunger-regulating c i r c u i t s involving alpha- and beta-adrenergic receptors located, respectively, i n the ventromedial and l a t e r a l hypothalamus. Proc. Nat. Acad. S c i . , 67 (1970) 1063-1070. Liebowitz, S.F., Hypothalamic alpha- and beta-adrenergic systems regulate both hunger and t h i r s t i n the rat. Proc. Nat. Acad. S c i . , 68 (1971) 332-334. Lyness, W.H., F r i e d l e , N.M. and Moore, K.E., Destruction of dopaminergic nerve terminals i n nucleus accumbens: E f f e c t on d-amphetamine self-administration. Pharmac. Biochem. Behav., 11 (1979) 553-556. 230 Mackintosh, N.J., The Psychology of Animal Learning. London, Academic Press (1974). Maricq, A.V. and Church, R.W., The d i f f e r e n t i a l e f f e c t s of haloperidol and methamphetamine of time estimation i n the ra t . Psychopharm., 79 (1983) 10-15. Maricq, A.V., Roberts, S. and Church, R.W., Methamphetamine and time estimation. J. Expt. Psychol.: Anim. Behav. Proc., 7 (1981) 18-30. Marshall, J.F. and Teitelbaum, P., New considerations i n the neuropsychology of motivated behaviors. In: Handbook of Psychopharmacology, Vol. 8, Eds: L.L. Iversen, S.D. Iversen and S.H. Snyder, New York, Plenum Press (1977). Martin-Iverson, M.T., Ortmann, R. and Fibiger, H.C, Place preference conditioning with methylphenidate and nomifensine, Brain Res., 332 (1985) 59-67. Mason, S.T., Beninger, R.J., Fibiger, H.C. and P h i l l i p s , A.G., Pimozide-induced suppression of responding: Evidence against a block of food reward. Pharmac. Biochem. Behav., 12 (1980) 917-923. McCall, R.B. and Appelbaum, M.I., Bias i n the analysis of repeated-measures designs: Some alternative approaches. Child Devel., 44 (1973) 401-415. Meltzer, D. andBrahlek, J.A., Quantity of reinforcement and f i x e d - i n t e r v a l performance. Psychon. S c i . , 12 (1968) 207-208. M i l l e r , N.E., Chemical coding of behavior, Science, 148 (1965) 328-338. M i t c h e l l , M.J., Nicolaou, N.M., Arbuthnott, G.W. and Yates, CM., Increases i n dopamine metabolism are not a general feature of i n t r a c r a n i a l s e l f - s t i m u l a t i o n . L i f e S c i . , 30 (1982) 1081-1085. Mogenson, G.J. and P h i l l i p s , A.G., Motivation: A psychological construct i n search of a physiological substrate. In: Progress i n Physiological Psychology and Psychobiology, Eds: J.M. Sprague and A.E. Epstein, New York, Academic Press (1976) 189-244. Mora, F., P h i l l i p s , A.G., Koolhaas, J.M. and Rol l s , E.T., Prefrontal cortex and neostriatum s e l f - s t i m u l a t i o n i n the ra t : D i f f e r e n t i a l e f f e c t s produced by apomorphine. Brain Res. B u l l . , 1 (1976) 421-424. 231 Morely, M.J., Bradshaw, CM. and Szabadi, E., The e f f e c t of pimozide on v a r i a b l e - i n t e r v a l performance: A test of the 'anhedonia' hypothesis of the mode of action of neuroleptics. Psychopharm., 84 (1984) 531-536. Moskowitz, H.R., Sensations, measurement and pleasantness: Confessions of a latent i n t r o s p e c t i o n i s t . In: Taste and Development: The Genesis of Sweet Preference, Ed: J.M. Weiffenbach, Bethesda, U.S. Dept. Health, Education and Welfare (1977). Neuringer, A.J., Effects of reinforcement magnitude on choice and rate of responding. J. Exp. Anal. Behav., 10 (1967) 417-424. Nielsen, J.A., Duda, N.J., Mokler, D.J. and Moore, K.E., Self-administration of central stimulants by rats : A comparison of the ef f e c t s of d-amphetamine, methylphenidate and McNeil 4612, Pharm. Biochem. Behav., 20 (1984) 227-232. Olds, J. and Milner, P. Positive reinforcement produced by e l e c t r i c a l stimulation of septal area and other regions of rat brain. J. comp. Physiol. Psychol., 47 (1954) 419-427. Pavlov, I.P., Conditioned Reflexes,' Oxford, Oxford University Press (1927). Pfaffman, C , Taste: A model of incentive motivation. In: The Physiological Mechanisms of Motivation, Ed: D.W. P f a f f , New York, Springer-Verlag, 1984. P h i l l i p s , A.G., Brain reward c i r c u i t r y : A case for separate systems. Brain Res. B u l l . , 12 (1984) 195-201. P h i l l i p s , A.G., Brooke, S.M., and Fibiger, H.C, Eff e c t s of amphetamine isomers and neuroleptics on s e l f - s t i m u l a t i o n i n from the nucleus accumbens and dorsal noradrenergic bundle. Brain Res., 85 (1975) 13-22. P h i l l i p s , A.G., Carter, D.A. and Fibiger, H.C, Dopaminergic substrates of i n t r a c r a n i a l s e l f - s t i m u l a t i o n i n the caudate-putamen. Brain Res., 104 (1976) 221-232. P h i l l i p s , A.G., Carter, D.A. and Fibiger, H.C, Decreased i n t r a c r a n i a l s e l f - s t i m u l a t i o n after neuroleptics or destruction of the nigroneostriatal bundle: performance or reinforcement d e f i c i t ? In: Brain Stimulation Reward, Eds: A. Wauquier and E.T. Roll s , Netherlands, North-Holland Publishing Co. (1977). 232 P h i l l i p s , A. G. and Fibiger, H.C, Dopaminergic and noradrenergic substrates of p o s i t i v e reinforcement: D i f f e r e n t i a l e f f e c t s of d- and 1-amphetamine. Science, 179 (1973) 575-577. P h i l l i p s , A.G. and Fibiger, H.C, Long-term d e f i c i t s i n stimulation-induced behaviors and s e l f - s t i m u l a t i o n after 6-hydroxydopamine administration i n r a t s . Behav. B i o l . , 16 (1976) 127-143. P h i l l i p s , A.G. and Fibiger, H.C, The role of dopamine i n maintaining i n t r a c r a n i a l s e l f - s t i m u l a t i o n i n the ventral tegmentum, nucleus accumbens and medial prefrontal cortex. Canad. J. Psychol. Rev. 32 (1978) 58-66. P h i l l i p s , A.G. and Fibiger, H.C, Decreased resistance to extinction a f t e r haloperidol: Implications for the role of dopamine i n reinforcement. Pharmac. Biochem. Behav., 10 (1979) 751-760. P h i l l i p s , A.G., Van Der Kooy, D. and Fibiger, H.C, Maintenance of i n t r a c r a n i a l s e l f - s t i m u l a t i o n i n hippocampus and o l f a c t o r y bulb following regional depletion of noradrenaline. Neurosci. Lett., 4 (1977) 77-84. Pickens, R.J. and Thompson, T., Cocaine-reinforced behavior i n r a t s : E f f e c t s of reinforcement magnitude and f i x e d r a t i o si z e , J. Pharmacol. Exp. Ther., 161 (1968) 122-129. Pizzolato, G., Soncrant, T.T. and Rapoport, S.I., Haloperidol and cerebral metabolism i n the conscious rat: r e l a t i o n to pharmacokinetics. J. Neurochem., 43 (1984) 724-732. Poulos, CX. and Hinson, R., Pavlovian conditional tolerance to haloperidol catalepsy: Evidence of dynamic adaptation i n the dopaminergic system. Science 218 (1982) 491-492. Prado-Alcala, R., Streather, A., Wise, R.A., Brain stimulation reward and dopamine terminal f i e l d s . I I . Septal and c o r t i c a l projections. Brain Res., 301 (1984) 209-219. Prado-Alcala, R. and Wise, R.A., Brain stimulation reward and dopamine terminal f i e l d s . I. Caudate-putamen, nucleus accumbens and amygdala. Brain Res., 297 (1984) 265-273. Premack, D., Toward empirical behavior laws, I: Positive reinforcement. Psychol. Rev., 66 (1959) 219-233. Premack, D., Reinforcement theory. In: Nebraska Symposium on Motivation, Ed: D. Levine, Lincoln, University of Nebraska Press (1965). 233 Premack, D., Catching up with common sense or two sides of a generalization: Reinforcement and punishment., In: The Nature of Reinforcement, Ed: R. Glaser, New York, Academic Press, 1971). Pubols, B.H., Constant versus variable delay of reinforcement. J. Comp. Physiol, 55 (1962) 52-56. Redgrave, P. and Dean, P., In t r a c r a n i a l s e l f - s t i m u l a t i o n , B r i t . Med. B u l l . , 37 (1981) 141-146. Rescorla, R.A. and Durlach, P.J., Within-event learning i n Pavlovian conditioning. In: Information Processing i n Animals: Memory Mechanisms, Eds: N.E. Spear and R.R. M i l l e r , H i l l s d a l e , Erlbaum (1981). Robbins, T.W., The potentiation of conditioned reinforcement by psychomotor stimulant drugs. A test of H i l l ' s hypothesis, Psychopharm. 45 (1975) 103-114. Robbins, T.W., Watson, B.A., Gaskin, M. and Ennis, C , Contrasting interactions of pipradol, d-amphetamine, cocaine, cocaine analogues, apomorphine and other drugs with conditioned reinforcement. Psychopharm., 80 (1983) 113-119. Roberts, D.C.S., Corcoran, M.E. and Fibiger, H.C, On the role of ascending catecholaminergic systems i n intravenous self-administration of cocaine. Pharmac. Biochem. Behav., 6 (1977) 615-620. Roberts, D.C.S., Koob, G.F., Klonoff, P. and Fibiger, H.C, Extinction and recovery of cocaine self-administration following 6-hydroxydopamine lesions of the nucleus accumbens, Pharm. Biochem. Behav., 12 (1980) 781-787. Roberts, S., Properties and function of an i n t e r n a l clock. In: Animal Cognition and Behavior, Ed: R.L. Mellgren, New York, North-Holland (1983). Roberts, S. and Holder, M.D., E f f e c t of c l a s s i c a l conditioning on an i n t e r n a l clock. J. Expt. Psychol.: Anim. Behav. P r o c , 11 (1985) 194-214. Robinson, P., Daley, M. and Wolff, P.C, Apomorphine induced reinforcement. Psychon. S c i . , 7 (1967) 117-118. Roscoe, J.T., Fundamental Research S t a t i s t i c s for the Behavioral Sciences, 2nd Ed. New York, Holt, Rinehart and Winston, Inc. (1975). Seeman, P., Brain dopamine receptors. Pharmacol. Rev., 32 (1980) 229-313. 234 Sevenster, P. Incompatibility of response and reward. In: Constraints on Learning, Ed. R. Hinde and J. Hinde, New York, Academic Press (1973). S h e f f i e l d , F.D., New evidence on the drive-induction theory of reinforcement. In: Current Research i n Motivation, Ed: R.N. Haber, New York, Holt, Rinehart and Winston (1966). Shettleworth, S.J., Food reinforcement and the organization of behavior i n golden hamsters. In: Constraints on Learning, Ed: R. Hinde and J. Hinde, New York: Academic Press (1973). Sinnamon, H.M. The reward-effort model: An economic framework for examining the mechanism of neuroleptic action. Behav. Brain S c i . , 5 (1982). 73-75. Smith, M.P. and Capretta, P.J., Eff e c t s of drive l e v e l and experience on the reward value of saccharine solutions. J. Comp. Physiol. Psychol., 49 (1956) 553-557. Smith, M.P. and Duffy, M., Consumption of sucrose and saccharine by hungry and satiated rats. J. Comp. Physiol. Psychol., 50 (1957a) 65-69. Smith, M.P. and Duffy, M. , Some physiological factors that regulate eating behavior. J. Comp. Physiol. Psychol., 50 (1957b) 601-608. Spear, N.E. and Kucharski, D. , Ontogenetic differences i n the processing of multi-element s t i m u l i . In: Animal Cognition, Eds: H.L. Roitblat, T.G. Bever and H.S. Terrace, H i l l s d a l e , Earlbaum (1984). Spetch, M.L. and T r e i t , D., The e f f e c t of d-Amphetamine on short-term memory for time i n pigeons. Pharmacol. Biochem. Behav., 21 (1984) 663-666. Spetch, M.L. and Wilkie, D.M. A systematic bias i n pigeons* memory for event duration. J. Exp. Psychol. (Anim. Behav.), 9 (1983) 14-30. Spyraki, C. and Fibiger, H.C, Intravenous self-adminis-t r a t i o n of nomifensine i n rats : Implications for abuse pot e n t i a l i n humans, Science, 212 (1981) 1167-1168. Spyraki, C , Fibiger, H.C. and P h i l l i p s , A.G., Dopaminergic substrates of amphetamine-induced place preference conditioning, Brain Res., 253 (1982a) 185-193. 235 Spyraki, C , Fibiger, H.C. and P h i l l i p s , A.G., Cocaine-induced place preference conditioning: lack of e f f e c t s of neuroleptics and 6-hydroxydopamine leaions. Brain Res., 253 (1982b) 195-203. Spyraki, C. , Fibiger, H.C. and P h i l l i p s , A.G., Attenuation by haloperidol of place preference conditioning using food reinforcement. Psychopharm. 77 (1982c) 379-382. Stein, L. , Effects and interactions of imipramine, chlorpromazine, reserpine and amphetamine on s e l f -stimulation: Possible neurophysiological basis of depression, In: Recent Advances i n B i o l o g i c a l Psychiatry, Ed: J. Wortis, New York, Plenum Press (1962). Stein, L. , Self-stimulation of the brain and the control stimulant action of amphetamine. Fedn. P r o c , 23 (1964) 836. Stein, L. and Wise, R.A., Release of norepinephrine from hypothalamus and amygdala by rewarding medial forebrain stimulation and amphetamine. J. comp. Physiol. Psychol., 67 (1969) 189-198. Stein, L. , Wise, R.A. and B e l l u z z i , J.D., Neuropharmacology of reward and punishment. In: Handbook of Psychopharmacology, Vol. 8, Eds: L.L. Iversen, S.D. Iversen and S.H. Snyder, New York, Plenum Press (1977). S t e l l a r , E., The physiology of motivation, Psychol. Rev., 61 (1954) 522. S t e l l a r , E., Sweet preferences and hedonic experience. In: Taste and Development: The Genesis of Sweet Preference, Ed: J.M. Weiffenbach, Bethesda, U.S. Dept. Health, Education Welfare (1977). S t e l l a r , E., Brain mechanisms i n hedonic processes. In: The Physiological Mechanisms of Motivation, Ed: D.W. P f a f f , New York, Springer-Verlag (1982). S t e l l a r , J.R., Kelley, A.E. and Corbett, D., Effects of peripheral and central dopamine blockade on l a t e r a l hypothalamic s e l f - s t i m u l a t i o n : Evidence for both reward and motor d e f i c i t s . Pharmacol. Biochem. Behav., 18 (1983) 433-442. Szostak, C. and Tombaugh, T.N., Use of a fixed consecutive number schedule of reinforcement to investigate the e f f e c t s of pimozide on behavior controlled by i n t e r n a l and external s t i m u l i . Pharmac. Biochem. Behav., 15 (1981) 609-617. 236 Szostak, C , Tombaugh, T.N. and Tombaugh, J., Examination of the e f f e c t s of pimozide on two conditional discrimination problems d i f f e r i n g i n levels of task complexity. Prog. Neuro-Psychopharmacol., 5 (1981) 615-618. Taylor, J.R. and Robbins, T.W., Enhanced behavioural control by conditioned reinforcers following microinjections of d-amphetamine into the nucleus accumbens. Psychopharm., 84 (1984) 405-412. Timberlake, W. and A l l i s o n , J ., An empirical approach to instrumental performance. Psychol. Rev., 81 (1974) 146-164. Tolman, E.C., Purposive Behavior i n Animals and Men. New York, Appleton-Century-Crofts (1932). Tolman, E.C., The nature and functioning of wants. Psychol. Rev., 56 (1949) 357-369. Tombaugh, T.N., Effects of pimozide on nondiscriminated and discriminated performance i n the pigeon. Psychopharm. 73 (1981) 137-141. Tombaugh, T.N., Anisman, H. and Tombaugh, J., Extinction and dopamine receptor blockade after intermittent reinforcement t r a i n i n g : F a i l u r e to observe functional equivalence. Psychopharm. 70 (1980) 19-28. Tombaugh, T.N., Grandmaison, L.J. and Zito, K.A., Establishment of secondary reinforcement i n sign tracking and place preference tests following pimozide treatment. Pharmac. Biochem. Behav., 17 (1982) 665-670. Tombaugh, T.N., Ritch, M.A. and Shepherd, D.T., Effects of pimozide on accuracy of performance and d i s t r i b u t i o n of correct responding on a simultaneous discrimination task i n the rat. Pharmac. Biochem. Behav., 13 (1980) 859-862. Tombaugh, T.N., Szostak, C. and M i l l s , P., Failu r e of pimozide to disrupt the acquistion of light-dark and s p a t i a l discrimination problems. Psychopharm., 79 (1983) 161-168. Tombaugh, T.N., Tombaugh, J. and Anisman, H. , E f f e c t s of dopamine receptor blockade on alimentary behaviors: Home cage food consumption, magazine t r a i n i n g , operant a c q u i s i t i o n and performance. Psychopharm. 66 (1979) 219-225. Ungerstedt, U., 6-hydroxy-dopamine induced degeneration of central monoamine neurons. Eur. J. Pharmacol., 5 (1968) 107-110. 237 Ungerstedt, U., Is interruption of the n i g r o - s t r i a t a l dopamine system producing the " l a t e r a l hypothalamus syndrome"? Acta. Physiol. Scand., 81 (1971) 84-95. Valenstein, E.S., Selection of n u t r i t i v e and non-nutritive solutions under d i f f e r e n t conditions of need. J. Comp. Physiol. Psychol., 63 (1967) 429-433. Vi t a l i a n o , P.P., Parametric s t a t i s t i c a l analysis of repeated measures experiments. Psychoneuroend. 7 (1982) 3-13. Wasserman, E.M., Gomita, Y. and G a l l i s t e l , C.R., Pimozide blocks reinforcement but not priming from MFB stimulation i n the ra t . Pharmac. Biochem. Behav., 17 (1982) 783-787. Wauquier, A., Clincke, G.H.C. and Fransen, J.F., Parameter selection i n a rate free test of brain s e l f - s t i m u l a t i o n : Towards an alte r n a t i v e i n t e r p r e t a t i o n of drug e f f e c t s . Behav. Brain Res., 7 (1983) 155-164. Wilson, M.C, Hitomi, M. and Schuster, C.R., Psychomotor stimulant self-administration as a function of dosage per in j e c t i o n i n the Rhesus monkey. Psychopharm., 22 (1971) 271-281. Wise, R.A., Catecholamine theories of reward: A c r i t i c a l review. Brain Res., 152 (1978) 215-247. Wise, R.A., Neuroleptics and operant behavior: The anhedonia hypothesis. Behav. Brain S c i . , 5 (1982) 39-87. Wise, R.A. and Bozarth, M.A., Brain substrates for reinforcement and drug self-administration. Prog. Neuro-Psychopharmacol., 5 (1981) 467-474. Wise, R.A. and Bozarth, M.A., Brain reward c i r c u i t r y : Four c i r c u i t elements "wired" i n apparent series. Brain Res. B u l l . , 12 (1984) 203-208. Wise, R.A. and Colle, L.M., Pimozide attenuates free feeding: Best scores analysis reveals a motivational d e f i c i t . Psychopharm. 84 (1984) 446-451. Wise, R.A., Spindler, J ., DeWit, H. and Gerber, G.J., Neuroleptic-induced "anhedonia" i n rats: pimozide blocks reward q u a l i t y of food. Science, 201 (1978) 262-264. Wise, R.A., Spindler, J. and Legault, L., Major attenuation of food reward with performance-sparing doses of pimozide i n the rat. Canad. J. Psychol., 32 (1978) 77-85. 238 Wise, R.A. and Stein, L., F a c i l i t a t i o n of brain s e l f -stimulation by central administration of norepinephrine. Science, 163 (1969) 299. Wise, R.A., Yokel, R.A. and deWit, H., Both p o s i t i v e reinforcement and conditioned aversion from amphetamine and apomorphine i n rats, Science 191 (1975) 1273-1274. Woolverton, W.L., Effects of antipsychotic compounds i n rhesus monkeys given a choice between cocaine and food. Drug and Alcohol Dep., 8 (1981) 69-78. Xenakis, S. and S c l a f a n i , A., The e f f e c t s of pimozide on the consumption of a palatable saccharin-glucose solution i n the rat. Pharmac. Biochem. Behav., 15 (1981) 435-442. Xenakis, S. and S c l a f a n i , A., The dopaminergic mediation of sweet reward i n normal and VMH hyperphagic rats. Pharmac. Biochem. Behav., 16 (1982) 293-302. Yokel, R.A. and Wise, R.A., Increased lever-pressing for amphetamine afte r pimozide i n ra t s : Implications for a dopamine theory of reward. Science, 187 (1973) 547-549. Yokel, R.A. and Wise, R.A., Attenuation of intravenous amphetamine reinforcement by central dopamine blockade i n rats, Psychopharm. 48 (1976) 311-318. Young, P.T., Hedonic organization and regulation of behavior. Psychol. Rev., 73 (1966) 59-86. Young, P.T., Evaluation and preferences i n behavioral development. Psychol. Rev., 75 (1968) 222-241. Young, P.T., Role of hedonic processes i n the development of sweet taste preferences. In: Taste and Development: The Genesis of Sweet Preference, Ed: J.M. Weiffenbach, Bethesda, Maryland, U.S. Dept. Health, Education and Welfare (1977). Young, P.T. and Greene, J.T., Relative a c c e p t a b i l i t y of saccharin solutions as revealed by d i f f e r e n t methods. J. Comp. Physiol. Psychol., 46 (1953) 295-298. Zarevics, P. and Setler, P.E., Simultaneous rate-independent and rate-dependent assessment of i n t r a c r a n i a l s e l f -stimulation: Evidence for the d i r e c t involvement of dopamine i n brain reinforcement mechanims. Brain Res., 169 (1979) 499-522. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0096734/manifest

Comment

Related Items