Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Factors affecting long-term habituation in Caenorhabditis elegans Beck, Christine Daily O’Brien 1995

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

Item Metadata

Download

Media
831-ubc_1995-059200.pdf [ 4.47MB ]
Metadata
JSON: 831-1.0088240.json
JSON-LD: 831-1.0088240-ld.json
RDF/XML (Pretty): 831-1.0088240-rdf.xml
RDF/JSON: 831-1.0088240-rdf.json
Turtle: 831-1.0088240-turtle.txt
N-Triples: 831-1.0088240-rdf-ntriples.txt
Original Record: 831-1.0088240-source.json
Full Text
831-1.0088240-fulltext.txt
Citation
831-1.0088240.ris

Full Text

FACTORS AFFECTING LONGTERM HABiTUATION INCAENORHABDITIS ELEGANSbyCHRISTINE DAILY O’BRIEN BECKB.Sc., University of Alberta, 1987M.A., University of British Columbia, 1991A THESIS SUBMiTTED IN PARTIAL FUFILLMENT OF THEREQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of PsychologyWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJune 1995© Christine Daily O’Brien Beck, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of cJ ôLi(’ry’jThe University of BritisF( Columbia”Vancouver, CanadaDate LAAIJ2DE-6 (2188)IIAbstractThe objective of these experiments was to explore long-term memory in Caenorhabditiselegans. This examination of memory in a simple organism with accessible genetics and a wellunderstood biology may permit later work to defme the cellular processes that underlie long-term memory.Habituation training with a vibrational stimulus was administered on Day 1, and theretention test of a block of stimuli was given 24 h after the end of training on Day 2. Long-term retention of habituation was evident as a lower level of responding on Day 2 relative to thelevel of responding on Day 2 of untrained controls or the initial level of responding of wormson Day 1.In Experiments 1 and 2, a habituation training protocol that produced long-termretention of habituation was established, and the effects of stimulus number, interstimulusinterval (1ST), and distribution of training on both short-term and long-term habituation wereexamined. In Experiment 1 (10-s 1ST), there appeared to be a floor effect which resulted in alow level of responding regardless of training on Day 1; thus no evidence for long-termhabituation after training at a 10-s 1ST could be found. In Experiment 2 (60-s ISI), worms thatreceived distributed and massed habituation training with 60 stimuli showed a significantlylower level of responding relative to untrained controls. The distributed habituation trainingappeared to be more effective at inducing long-term habituation and was used in the subsequentexperiments.To characterize the effects of heat shock treatments used in the behavioral experimentsthat follow, the effects of heat shock on two assays, the induction of a heat shock protein gene,hspl6, and the rate of egg-laying were measured in Experiment 3. All heat shock treatmentsused caused the induction of hspl6. In addition, the number of eggs laid during a fixed intervalafter heat shock was sensitive to the heat shock treatments given in Experiments 4 through 8.In Experiments 4 through 8, the effects of heat shock on short- and long-termhabituation were examined. Heat shock, which acts as a general cellular stressor, wasifiadministered at different times before, during and after training. In Experiment 4, heat shock(45 mm, 32°C) was administered, ending 2 h before training on Day 1. Heat shock beforetraining did not affect the initial level of responding on Day 1, habituation during training,short-term retention of habituation between blocks of training or long-term retention ofhabituation. In Experiment 5, heat shock (45 mm, 32°C) was administered during the restperiods of distributed training in the 1-h interval after each training block. While heat shockduring training had no significant effect on responding on Day 1, long-term habituation wasblocked.In Experiment 6, the possibility that heat shock before training would prevent thedisruption of long-term habituation by heat shock during training by inducing thermal tolerancewas examined. This was tested by administering heat shock (45 mm, 32°C) that ended 2 hbefore training and heat shock during training. It was found that heat shock before training didnot prevent the disruption of long-term habituation by heat shock during training.In Experiment 7, the effect of heat shock that ended 2 h before the retention test on Day2 on the retention of long-term habituation was examined. It was found that heat shock on Day2 did not disrupt the retention of habituation.Finally, in Experiment 8, the effect of brief heat shock (15 mm, 32°C) at differentintervals in the rest period following the training blocks was examined in an attempt to morenarrowly defme a critical period for consolidation of long-term habituation. Although therewas no significant effect of brief heat shock on retention of habituation, the pattern of the datasuggests that there may be a period of greater vulnerability worth further investigation.In summary, heat shock given before training or before the retention test did not affectlong-term habituation, while heat shock during training disrupted consolidation of long-termhabituation. Taken together, these behavioral results provide the foundation for an investigationof the cellular processes underlying long-term memory in . elegans. By exploring thedynamics of the formation of long-term habituation, intervals of time critical to the formation oflong-term habituation were defmed. This in turn will help to focus attention on the cellularivprocesses whose activity during those intervals of time may be important to the consolidationof long-term memory.VTable of ContentsAbstract iiList of Tables viiList of Figures viiiAcknowledgments xINTRODUCTION 1Habituation is the Product of Multiple Processes 1Long-term Habituation as a Form of Memory 9. elegans as Simple Model System for Learning 15Long-term Habituation in . elegans 22Interference as a Tool to Defme Memory Consolidation 23Overview of Experiments 28General Methods 28Subjects and Materials 28Stimulation and Behavioral Observations 29Habituation Training Procedure 29Heat Shock 30Scoring and Statistical Analysis 31Chapter One Experiment 1: Short- and Long-term Habituation at a Short ISI (10-s ISI) 39Methods 39Results 40Chapter Two Experiment 2: Short-term and Long-term Habituation at a Long ISI(60-s ISI) 47Methods 48Results 48Chapter Three Experiment 3: The Effects of Heat Shock on hspl6 Induction andthe Rate of Egg-laying 56Experiment 3A: The Effect of the Heat Shock Treatments on hspl6Induction 56Methods 57Results 57Experiment 3B: The Effect of the Heat Shock Treatments on the Rate ofEgg-laying 60Methods 60Results 61Chapter Four Experiment 4: Short-term and Long-term Habituation with Pre-exposure toHeat Shock 64Methods 64Results 65Chapter Five Experiment 5: Short-term and Long-term Habituation with Heat ShockDuring Training 71viMethods 71Results 72Chapter Six Experiment 6: Short-term and Long-term Habituation with Pre-exposureto Heat Shock and Heat Shock During Training 78Methods 78Results 80Chapter Seven Experiment 7: The Effects of Heat Shock Just Prior to Testing on theRetention of Long-term Habituation 87Methods 87Results 91Chapter Eight Experiment 8: Titration of the Effects of Heat Shock During the Rest Period 91Methods 92Results 93Chapter Nine Synthesis of the Results: An Analysis Including Experiment 2,and Experiments 4 through 8 96Chapter Ten Discussion 106Interference as a Tool to Examine Memory Consolidation 109The Distributed Training Effect: A Psychological Perspective 114Differentiation of Types of Long-term Memory Through InterferenceTreatments 117The Use of Heat Shock as an Interference Treatment 122List of Abbreviations 123Bibliography 125Appendix 1 The Distribution of Missing Values 136vilList of TablesTable 1. The post-hoc comparisons between the LTH and LTH-HS groups inExperiments 2, 4, 5, 6, 7 and 8 of the mean block magnitude of thefirst twenty stimuli on Day 1. 106vmList of FiguresFigure 1. The nematode Caenorhabditis elegans with the apparatus usedbehavioral observations. 17Figure 2. The tap-withdrawal circuit. 20Figure 3. The countemull value and the calculation of effect size. 35Figure 4. Habituation curves at a 10-s 1ST. 42Figure 5. Habituation by block at a 10-s 1ST. 46Figure 6. Habituation curves at a 60-s 1ST. 50Figure 7. Habituation by block at a 60-s 1ST. 54Figure 8. The expression of JZ in transformed . elegans. 59Figure 9. The number of eggs laid by five worms in a 30 mm test period,15 to 45 mm after the end of the heat shock treatment. 63Figure 10. Habituation curves: distributed habituation training withpre-exposure to heat shock. 67Figure 11. Habituation by block: distributed habituation training withpre-exposure to heat shock. 70Figure 12. Habituation curves: distributed habituation training withheat shock during training. 74Figure 13. Habituation by block: distributed habituation training with heat shockduring training. 77Figure 14. Habituation curves: distributed habituation training withpre-exposure to heat shock and heat shock during training. 82Figure 15. Habituation by block: distributed habituation training with pre-exposureto heat shock and heat shock during training. 86Figure 16. Habituation by block: distributed habituation training with heat shock onDay 2, 2 hr before testing. 90Figure 17. Habituation by block: distributed habituation training alone or with briefheat shock during training either early, mid or late in the 1-h rest periodfollowing each training block. 95Figure 18. The mean Day 1 and Day 2 response levels across experiments:LTH groups. 98Figure 19. The mean Day 1 and Day 2 response levels across experiments:LTH-HS groups. 100xFigure 20. The number of missing values in Experiments 1 and 2 139Figure 21. The number of missing values in the three blocks of training on Day 1 in themassed and distributed habituation training group. 142Figure 22 The number of missing values in the twenty stimuli control group, and themassed and distributed training groups. 144Figure 23. The distribution of missing values during distributed habituation training onDay 1 and on Day 2. 146Figure 24. The distribution of missing values during massed habituation training onDay 1 and on Day 2. 148Figure 25. The number of missing values in Experiment 4: pre-heat shock. 152Figure 26. The distribution of missing values during training in the LTH and PRE HS ILTH groups on Day 1 and on Day 2. 155Figure 27. LTH-HS (Exp. 5) and LTH (Exp. 2; distributed training) and thedistribution of missing values during habituation training on Day 1. 158Figure 28. The distribution of missing values during training in the LTH (Experiment2) and LTH-HS (Experiment 5) groups on Day 1 and on Day 2. 160Figure 29. The distribution of missing values during training in the LTH, LTH-HS andthe PRE HS / LTH-HS groups on Day 1 and on Day 2. 164Figure 30. The distribution of missing values during training in the LTH and LTH /D2HS groups on Day 1 and on Day 2. 166Figure 31. The distribution of missing values during training in the LTH and LTHearly HS groups on Day 1 and on Day 2. 169Figure 32. The distribution of missing values during training in the LTH-mid HS andLTH-late HS groups on Day 1 and on Day 2. 171Figure 33. The distribution of missing values during training in all the LTH groups(Exp. 2, 4, 6, 7 and 8) on Day 1 and on Day 2. 174Figure 34. The distribution of missing values during training in both of the LTH-HSgroups (Exp. 5 and 6) on Day 1 and on Day 2. 178xAcknowledgmentsFirst, I want to acknowledge the tremendous support and encouragement I receivedfrom my advisor, Cathy Rankin, whose patience with my struggles during this time seemedlimitless. I also owe thanks to the other members of my thesis committee, Eric Eich and TonyPhiffips, who helped keep the standards of the work high.I wish to acknowledge the help of the others that were involved in the experiments Ireport here. Tracey Smillie, Richard Faber and Michael McCready, as undergraduate students,all ran projects which were part of this work. Peter Candido (Department of Biochemistry,UBC) kindly supplied me with the transgenic strain of. elegans used in Experiment 3A, andDon Jones offered his expertise in staining technique. Finally, I would like to thank SteveWicks for staining and photographing the transgenic worms in Experiment 3A.Last but not least, I would like to gratefully acknowledge the support, encouragementand patience of my family, Sean, Caitlin and Rowan— without whom, I cannot imaginehaving come this far.I would like to dedicate this thesis to my mother, Kathleen Beck, and my father,Charles Beck. Thank you both for passing your love of science on to me.1IntroductionLong-term memory may be defined as the lasting changes in behavior that are broughtabout by experience and are mediated by cellular processes in the nervous system. Studyinglong-term memory in a simple model system provides an opportunity to defme both thebehavioral patterns that indicate the operation of memory and the cellular mechanisms that mustsupport memory. The experiments presented here using the nematode Caenorhabditis elegansbegin an analysis of the behavioral changes produced by a form of long-term memory in thissimple organism.Habituation is the product of multiple processesThe focus of this work is the analysis of memory for a simple form of learning,habituation. Habituation is a ubiquitous form of learning, found in organisms ranging fromprotozoa to mammals (Harris, 1943; Thompson & Spencer, 1966; Wood, 1988) and has beendescribed in a variety of behavioral responses, such as bristle-cleaning in the fruit flyDrosophila melanogaster (Corfas & Dudai, 1989), desynchronization ofelectroencephalography (BEG) to auditory stimuli in rats (Dworkin & Dworkin, 1990), therespiratory startle response to light in goldfish (Laming & McKinney, 1990), the galvanic skinresponse to visual stimuli in humans (Barry & Sokolov, 1993), and the escape response toelectric shock in the crab (Rakitin, Tomsic & Maldonado, 1991). Habituation may be definedas a decrease in responding seen with repeated stimulation (Groves & Thompson, 1970). It isdistinguished from simple fatigue by a number of features including the expression ofdishabituation, which is the facilitation of the habituated (decremented) response by a novel ornoxious stimulus, and a sensitivity of both habituation and spontaneous recovery fromhabituation to interstimulus interval (Staddon, 1993; Groves & Thompson, 1970). Retention ofhabituation may be observed after a repeated series of habituation training and spontaneous2recovery; this training leads to progressively greater habituation (Groves & Thompson, 1970;Petrinovich, 1984).Groves and Thompson (1970) proposed that the habituation observed during a singletraining session is composed of two antagonistic processes: habituation, which is stimulus-specific, and sensitization, which affects the state of the whole organism. Basing their dual-process theory of habituation on work with the habituation of spinal reflexes in the cat, theyargued that the connection between stimulus (S) and response (R) is affected by these twoindependent processes; habituation weakens the connection specifically between the stimulusused in habituation training and its response, while sensitization facilitates all S - R connectionsby altering the state of the organism (Groves, DeMarco & Thompson, 1969; Groves,Glanzman, Patterson & Thompson, 1970; Groves, Lee & Thompson, 1969; Groves &Thompson, 1970). Generally, the habituation process starts small and builds gradually; thesensitization process starts relatively more strongly, and in some paradigms increases for ashort time before fading gradually. They proposed that the dynamics of habituation are theproduct of the interaction between the dynamics of these two processes, habituation andsensitization (Groves & Thompson, 1970). Although there is some controversy about the roleof stimulus perception in habituation (Hall, 1991), there appears to be a consensus thathabituation is the product of multiple processes (Groves & Thompson, 1970; Hall, 1991;Peeke and Petrinovich, 1984; Staddon, 1993; Barry & Sokolov, 1993).The notion of multiple behavioral processes has implications for the cellular processesunderlying the behavior. One question that must be answered is where in the nervous systemhabituation is mediated and whether different components are localized in different loci. Theprocesses underlying habituation may be highly localized, for example, in the sensory neuronsynapses (e.g. Castellucci, Pinsker, Kupfermann & Kandel, 1970) or may be distributedthroughout the nervous system (e.g. Falk, Wu, Cohen & Tang, 1993). Studies correlating3electrophysiological changes in neuronal activity have focused on this question. The efforts tocharacterize the cellular processes underlying habituation have been confounded by thecomplex relationship between the decrement in response expressed in the sensory neuronstransducing the stimuli and the behavioral habituation induced by the stimulation of the sensoryneuron.In support of a simple and direct relationship between the activity of the sensoryneurons and habituation, Hernández-Peón (1960) showed in the cat that the induction ofhabituation was accompanied by presynaptic inhibition of neurons in the periphery; heproposed that habituation results from inhibition of sensory input which may occur as early asthe first sensory relay nucleus. However, Groves, Glanzman, Patterson and Thompson(1970) showed that habituation of spinal reflexes in the cat was not due to presynapticinhibition of sensory fibers; there are no necessary changes in excitability of afferent terminalsduring the development of response habituation. This finding argues against a peripheral locusfor habituation (Groves & Thompson, 1970). In addition, Weinberger, Goodman and Kitzes(1969) did not fmd a relationship between habituation of the eye-movement response toauditory stimuli and the ascending auditory evoked potentials; although they did find asignificant relationship between reticular formation activity and eye-movement habituation,which may indicate the role of arousal state in habituation. It may be that the locus ofhabituation depends on the system that is mediating the plasticity.A traditional way of probing for the locus of processes that support learning is to lesionstructures of the brain that may mediate those processes. There are few examples of disruptionof habituation without the loss of sensory or response ability by lesions of specific neuralstructures; however, lesions of the telencephalic area blocks habituation in both the startleresponse to light stimulation of the goldfish and the attack response to visual stimuli in larvalsalamander (Laming & McKinney, 1990; Pietsch & Schneider, 1990). Pietsch and Schneider4(1990) noted that while the telencephalon was not formerly believed to play an important role invision in the salamander, it may mediate an active-negative component of the visual behavior ofthe salamander. Laming and McKinney (1990) hypothesized that in the goldfish, thetelencephalon acts as a holding center for short- and long-term memories, and may also play arole in processing temporal stimuli. In both organisms, the role of novelty in stimulus-processing appears to have been disrupted by these lesions.Likewise, in the honeybee, there is evidence for inhibition of the proboscis extensionreflex by central mechanisms (Braun & Bicker, 1992). The proboscis extension reflex is anappethive response to sugar solution applied to one of the bee’s antennae; Menzel, Hammerand Sugawa (1989) demonstrated that hunger improved the learning index in the classicalconditioning of the proboscis extension reflex. Braun and Bicker (1992) examined whetherhunger and satiation would influence habituation of the reflex. They found that both the initialresponse to the stimuli was higher and the rate of habituation was slower in hungry bees thanin satiated bees, supporting the theory that the internal state of the organism (i.e. sensitization)affects habituation. Habituation of the reflex was limited to the antenna stimulated; thecontralateral antenna did not show response decrement when tested. Dishabituation of thehabituated reflex could be evoked by stimulating the contralateral antenna. Throughpharmacological blockers of neural pathways it was determined that expression and habituationof the reflex depended upon the functioning of the antenna lobes, but not mushroom bodies,which mediate olfactory conditioning in the honeybee (Braun & Bicker, 1992).Although there appears to be no universal locus for habituation, it is possible that thecellular processes that support synaptic plasticity share characteristics across systems. Acandidate for a cellular analogue of habituation is the phenomenon of synaptic depression,which is the decrease in post-synaptic activity with a consistent level of stimulation of the presynaptic neuron (Zucker, 1989). Quantal analysis in the mollusk Aplysia californica,5demonstrated that synaptic depression of sensory neurons is accompanied by a decrease in thenumber of quanta of neurotransmitter released (Castellucci & Kandel, 1974). In addition, inthe siphon sensory neuron - motor neuron synapse, the development of synaptic depression iscorrelated with the kinetics of a long-lasting presynaptic calcium current inactivation (Klein,Shapiro & Kandel, 1980). The calcium current is responsible for presynaptic neurotransmitterrelease; with calcium current inactivation, the amount of neurotransmitter release would bediminished, resulting in synaptic depression (Klein, Shapiro & Kandel, 1980; Byrne, 1982).Further analysis demonstrated that a combination of transmitter depletion and inactivation of thepresynaptic calcium current are likely responsible for gill sensory neuron synaptic depression(Gingrich & Byrne, 1985).Interestingly, it now appears that the gill-withdrawal reflex of Aplysia is not mediatedsolely by the monosynaptic pathway of the siphon sensory neuron to motor neuron connection(Lukowiak, 1978; Hawkins, Castellucci & Kandel, 1981; Frost, Clark & Kandel, 1988); it isclear that the circuit that supports the gill-withdrawal reflex involves polysynaptic pathwayswhich may differentially modulate the response to siphon stimulation (Trudeau & Castellucci,1992). Optical measurements of neuronal activity in the abdominal ganglion show that over200 neurons are activated by the sensory stimulus that produces the gill-withdrawal reflex(Zecevic, Wu, Cohen, London, Hopp & Falk, 1989; Nakashima, Yamada, Shiono, Maeda &Satoh, 1992). In an additional study, the activity of neurons in the abdominal ganglion wascharacterized during habituation of the gill-withdrawal reflex with optical measurements (Fallc,Wu, Cohen & Tang, 1993). It was found that habituation was expressed nonuniformly byneurons in the abdominal ganglion; the neurons could be separated into classes based on theiractivity (e.g. remaining steady during habituation, decreasing with habituation, increasing withhabituation) during the formation of habituation (Falk et al., 1993). While the synapticdepression exhibited by the sensory neuron-to-motor neuron synapse may still be important to6the formation of habituation of the withdrawal response in Aplvsia, it may not completelydescribe the cellular processes underlying habituation.An additional problem with homosynaptic depression as a cellular analogue forhabituation in Aplysia is the finding that tail sensory neuron activity in the intact organismduring habituation does not seem to decrease with stimulation (Stopfer & Carew, 1994).Stopfer and Carew (1994) defmed the stimulation necessary to produce behavioral habituation,decreased activity in the tail motor neurons, and homosynaptic depression in cell culture. Whenstimulation with these parameters was applied to tail sensory neurons while theysimultaneously recorded from that sensory neuron and the motor neuron it was contacting, itactually facilitated rather than depressed the excitatory presynaptic potentials (EPSPs) of the tailsensory neuron while decreasing motor neuron activity (Stopfer & Carew, 1994). In addition,other tail sensory neurons also showed facilitation, demonstrating that the effect washeterosynaptic (Stopfer & Carew, 1994). These results indicate that homosynaptic depressionis not an adequate description of the all of the cellular processes underlying this form ofhabituation (Stopfer & Carew, 1994). Stopfer and Carew (1994) proposed that either increasedsensory neuron output to inhibitory intemeurons, or increased inhibition at intemeuronal sitesmay contribute to habituation in the tail withdrawal reflex.On the whole, work with Aplvsia has given us a tremendous amount of power toexamine plasticity in the elements of the nervous system. However, it seems from the recentevolution of understanding of the nature of the gffl-withdrawal circuit (Fallc et al., 1993), andthe fmding that homosynaptic depression may not be responsible for habituation of the tailwithdrawal response (Stopfer & Carew, 1994), that information about cell-to-cell interactionsmust be examined in the context of the whole organism to test the validity of their participationin mediating a form of behavioral plasticity.7An alternative approach to the investigation of the cellular processes underlying learningis the isolation and cloning of mutants with impaired learning. By identifying the gene productcoded by the gene affected by these mutations and defining the physiological role of the geneproduct in the nervous system, the cellular processes underlying learning may be probed(Heisenberg, 1989). This approach has been successfully used in the fruit fly Drosophilamelanogaster (Dudai, 1988; Heisenberg, 1989; Tully, Preat, Boynton & Del Vecchio, 1994).The biochemical characteristics of the classic Drosophila learning mutants dunce andrutabaga have been defined (Byers, Davis & Kiger, 1981; Dudai & Zvi, 1984; Livingstone,Sziber & Quinn, 1984). Both mutants affect the level of cAMP, an intracellular secondmessenger believed to be involved in sensitization and classical conditioning in Aplysia (Byrneet al., 1993). dunce is deficient in cAMP phosphodiesterase; thus it has reduced cAMPhydrolysis (Byers, Davis & Kiger, 1981). rutabaga is deficient in the Ca2 stimulation ofadenylate cyclase, the enzyme that controls the production of cAMP; thus rutabaga hasabnormally low levels of cAMP.These mutants were isolated through a screen for learning defective mutants in theolfactory classical conditioning paradigm; both show deficits in the retention of olfactoryconditioning (Dudai, Jan, Byers, Quinn & Benzer, 1976; Aceves-Pina et al., 1983). Theirperformance in a variety of habituation paradigms has been examined; no consistent pattern ofeffects has emerged (Heisenberg, 1989; Corfas & Dudai, 1989, 1990; Wittekind & Spatz,1988; Duerr & Quinn, 1982; O’Dell, 1994). Duerr and Quinn (1982) found that habituation ofthe proboscis extension reflex was slower and less profound in dunce and rutabaga flies than inwild-type (normal) flies. In contrast, the habituation of the landing response to visual stimuli indunce and rutabaga flies was more rapid and deeper than that seen in wild-type flies. Thehabituation of the male courtship response was more rapid and deeper in dunce and inactive, amutant strain deficient in octopamine, a neurotransmitter with a putative role in learning8(O’Dell, 1994). Clearly, there is no simple relationship between the deficits in olfactoryconditioning exhibited by the classic learning mutants and abnormal expression of habituation.In the bristle-cleaning reflex of Drosophila, the relationship between the responsedecrement seen at the cellular level and the response decrement seen at the behavioral level wasexamined. The rate and depth of habituation of the bristle-cleaning reflex in response to adeflected bristle was unchanged in dunce and rutabaga when compared to wild-type ffies;however, memory for habituation was attenuated (Corfas & Dudai, 1988). Dishabituation ofthe habituated bristle-cleaning reflex was evoked by stimulation of the dorsalcentralmicrochaetae (Corfas & Dudai, 19989). The response decrement of the sensory neuron thattransduces the bristle-bending stimulus, which might have been considered a cellular analogueof the habituation of the bristle-cleaning reflex, actually proved to be affected differently by thedunce and rutabaga mutants and was not facilitated by stimulation that produced dishabituationof the behavioral response (Corfas & Dudai, 1990). dunce showed a more rapid and deeperdegree of fatigue to very brief stimulation at a short ISI (stimulus duration 400 ms; 5-s 1ST)than wild-type; rutabaga showed slower fatigue with longer stimuli at a longer 1ST (stimulusduration 6 5; 20-s 1ST) when compared to wild-type flies (Corfas & Dudai, 1990). In addition,in dunce, the degree of fatigue was not 1ST-dependent when ISIs were varied between 1 and 2s, while in both rutabaga and wild-type flies, fatigue was deeper at the shorter 1ST.In addition, the injection of cAMP phosphodiesterase inhibitors produced a dunce-likephenotype in fatigue exhibited by wild-type flies, suggesting that dunce has its effect on fatiguethrough the deficient cAMP phosphodiesterase (Corfas & Dudai, 1990). The phenotype ofrutabaga is not affected by cAMP phosphodiesterase inhibitors, as would be expected sincerutabaga has a low level of cAMP available. Because of the differential expression ofhabituation and fatigue in these mutant strains, Corfas and Dudai (1990) concluded that thecellular fatigue exhibited by the sensory neuron does not have a simple, direct relationship9with the behavioral habituation of the bristle-cleaning reflex, and that central mechanisms maybe involved.In Drosophila, as with the other systems examined here, the relationships betweenlearning processes and neural processes are rarely simple and direct. Even so, in most casesexamined, a mutant strain that was isolated through a screen for abnormal olfactoryconditioning exhibits some form of abnormal habituation as well. The continued analysis ofthe genetics of habituation would be enhanced by precise descriptions of the learningprocesses and the neural pathways that different learning paradigms share and ways in whichthey are unlike each other. For example, the disparate effects of dunce and rutabaga on variousforms of habituation may be influenced by the state of the organism and how it affects thatparticular form of habituation.In summary, habituation is a highly significant and commonly observed form oflearning. Despite its ubiquitous nature, the cellular processes underlying habituation are not yetwell understood. It may be that investigations focused on the processes underlying long-termmemory for habituation will help to elucidate the common elements of the processes thatunderlie habituation.Long-term habituation as a form of memoryLong-term habituation has been demonstrated in a range of organisms and responses,such as the acoustic startle response and the response to spatial novelty in rats (Leaton &Supple, 1991; Cerbone & Sadile, 1994), the gill-withdrawal reflex in Aplysia (Carew, Pinsker& Kandel, 1972), the response of PC 12 cells (cultured rat pheochromocytoma cells) to ATP(Cheever & Koshland, 1992), and the proboscis extension reflex in the honeybee (Bicker &Hähnlein, 1994).Groves and Thompson (1970) define long-term habituation as the progressively greaterhabituation observed with repeated habituation training. In work with Aplysia and other10invertebrates, memory for training lasting at least 24 h is considered a characteristic of long-term memory for various types of learning including habituation (Carew, Pinsker & Kandel,1972; Carew & Kandel, 1973; Tully, Preat, Boyton & Del Vecchio, 1994), although there aresome examples of consolidated memory that last less than 12 h (Tully et aL, 1994; Tempel,Bonini, Dawson & Quinn, 1983; Ghirardi, Montarolo & Kandel, 1995). Long-termhabituation has been observed as a decrease in the magnitude of the initial stimulus of ahabituation run and as a more rapid rate of habituation during the retention test (e.g. long-termhabituation was measured as a decrease in overall responsiveness, Carew, Pinsker & Kandel,1972; increased rate of habituation, Cheever & Koshland, 1992; decrease in initial responseand overall responsiveness, Leaton & Supple, 1991; increase in the rate of habituation and adecrease in response level, Lozado, Romano & Maldonado, 1990; decrease in the number ofstimuli required for habituation, Bicker & Hahnlein, 1994; decrease in the overallresponsiveness, Cerbone & Sadile, 1994). Long-term habituation may occur in the same cellsor structures that are involved in short-term habituation or it may involve additional cells orstructures. Likewise, it may involve the same cellular processes as short-term habituation, or itmay include novel ones.Long-term habituation of the acoustic startle response in the rat appears to be mediatedby the cortex and nuclei of the cerebellum, as lesions to these areas block the formation oflong-term habituation of that response without disrupting short-term habituation, while lesionsto other areas of the cerebellum did not disrupt long-term habituation of the acoustic startleresponse (Leaton & Supple, 1991). In addition, long-term habituation of the lick suppressionresponse was not affected by the lesions to the medial cerebellum, so it appears that onlyprocesses associated with the formation of long-term memory for habituation of acoustic startleare localized there (Leaton & Supple, 1991).11Because short-term habituation of the acoustic startle reflex was unaffected by medialcerebellar lesions, Leaton and Supple (1991) suggested that the processes mediating long-termhabituation of this response are separate from the processes that mediate the response itself.However, the finding that long-term habituation of the lick-suppression response was notaffected by the lesions that disrupted long-term habituation of the acoustic startle responseindicates that there is stimulus-specificity in the processes that mediate this form of long-termhabituation (Leaton & Supple, 1991). It has also been found that lesions to the mesencephalicreticular formation block long-term habituation without disrupting short-term habituation,suggesting that this structure also plays a part in the circuit that mediates long-term memory forhabituation of acoustic startle (Jordan, 1989).As with habituation during training, the internal state of the animal may contribute tolong-term memory for habituation. Long-term habituation of the exploratory response tospatial novelty in the rat appears to be strongly influenced by factors that alter the internal stateof the animal, such as wakefulness, stimulants, tranquilizers and strain differences invasopressin levels (Cerbone & Sadile, 1994; Sadile, Cerbone & Cioffi, 1981). Stimulation ofthe medial forebrain bundle, a manipulation that enhances learning in some active and passiveavoidance paradigms (Huston, Mueller & Mondadori, 1977; Mondadori, Huston, Ornstein &Waser, 1976), disrupts the formation of long-term habituation (Cerbone & Sadile, 1994).Cerbone and Sadile (1994) suggested that long-term habituation is the product of the interactionof subcortical structures which supply excitation and cortical structures which supplyinhibition.The cellular processes underlying long-term memory may involve lasting changes in thestate of a circuit or the activity of a structure (Dudai, 1988). There is evidence for responses atthe level of the genome accompanying induction of long-term habituation of the response tospatial novelty (Cerbone & Sadile, 1994). Cycloheximide, a protein synthesis inhibitor,12blocked long-term habituation (Cerbone & Sadile, 1994). Arousal and habituation to noveltywere accompanied by distributed changes in the expression of c-fos and c-junimmunoreactivity in the brain, possibly reflecting a change in gene expression induced by thesealtered internal states (Papa, Pellicano, Welzl & Sadile, 1993). Exposure to novelty altered thedistribution of unscheduled DNA synthesis, which is thought to underlie DNA remodeling(Sadile et al., 1995). Taken together, the work on the habituation to spatial novelty in ratsprovides evidence that the internal state of the organism is capable of influencing molecularevents, and thus may help mediate long-term habituation (Cerbone & Sadile, 1994).Long-term habituation may involve a number of neural structures in most systems, butit may be the case that the cellular processes underlying long-term habituation may be containedwithin individual cells as well. Cultured, neuronally differentiated PC12 cells from the ratrespond with neurosecretory responses to the application of acetyicholine or ATP, and withrepeated stimulation will exhibit response decrement with many of the characteristics ofhabituation (McFadden & Koshland, 1990a; McFadden & Koshland, 1990b). Cheever andKoshland (1992) found that with repeated training, the rate of response decrement increased.This increase in rate was dependent on the number of stimuli received, and was retained for atleast 90 mm. It would be interesting to examine the interaction between the interstimulusinterval, and retention of response decrement as evidenced by an increased rate of habituationwith repeated training sessions. Although the plasticity expressed in this preparation cannot beviewed as a form of learning, as the sensitivity of spontaneous recovery from responsedecrement to ISI and dishabituation of the decremented response have not yet been described, itmay be a useful tool in understanding what cellular events may accompany habituation in intactsystems.The long-term memory for habituation training in Aplysia has been studied at both thebehavioral and cellular levels (e.g.. Carew et al., 1972; Montarolo, Kandel & Schacher, 1987;13Bailey & Chen, 1983, 1988a, b). Carew et al. (1972) demonstrated that Aplysia is capable oflong-term retention of habituation of two defensive withdrawal responses, the siphonwithdrawal response and the gill withdrawal response, and demonstrated that the retention oflong-term habituation lasts at least three weeks. Furthermore, they showed that long-termretention of habituation in the siphon withdrawal response was greater after distributed trainingthan after massed training. Distributed training was given over a series of four days, whilemassed training was given on a single day. Distributed training contributed to better retentionon the first test, one day later, and better retention after one week. In addition, Carew andKandel (1973) showed that distributed training with blocks of training separated by only 1.5hours rather than by entire days was also effective at producing long-term habituation.The study of long-term memory in Aplysia has led to an examination of thephysiological changes in the cells and morphological changes at the synapse (Castellucci,Carew & Kandel, 1978; Bailey & Chen, 1983). Castellucci et al. (1978) showed that theincidence of excitatory postsynaptic potentials of the sensory neuron - motor neuron synapsedecreased significantly with long-term habituation. Bailey and Chen (1983) used electronmicroscopy to survey the ultrastructure of the siphon sensory neuron synapses in animals inwhich long-term memory for habituation and sensitization had been induced. They found thatthe number and size of presynaptic active zones and the complement of vesicles increased withlong-term sensitization and decreased with long-term habituation.In view of the demonstrations that polysynaptic pathways play a role in the gill-withdrawal and the tail-withdrawal circuit, it is important to consider the possibility that long-term memory for plasticity expressed in the gill-withdrawal reflex may be mediated by cellsother than the stimulated sensory neurons. In support of this possibility is the finding that long-term homosynaptic depression may not be demonstrable in cell culture, while heterosynapticdepression was successfully induced by the application of FMFRamide (Montarolo, Kandel &14Schacher, 1987). Work has continued to characterize long-term memory for sensitization inAplysia. However, little further progress in the investigation of the processes underlying long-term habituation has occurred.In summary, a number of manipulations that affect the expression of long-termhabituation have been explored in various systems; however, there is as yet no consensus onthe loci or mechanisms of long-term habituation. In the rat, long-term habituation of theacoustic startle response can be abolished with a lesion to the medial cerebellum withoutdisrupting short-term habituation of the acoustic startle response or long-term habituation of thelick-suppression response (Leaton & Supple, 1991). Also in the rat, long-term habituation ofthe exploratory response to spatial novelty was affected by agents affecting the arousal level ofthe organism and depressing activity in the forebrain (Cerbone & Sadile, 1994). In addition,there is evidence that events at the level of the genome may be triggered by changes in theinternal state of arousal or habituation of the organism (Cerbone & Sadile, 1994). A simplemodel of habituation with cultured neuronally differentiated PC12 cells from the rat may helpelucidate the activity of individual cells during habituation and long-term habituation (Cheever& Koshland, 1994). Finally, work with Aplysia on long-term habituation of the gill-withdrawal reflex has demonstrated that uhrastructural changes at the sensory neuron synapseaccompany the induction of long-term memory for habituation (Bailey & Chen, 1983).Although a number of manipulations which block the consolidation of long-termhabituation in the preparations have been discussed above, as yet little is known about thespecific time during which memory consolidation for long-term habituation occurs. If criticalperiods during which memory consolidation was vulnerable were known, the roles of changesin gene expression or protein synthesis during those periods in the mediation of long-termhabituation could be probed.15C. elegans as a simple model system of learningThe purpose of the experiments proposed here is to examine the characteristics of aform of long-term habituation in the free-living (i.e. nonparasitic) nematode C. elegans. inorder to investigate the processes proposed to underlie learning and memory in a new systemand search for insights into these processes (see Figure 1 A). C. elegans has been successfullyused as a model for the genetic control of development (Wood, 1988), and the extensivebackground of information on the organism’s biology make it an excellent candidate for theinvestigation of cellular and molecular mechanisms of learning and memory (Rankin, Beck &Chiba, 1990). Compared to the nervous systems of other organisms studied, the nervoussystem of C. elegans is extremely simple; all 302 neurons and their cell lineages have beenidentified, and the connectivity of all the neurons is known (White et al., 1986; Hall & Russell,1991). Despite this simplicity, C. elegans has a rich behavioral repertoire and is capable of avariety of forms of behavioral plasticity (Bargmann, 1993; Rankin et al., 1990).Short-term habituation has been studied in the tap withdrawal reflex in C. elegans.The tap-withdrawal reflex is a set of behavioral responses that are evoked by vibrational stimuli(Ranicin, Beck & Chiba, 1990; see Figure 1 B and C for apparatus). This behavioral reflex ismediated by the tap-withdrawal circuit in the nervous system of C. elegans; this circuit hasbeen functionally described through lesion studies (for a diagram of the tap-withdrawal circuit,see Figure 2; Wicks & Ranicin, 1995). The tap-withdrawal circuit is composed of sevenmechanosensory neurons, nine interneurons, and a pooi of 60 motorneurons that control thelocomotion of the worm (Wicks & Rankin, 1995). The tap-withdrawal circuit can be separatedinto two functional circuits that drive forward and backward swimming (Wicks & Rankin,1995). A direct touch to the tail of the animal consistently evokes an acceleration in forwardswimming; a direct touch to the head consistently evokes a reversal response, in which the16Figure 1. A. The nematode Caenorhabditis elegans. B. The behavioral apparatus. Theworm rests on an agar-filed plate. Behavioral observations are made through thestereomicroscope. Behavioral responses are recorded on video-tape for future scoring.C. Vibrational stimuli are given by a mechanical tapper with an electromagnetic relaycontrolled by an S88 generator. (Adapted from Rankin, Beck & Chiba, 1990.)ABCpharynx17tallIntestineheadeggs0.1 mm anusGrass S-88stimulus generatortapPet r Iplate holder18worm stops, swims tail-first or backward for a distance and then turns to resume forwardswimming (Chalfie & Sulston, 1981).The tap is a vibrational stimulus that can evoke accelerations or reversals, and thus mustactivate both forward- and backward-swimming subcircuits. These two subcircuits of the tap-withdrawal circuit appear to functionally inhibit one another, so that the locomotory behavioremitted is the interaction of the activity of both subcircuits (Ranlcin, 1991; Wicks & Rankin,1995). Making use of the worm’s transparent body structure and the knowledge of the somaticcell lineage, it is possible in . elegans to lesion individual neurons through laser microsurgeryearly in development, and then test the behavior of these worms at maturity (Sulston & White,1980). When the posterior mechanosensory neurons are ablated early in development, wormsmove normally at maturity but are insensitive to direct touches to the tail (Chalfie & Sulston,1981). Worms with these neurons ablated responded to the vibrational stimulation of the tapwith only backward swimming or reversal responses (Wicks & Rankin, 1995). The reversalresponses from worms in which the posterior mechanosensory neurons were ablated weresignificantly larger in magnitude than those elicited from control worms by the same tapstimulus (Wicks & Rankin, 1995). The greater magnitude of the reversal responses in wormsthat have received lesions to the posterior mechanosensory cells indicates that with the loss ofthe acceleration response, a source of inhibition had been removed from the circuit driving thereversal response. Likewise, when the anterior mechanosensory neurons are ablated, wormsare insensitive to touch to the area behind to phaiynx (Chalfie & Sulston, 1981). In wormswith these neurons ablated, the tap stimulus evoked only accelerations (Wicks & Ranldn,1995). A computational model based on the known anatomical structure of the tap-withdrawalcircuit has been developed to confirm that the known configuration of neurons can indeedsupport this set of competing responses to tap (see Figure 2; Wicks, Roehrig & Rankin,unpublished observations).19Figure 2. The simplified tap-withdrawal circuit. The hypothesized circuit that mediates thenematode tap withdrawal reflex consists of seven sensory neurons (squares), nineinterneurons (circles), and two motoneuron pools (not shown) that produce forward andbackward locomotion (triangles). All cells represent bilateral classes of cells except AVMand DVA, which are single cells. Chemical connections are represented by arrows; thewidth of the arrows is proportional to the number of synaptic contacts as determined inelectron microscopic studies of the . elegans nervous system (White et al., 1986). Gapjunctions are indicated by dashed lines. Putative excitatory (green) and inhibitory (red)chemical connections are indicated, based on anatomical work mentioned above and acomputational model of the circuit (Wicks, Roehrig & Rankin, unpublished observations).This circuit has been simplified for ease of presentation in two ways. First, the bilateralsymmetry of the circuit has been collapsed, and second, only connections with an averageof greater than five synapses are shown. (Adapted from Wicks and Rankin, 1995.)The Tap Withdrawal Circuit20_____________Chcmica— -- — — --Electrical__________-lflhibitoryExcitatory21The competitive relationship between the acceleration response and reversal response totap brings one to the question of whether the analysis of just the reversal response, as in theseexperiments, is an adequate assay of habituation of the tap-withdrawal reflex. To answer thisquestion, it is important to understand the relationship between the competing responses duringhabituation. In normal worms, accelerations do not occur at a high enough frequency duringhabituation to make an analysis of the habituation of the magnitude of accelerations useful(Rankin, Beck & Chiba, 1990). However, through habituation studies of worms that havebeen subject to laser ablations of mechanosensory neurons such that they emit only reversalresponses or only acceleration responses, the habituation of the two subcircuits may beanalyzed independently (Wicks & Rankin, unpublished observations). The two competingresponses of the tap withdrawal circuit do, in fact, exhibit different dynamics duringhabituation. In worms with only posterior touch receptors, the acceleration response did exhibithabituation, but very slowly and with evidence of an early facilitation at short ISIs. In wormswith only anterior touch receptors, the reversal response habituated more slowly than inunaltered controls. When summated, the dynamics of these two competitive responsesapproximate the dynamics of habituation exhibited in the reversal response to tap in normalworms (Wicks & Rankin, unpublished observations). Thus, an analysis of the habituation ofthe reversal response, the predominant response during habituation training in normal worms,actually reflects the habituation of both responses to the tap-withdrawal reflex.Habituation of the reversal response in C. elegans exhibits many of the characteristicsof habituation described by Groves and Thompson (1970), including dishabituation,spontaneous recovery, and sensitivity to interstimulus interval (1ST). A major finding in ourwork is that interstimulus interval has important effects on the dynamics of a number ofcomponents of habituation, affecting the rate of habituation, depth of habituation, and the rateof spontaneous recovery from habituation (Broster & Rankin, 1994; Rankin & Broster, 1992).22Long ISIs (e.g. 60-s 1ST) lead to slower, shallower response decrement than short ISIs (e.g.10-s ISI), but also lead to slower recovery from habituation (Rankin & Broster, 1992). Theslower recovery exhibited after training with a long ISI suggests that short-term retention ofhabituation is better with longer ISIs.Long-term habituation in C. elegansIn addition to short-term habituation, Rankin et al. (1990) showed that habituationtraining could be retained and affect behavior up to 24 hours after the training, a signfficantlength of time in an organism whose reproductive cycle is three days. This fmdingdemonstrated that.elegans is capable of a form of long-term memory, long-term habituation.To demonstrate long-term habituation, Rankin et al. (1990) used a distributed trainingprocedure in which the experimental group received 100 stimuli in five blocks of 20 stimuli at a10-s 151 separated by periods of rest on Day 1, while the control group received only one blockof 20 stimuli. A test block of 20 stimuli was given on Day 2. The Day 2 performance of eachgroup was compared to its Day 1 performance; the experimental group, receiving 100 stimuli,showed a significant decrease from Day 1 to Day 2 while the control group, receiving 20stimuli, did not. This decrease on Day 2 exhibited by the experimental group was consideredevidence of long-term habituation. Attempts to replicate these results using a variety ofprocedures have met with mixed success — in some experiments long-term habituation wasseen, and in others it was not (Rankin, unpublished observations). The objectives ofExperiments 1 and 2 were to establish a procedure that reliably produces long-termhabituation, and to examine the role of several stimulus parameters (stimulus number, stimulusgrouping and interstimulus interval) in long-term habituation in C. elegans.23Interference as a tool to define memory consolidationOnce a habituation training protocol that produces long-term retention of habituation hasbeen established, the next step in the investigation of this form of LTM is to attempt to perturbthe formation and retention of LTM so that the dynamics of these processes can be examined.Work with both vertebrates and invertebrates has suggested that protein synthesis is necessaryfor the cellular processes of LTM (Flood, Bennett, Orme & Jarvik, 1977; Mizumori,Rosenzweig, & Bennett, 1985; Frey, Krug, Reymann & Matthies, 1988; Davis & Squire,1984; Montarolo, et al., 1986). Thus, protein synthesis inhibitors have been used to probe forcritical periods in memory formation.The investigation of the time course of the cellular processes underlying formation oflong-term memory for sensitization in Aplysia began with the demonstration that proteinsynthesis inhibitors administered just before training and lasting for 1 to 3 hr after trainingblock both long-term sensitization of the gill-withdrawal response induced by repeated electricshocks to the tail in a reduced preparation and long-term facilitation in cultured cells byinfusions of 5-HT (Castellucci, Blumenfeld, Goelet & Kandel, 1989; Montarolo et al., 1986).The structural changes at the synapse associated with sensitization, such as increases in thenumber of presynaptic active zones, the size of each zone and the vesicle complement of eachzone (Bailey and Chen, 1983; 1988a, b), are also blocked by transcriptional and translationalinhibitors in cell culture (Schacher, Montarolo, Kandel, Chen & Bailey, 1991), strengtheningthe argument that the protein synthesis inhibitors may be directly affecting memory formation.The results from experiments on memory and protein synthesis inhibitors with Aplysiahave led investigators to examine the cellular processes that are altered during the period of timewhen protein synthesis inhibitors are effective at blocking the induction of long-termsensitization or its analogues (Castellucci, Kennedy, Kandel & Goelet, 1988; Barzilai,Kennedy, Sweatt & Kandel, 1989; Eskin, Garcia & Byrne, 1989). This work has led to the24identification and classification of a number of proteins, the synthesis of which alters duringthe induction of long-term sensitization or its analogues; these proteins have been classified bytheir time of induction as early, intermediate and late proteins (Barzilai et al., 1989; Castellucciet al., 1988; Kuhi, Kennedy, Barzilai & Kandel, 1992). Thus, the work with Aplvsia on theeffect of protein synthesis blockers on LTM has led to the investigation of the roles of specificproteins in the formation of memory.Although in Aplysia there appears to be clear evidence for the blockage of long-termsensitization with protein synthesis blockers, not all areas of invertebrate research support thehypothesis that protein synthesis is always necessary for LTM. In the Drosophila olfactoryclassical conditioning paradigm, a number of treatments have been used to distinguish betweendifferent types of memory. Anesthetic in the form of cold shock (4 °C for 1 to 2 mm) within30 mm of training disrupted memory, but did not disrupt memory if administered after 2 h(Tempel, Bonini, Dawson & Quinn, 1983; Tully et al., 1994). However, cycloheximideadministered before training such that protein synthesis was inhibited by 90 to 95% in the brainduring and immediately after training failed to block the formation of anesthesia-resistantmemory (lasting about 4 days) for associative training, but blocked long-term memory (lastingover 7 days) from distributed training but not massed training (Tully et al., 1994). In addition,Wittstock, Kaatz and Menzel (1993) showed that when protein synthesis is inhibited by morethan 99.5% in honeybees by cycloheximide, there was still no impairment of the one-trial long-term olfactory learning exhibited by bees — support for the notion that there may be multiplemechanisms for LTM, some of which require protein synthesis, others of which do not.Olfactory learning resists disruption by protein synthesis inhibitors in vertebrate systems aswell (Staubli, Faraday & Lynch, 1985). Clearly, the issue of the role of protein synthesis andgene activation in learning is far from resolved.25One problem that has interfered with the interpretation of the results from studiesexamining the blocking of memory formation with protein synthesis inhibitors is that manyagents that affect protein management have systemic effects; the changes in memory formationobserved after treatment with the agent may be a result of these effects, not a specific failure toform LTM (Davis, Rosenzweig, Bennett & Squire, 1980; Flexner, Flexner, & Roberts, 1966;Squire & Barondes, 1974; Squire, Geller & Jarvik, 1970; Gold, 1989). In work withvertebrates, attempts have been made to disassociate the systemic effects of the proteinsynthesis inhibitors such as depletion of constitutive proteins in the nervous system, sickness,lethargy, etc., from the specific ability of protein synthesis inhibitors to inhibit de novo proteinsynthesis at the time of training (Davis, Rosenzweig, Bennett & Squire, 1980; Squire,Emanuel, Davis & Deutch, 1975; Pilcher & Booth, 1975; Segal, Squire & Barondes, 1971;Squire & Barondes, 1974). In work with Aplysia, this problem is partly dealt with by usingreduced preparations or cell culture to test long-term sensitization or its analogues (Byrne,Zwartjes, Homayouni, Critz & Eskin, 1993); but with these approaches, behavioralinformation from the whole organism is not accessible. In addition, the studies with Aplsia onthe role of protein synthesis focus on sensitization. Sensitization involves the enhancement ofa response (Groves & Thompson, 1970); thus, a treatment which blocks long-termsensitization decreases the responsiveness of the organism. Although protein synthesisblockers permit short-term sensitization while preventing long-term sensitization (Castellucci etal., 1989), it is possible to imagine that the history of the treatment may have a long-term effectof depressing the organism’s responsiveness, and thus its ability to express long-termsensitization, by impeding its overall functioning. One advantage of using habituation as alearning paradigm in this context is that the successful block of LTM for habituation by anagent would result in a higher Day 2 level of responding relative to trained, untreated controls,while systemic effects from the treatment impairing the ability of the organism to function26normally should result in a lower Day 2 level of responding. Thus, in a habituation paradigm,it is less likely that the systemic effects of the treatment will confound a positive effect on theretention of training.The agent used to disrupt LTM in the present experiments is heat shock (32°C). Thecellular response to heat shock (body temperature elevated at least 2 to 5 °C above the optimallevel) was first observed in Drosophila (Schlesinger, Tissieres & Ashbumer, 1982) and hassince been described in every system examined, from bacteria to humans (Schlesinger,Tissieres & Ashburner, 1982; Lindquist, 1986; Nowak, 1993). The cellular response to heatshock is also evoked by other types of cellular stress such ischemia, glucose deprivation,toxins, cold, viral infections and other damaging agents (Lindquist, 1986). Heat and theseother disruptive agents cause damage in the form of the denaturation and agglutination ofproteins. In response, cells stop protein synthesis of all proteins except for a class of proteinscalled heat shock proteins (HSPs), the production of which are massively increased. TheseHSPs are the most widely conserved families of proteins described and appear to play a role indamage-control (Schlesinger, Tissieres & Ashburner, 1982; Lindquist, 1986). They areclassified into three loosely grouped families by molecular weight (the HSP2Os or lowmolecular weight HSPs, the HSP7Os or intermediate molecular weight HSPs, and the HSP9Osor high molecular weight lISPs). Some HSPs are present constitutively and are not inducedby cellular stress, others are present constitutively and are also induced by cellular stress, andstill others are not present constitutively and are only induced by cellular stress (Feige & Polla,1994; Schlesinger, Tissieres & Ashburner, 1982; Lindquist, 1986). In a cell, the presence oflast kind of HSPs, the strictly inducible HSPs, is an indicator of cellular stress (Nowak, 1994;Sharp & Sagar, 1994; Koroshetz & Bonventre, 1994; Stringham, Dixon, Jones & Candido,1992; Stringham & Candido, 1993).27The optimal temperature range for growth and development in C. elegans is 15° to 25°C (Wood, 1988). C. elegans reacts with the production of HSPs to temperatures over 29° C(Snutch & Baillie, 1983). In C. elegans, the production of HSPs occurs in all cell typesincluding neurons, demonstrating that all tissues are affected by the stress of the heat shock(Stringham et a!., 1992). Heat shock proteins belonging to the HSP2O, HSP70 and HSP9Ofamilies are expressed in C. elegans (Snutch & Baihie, 1983; Heschl & Baillie, 1990; Russnak& Candido, 1985; Dailey & Golomb, 1992). As little as 15 mm of exposure to 33°C results inthe production of HSPs (Stringham et al., 1992). Yet after strong heat shock (33°C, for twoshifts of 2 hr each), the worms are still viable; after heat shock they continue to reproduce andrespond normally to tactile stimuli (Beck, unpublished observations). When observed within 2mm after heat shock (45 mm, 32°C), worms are eating and moving in a coordinated fashion;they respond normally to tactile and vibrational stimuli, but their spontaneous behavior remainsslightly lethargic for 3 to 5 mm (Beck, unpublished observations). For this reason, in theexperiments presented here, an interval of 10 mm at room temperature was allowed after heatshock before habituation training resumed.The key characteristic of heat shock as an agent to disrupt LTM formation is itsnonspecificity, as it both causes damage directly to cellular proteins through denaturation; and itevokes an active response from the affected cells, in the form of a reduction of general proteinsynthesis accompanied by the stress-induced production of HSPs (Lindquist, 1986; Morimoto,Tissieres & Georgopoulos, 1990). The number of cellular processes that might be affected byone or more of these effects of heat shock is great. This breadth of effect limits the narrownessof the assertions that may be made about the cellular processes underlying LTM based on theresults of this study, but greatly increases the chance of disrupting some essential process.Importantly, heat shock can be delivered for short, defined periods of time, which permits anexamination of the effects of a disrupting stimulus lasting as little as 15 min. Thus, using heat28shock, it is possible to examine the effects of a potentially disrupting treatment on LTM notonly before, during and after training, but also to examine critical periods of memory formationmore closely by shifting short duration treatments through them systematically.An additional advantage of heat shock is that its effect may be measured in more than itspossible behavioral consequences. Heat shock causes the induction of heat shock proteins; thisinduction may be demonstrated in transgenic worms with a reporter gene after a heat shockpromoter so that induction of heat shock proteins results in the induction of that reporter gene(Stringham et al., 1992). Heat shock affects the rate of egg-laying; this is a sensitive assay tothe severity of the heat shock. Thus, a profile of the consequences of heat shock on theorganism and on the consolidation of long-term memory may be developed.Overview of ExperimentsIn Experiments 1 and 2, a protocol that produces long-term habituation was defined byexamining factors affecting LTM for habituation in . elegans such as stimulus number,interstimulus interval, and distribution of training. In Experiments 3, the effect of heat shockon the induction of the heat shock gene hspl6 and the rate of egg-laying was examined. InExperiments 4 through 8, the effects of heat shock before, during and after habituation trainingon immediate habituation, short-term retention of habituation and long-term retention ofhabituation were investigated with the objective of defining a critical period for theconsolidation of LTM for habituation.General MethodsSubjects and Materials£. elegans Bristol (N2) were maintained in 4 cm diameter Petri plates filled with 10 mlNematode Growth Medium agar. The data from a total of 399 worms were included in theanalyses of these experiments. The nematodes fed on E i (0P50), as described in Brenner(1974). At four days post-hatching (the peak of egg-laying), subjects were individually placed29on labeled agar-filled plates at least two hours before training and were maintained on the sameplates throughout the procedure. Worms were transferred individually from the breeding plateto the labeled plates using a bent wire pick; simultaneously, small amounts of bacteria sufficientto feed each subject during the study were seeded on each plate.Stimulation and Behavioral ObservationsObservations were made through a stereomicroscope with attached videorecordingequipment (see Figure 1B; Wild M3Z, Wild Zeiss Canada; Panasonic Digital 5100 camera;Panasonic AG 1960 VCR; NEC monitor). A time-date generator was used to superimpose theexperimental time, time of day and date on the video record (Panasonic WJ-810). Thevibrational stimulus used in this work was a 6 Hz train of six taps delivered to the side of thePetri plate holding the subject (see Figure 1C). Each tap exerted approximately 1 - 2 N of forceon the plate. The stimuli were produced by a mechanical tapper with an electromagnetic relaytriggered by a Grass S88 stimulus generator controlled by the experimenter.The dependent measure used in this work was a measure of the magnitude of thereversal response to tap, in which the worm, either lying still or swimming forward, respondsto a vibrational stimulus by initiating tail-first or backward swimming for a distance. Thedistance traveled during the reversal response can be quantified using stop-frame video analysisand computer-driven digitizing equipment (Macintosh computer; Bit Pad Plus digitizing tablet;Macmeasure software).Habituation Training ProcedureIn Experiments 1 and 2, worms were trained using one of four different trainingprocedures on Day 1 (4 days post-hatching) and tested on Day 2 (5 days post-hatching). Thefour Day 1 training groups were: (a) a distributed training group that received 60 stimuli inthree blocks of twenty stimuli with one hour rest periods between blocks of training, (b) amassed training group that received 60 stimuli together, (c) a twenty stimuli control group that30received 20 stimuli together, and (d) a single stimulus control group that received only onestimulus.The twenty stimuli control group was included to examine the effects of a smallnumber of stimuli during training on long-term habituation. The single stimulus control group,that received no habituation training on Day 1, was included to control for the effects of ageand handling. On Day 2, at least 24 hours (24 to 28 hours) after the end of training on Day 1,all subjects received 20 stimuli. The performance on Day 2 of the single stimulus controlgroup was compared with the performance on Day 1 and on Day 2 of any group that showed alower level of responding on Day 2. Long-term habituation was evident when there was adecrease in the level of responding from Day 1 to Day 2 in a trained group and at the same timethe Day 2 performance of the untrained single stimulus group was comparable to the Day 1performance of the trained groups.The results from Experiments 1 and 2 indicated that habituation at a 60-s 1ST with 60stimuli was the most effective at producing LTM for habituation, and that with an inspection ofFigure 4, distributed habituation training with 60 stimuli appeared slightly more effective thanmassed habituation training. Therefore, in Experiments 4 through 8, which involved heatshock, the 60-s 1ST distributed habituation protocol and the single stimulus control protocolwere used.Heat shockHeat shock (32°+I- 1°C), used in Experiments 4 through 8, was delivered byimmersing the plate sealed with Parafilm in which the worm rested in a temperature controlledwater bath. When measured with a probe thermometer, the temperature of the agar when a platewas submerged in the water bath reached 32 ° C in about 2 mm. When removed from the bathand returned to a room-temperature environment, the temperture of the agar returned to roomtemperature in about 30 s. Tn different experiments, heat shock was given before training,31during training or after training on Day 2. When heat shock was delivered on Day 1 duringtraining, a minimum of 10 mm was allowed after the end of heat shock exposure beforetraining was resumed.In Experiment 3, the effects of the heat shock treatments used in the subsequentbehavioral experiments were examined in two contexts. In Experiment 3A the induction ofhspl6, the gene of a low molecular-weight heat shock protein HSP16, by the heat shocktreatments used in the subsequent experiments was tested. In Experiment 3B, the effect ofthese heat shock treatments on the rate of egg-laying, a sensitive and easily quantifiable assay,was tested.Scoring and statistical analysesThe statistical package, SPSS (Version 4.0.5 Macintosh), was used for these analyses.Planned comparisons were done by hand.General analyses. In general, ANOVAs were employed to test for differences betweenthe means of groups and different stages of habituation (see below for a description of theapplication of ANOVAs to the analysis of short- and long-term habituation). The alpha levelwas set at .05. The error rate of each analysis was controlled by dividing the alpha level by thenumber of comparisons within the analysis; for example, in the examination of long-termhabituation in Experiments 1 and 2, the asymmetry of the design and experimental hypothesesmade the application of two ANOVAs, one at each level of day of training, appropriate in theanalysis of long-term habituation. The alpha level of these ANOVAs was reduced: .05 /2 =.025.Between-subject, within-subject and mixed-design ANOVAs were employed asappropriate. When a repeated-measures design was used, Mauchly’s test for sphericity, whichexamines the validity of the assumption of sphericity in the repeated-measures data, was done(Lindman, 1991). When a significant result from this test indicated that the assumption of the32homogeneity of variance and covariance was violated, the degrees of freedom of thecorresponding E-test were adjusted downward using the Huynh-Feldt epsilon generated bysPss.Planned comparisons focusing on contrasts that have significance in the context of thestudy were employed when the overall ANOVA reached significance. When multiple plannedcomparisons were used in one analysis, the alpha level was reduced by dividing by the numberof comparisons (Lindman, 1991).As planned comparisons are not robust to violations of the homogeneity of variance,Bartlett’s test of homogeneity of variance was applied across between-subjects factors(Lindman, 1991). When the test indicated that the assumption of homogeneity was violated,the error term of the E ratio for that planned comparison, which would nonnally be the errorterm representing the variance from the entire analysis, was replaced with the error termreflecting only the groups being compared (Lindman, 1991). if, in a repeated measuresdesign, Mauchly’s test of sphericity indicated that the assumption of the homogeneity ofvariance and covariance was violated, planned comparisons were not employed, as plannedcomparisons are not robust to violations of this assumption. The results of Bartlett’s test of thehomogeneity of variance and Mauchly’s test for sphericity are only reported here when they aresignificant.The measurement of effect size. An estimate of effect size (ES) which is independentof the units of measurement used in the dependent measure may be calculated from theinformation in an ANOVA table with the following formula:effect size (ES) ((MSeffect I fl) / MS error) 1/2As can be seen, the effect size has a direct relationship with the value of the f test.F = (MSeffect / (MS error) = ES2 (fl)33The countemull value of the observed mean and the interpretation of statisticalsignificance. When the contrast of two means is not statistically significant (for example, ifalpha is set at .05, and p> .05), the common practice is to infer that the experimentalmanipulation had no effect at all on the group receiving treatment. Likewise, when the contrastbetween two means is significantly different, the common practice is to assume that thestatistical significance automatically signals scientific importance (Rosenthal & Rubin, 1994).Both of these assumptions may be invalid in some situations. A consideration of thecounternull value of the observed mean, a statistic which corresponds to the value of the meanfor which there is as much evidence as the mean of the null hypothesis, but on the oppositeside of the observed mean, may help clarify these situations (Rosenthal & Rubin, 1994).In Figure 3A, the distribution of the mean under the null hypothesis is shown. Thevalue of the obtained mean corresponds to the probability of finding that mean on thedistribution of the expected mean, which is determined by the null hypothesis. However, thereis just as much evidence (the same probability) for a mean with twice the effect size of theobserved mean, assuming the expected effect size of the mean of the null hypothesis is zero.As can be seen in Figure 3B, the observed mean corresponds to the same probability regardinga distribution the mean of which is symmetrical and in the opposite direction from the mean ofthe null hypothesis (Rosenthal & Rubin, 1994). The mean of this symmetric but oppositedistribution is the counternull value of the observed mean (Rosenthal & Rubin, 1994).A consideration of the countemull value of the observed mean may be helpful in twosituations. First, when the observed mean is not found to be statistically significant, thecounternull value, for which there is as much evidence as the expected mean of the nullhypothesis, may be considered; if the effect size of the countemull value would be significantthere is no justification for saying conclusively that the experimental manipulation had nj effecton dependent measure. In other words, considering the countemull value of the observed34Figure 3. The counternull value of the obtained effect size. In Fig. 3A, the distributioncorresponding to the null hypothesis with the expected mean is shown; the probability ofthe obtained mean is the area under the curve away from the center of the distribution(shaded area). In Fig. 3B, the obtained mean has the same probability regarding thedistribution of the counternull mean; again the probability of the obtained mean is the areaunder the curve away from the center of the distribution.35X expected X obtainedX expected X obtained X counternullFor example:ES expected = 0ES obtained = .68ES counternull = 2 (ES obtained- ES expected)= 2 (.68)- 0= 1.3636mean may prevent the misinterpretation of a failure to observe significance as incontrovertibleevidence for a lack of a relationship between the experimental manipulation and the result(Rosenthal & Rubin, 1994).A second situation in which the counternull value of the observed mean may be helpfulin the interpretation of statistical results, is in judging the scientific significance of a differencebetween the expected mean and the observed mean. The counternull mean corresponds to thevalue of mean when the effect size is doubled (see Figure 3B). If the difference between theobserved mean and the counternull mean is not big enough to be of scientific importance in thecontext of the work, then a statistically significant difference between the expected mean andthe observed mean probably does not correspond to a scientifically important difference either(Rosenthal & Rubin, 1994).The countemull value is not a test of significance, but rather an aid to the interpretationof statistical results and their importance in the context of the work (Rosenthal & Rubin, 1994).The counternull value of an obtained effect size will be considered here when the interpretationof a fmding is in doubt or is of special theoretical significance.Scoring of responses. Reversal responses were quantified by measuring the distancetraveled backward in response to the tap stimulus. A pause in forward swimming was scoredas a reversal response of zero, as was no response to the tap stimulus. Increases in forwardswimming immediately following the tap (accelerations), less than 30% of the responses, wereconsidered missing data since the magnitude of accelerations cannot be directly compared withthe magnitude of reversal responses. In Appendix 1, the significance of the distribution ofmissing values to the interpretation of the habituation data is explored in Experiments 1 and 2and 4 through 8.Treatment of data. Differences in performance during habituation training wereexamined through analyses of the response magnitudes averaged over blocks of either five or37twenty stimuli. Obtaining these means decreased some the variability exhibited in these data,and helped deal with the missing data points. The means of responses were calculated for everysubject and then used in the analyses. The comparisons of the mean block magnitudes wereused in the analyses of habituation, short-term retention of habituation and long-term retentionof habituation.Analysis of short-term habituation. Short-term habituation was examined duringhabituation training on Day 1. The mean of five responses is more sensitive to short-termchanges in the levels of responding than the mean of twenty responses, so it was used as thedependent measure. The effects of different types of training on habituation during training onDay 1 were compared across the trained groups in Experiments 1 and 2 with a mixed-designANOVA; the effect of training was measured by comparing the mean of the first five responsesto the mean of the last five responses. A significant decrease from the initial response level tothe final response level on Day 1 was considered evidence of short-term habituation.The nature of the distributed habituation training protocols used in Experiments 1 and 2made it appropriate to expand the analysis of short-term habituation of the distributed traininggroups. As this approach was used in the analyses of Experiments 4 through 6 as well asExperiments 1 and 2, it will be described here. In distributed training, stimuli are delivered inthree blocks of twenty stimuli with rest intervals between them. A two-way repeated-measuresANOVA with Block (3 levels: blocks 1, 2 and 3) and Training (2 levels: the initial mean of fiveresponses, INIT, and the final mean of five responses, HAB, in each block) was employed.Thus, habituation within blocks (a significant main effect of Training, in which the mean of theinitial five responses is higher than the mean of the fmal five responses), the short-termretention of habituation between blocks (a significant main effect of Block, in which the levelsof responding decreased across blocks), and differences in the habituation expressed in eachblock (a significant Training x Block interaction) were examined through this analysis.38Because the expression of spontaneous recovery from habituation was important in the contextof the work, a planned comparison was employed to contrast the mean of last five responses ofBlock 1 with the mean of the first five responses of Block 2 and likewise with Blocks 2 and 3.In Experiments 4, 6,7 and 8, the distributed training procedure was used in more thanone group; in Experiments 4 and 6, the expression of short-term habituation was of particularinterest. In these experiments, short-term habituation was compared across groups with amixed-design ANOVA (Block x Training x Group); again spontaneous recovery fromhabituation between blocks of training was examined with planned comparisons.Analysis of long-term habituation. Long-term retention of habituation was defmed as asignificant decrease in the level of responding from Day 1 to Day 2. In the analysis of long-term memory, the mean of twenty responses was used as the dependent measure as it is lessvariable than the mean of five responses used in the analysis of short-term habituation. Foreach worm, the mean of the responses to the first twenty stimuli on Day 1 and the twentystimuli on Day 2 were calculated. In Experiments 1, 2, 5 and 7, control groups that receivedone stimulus on Day 1 were included to control age, treatment, or handling effects. Worms inthese groups received habituation training only on Day 2, thus the mean of twenty responseson Day 2 was calculated for these subjects. In Experiments 1, 2, 5 and 7, two ANOVAs wereemployed at the different levels of day (Day 1 and Day 2).In Experiments 4, 6 and 8, no untrained control group was included in the experimentaldesigns. In these experiments, long-term habituation was analyzed across groups with amixed-design ANOVA (Day x Group). The level of responding on Day 1 and Day 2 werecontrasted using planned comparisons.Experimental design affected the particular planned comparisons done on the data fromDay 2 in all experiments. These planned comparisons are discussed in the methods and resultssections of each experiment.39Experiment 1Short- and long-term habituation at a short ISI (10-s 151)To clarify some of the inconsistent findings from a variety of 10-s 1ST long-termhabituation experiments with different handling and training procedures (Rankin, unpublishedobservations), both short and long-term habituation after training at a 10-s 1ST with massed anddistributed procedures was examined.MethodsSubjects and Materials. Worms were maintained as described in the general methods.Twenty-one subjects were run in each of four conditions for a total of 84 subjects, and the datafrom 83 subjects were used in the analyses (one subject was dropped because of a variety oftechnical errors such as failure to record trials or equipment difficulties). The use of theapparatus, behavioral observations and scoring of behavioral responses were performed asdescribed in the general methods section.Procedure. The Day 1 training procedures for the four groups are described in thegeneral methods (distributed training, ), 21; massed training, j], = 21; twenty stimuli controltraining, j, = 21; and single stimulus control, n = 20. Testing was 24 - 28 hr later on Day 1.Statistics. The analysis of short-term habituation and long-term habituation wasperformed in the same manner for Experiment 1 (10-s ISI) and Experiment 2 (60-s ISI). Theanalysis of short-term habituation, using means of the first five responses and last fiveresponses as the measure, was performed as described in the general methods. A two-way,mixed design ANOVA (Training x Group; alpha = .025) compared the initial response leveland final response levels of the trained groups. A two-way, repeated-measures ANOVA (Blockx Training; alpha = .025) compared the initial and final level of responding across blocks oftraining in the distributed training procedure. Mauchly’s test of sphericity indicated that theassumption of homogeneity of variances and covariances was violated (Mauchly’s sphericity40test, chi-square approximation = 7.17, = .03), so planned comparisons examining recoveryfrom habituation were not performed, and the degrees of freedom of the relevant E tests wereadjusted with the Huynh-Feldt epsilon.Long-term habituation, using means of the first twenty stimuli on Day 1 and the twentystimuli on Day 2, was analyzed with pair of ANOVAs (each with an alpha = .025), one oneach level of Day, and were followed with planned comparisons designed to test specifichypotheses. A planned comparison was used to compare the Day 2 response level of singlestimulus control subjects to the average response level of the trained subjects on Day 1 (alpha =.025). This contrast was done to examine whether the untrained subjects of the single stimuluscontrol group were responding on Day 2 at the same level as the other groups at the beginningof training on Day 1 as would be expected if age and handling did not affect the responding ofthe subjects of the single stimulus group.To examine the effects of training, the average of the Day 2 level of responding of thedistributed and massed training groups was compared to the Day 2 level of responding of thesingle stimulus controls (alpha = .0 1). Long-term habituation may be expressed as asignificantly lower response level on Day 2 in the distributed and massed group than the singlestimulus control group. To examine the effects of the different training procedures, the Day 2response of the distributed and massed training groups were contrasted with each other (alpha= .01).ResultsHabituation 10-s 1ST. Short-term habituation was examined across groups comparingthe mean of the initial five responses with the mean of the final five responses on Day 1.Distributed training (Fig. 4A), massed training (Fig. 4B) and twenty stimuli control training(Fig. 4C) all produced significant short-term habituation on Day 1 (mixed-design Group xTraining ANOVA (alpha = .05 I 2 = .025): Training: E(1, 55) = 173.85, < .01). The41Figure 4. Habituation curves at a 10-s 1ST. The mean response magnitude (mm) to eachstimulus given during training on Day 1 and testing on Day 2 is shown for subjectsreceiving distributed training (Fig. 4A; j = 21; 60 stimuli in three blocks of 20 stimuli withone hour rests between blocks), massed training (Fig. 4B; n = 21; 60 stimuliconsecutively), and twenty stimuli control training (Fig. 4C; n = 21; 20 stimuliconsecutively).42DAY1 DAY2A3-u-I-—cnE 2.5-ZEb i.::1•0.50 20 21 40 41 60 0 20STIMULIB3-LLJ25-ZE -1•STIMULIC3-Ui—øE 2.5ZE2-1.:- L1-W 0.5-0-ST I MU LI43groups did not differ from each other in this measure of the expression of short-termhabituation (Group: E(2, 55) = .02, n& Group x Training: E(2, 55) .41, a).However, it is worth noting that the lack of difference observed between the threeprocedures may have been produced by a floor effect, as for all groups the mean blockmagnitude was close to zero by the end of the training runs (mean block magnitude (in mm) +1-SE to the last five stimuli in distributed training: 0.20 +1- .07; massed training: .09 +1- .03,twenty stimulus control: 0.20 +I .05).To examine habituation duing distributed training on Day 1, the means of the initialfive responses and final five responses of each block were compared in a two-way repeated-measure Block x Training ANOVA (alpha = .025). There was significant habituation withinblocks overall (Training: (1, 12) = 9.39, = .01), and this habituation did not varysignificantly between blocks (Training x Block: E( 2, 24) = .64, n). Interestingly there wasevidence for short-term retention of habituation; the level of responding within blocks exhibiteda significant decrease during training on Day 1 (Block: E(1.46, 17.5) = 5.16, = .02; degreesof freedom adjusted with the Huynh-Feldt epsilon). As the Mauchly test for sphericity(Training x Block: chi-square approximation = 7.17, = .03; Huynh-Feldt epsilon = .73)indicated that the assumption of the homogeneity of variances and covariances had beenviolated, planned comparisons examining recovery from habituation between blocks of trainingwere not carried out. However, an inspection of Figure 4A indicates that there appeared to berecovery from habituation between Blocks 1 and 2 and Blocks 2 and 3.Long-term retention of habituation 10-s 1ST. Retention of habituation training shouldbe evident as a lower level of responding in trained groups than the untrained control group onDay 2, while the level of responding on Day 1 of the trained groups should be the same as thelevel of responding on Day 2 of the untrained control group. Two factorial ANOVAs acrossgroups on the level of response as measured by the mean of the first twenty responses on Day441 and the twenty responses on Day 2, including the single stimulus control group werecalculated, one at each level of Day (Day I and Day 2; alpha .05 / 2 = .025).No effect of group was evident on Day 1 or Day 2 (Day 1: E(3, 79) = 2.39, n; Day2: E(3, 79) = 2.55, n), although the counternull values of the effect sizes were significantlydifferent from the null effect size (Day 1, Group: obtained ES = .34, counternull ES = .68,E(3, 79) = 9.59, < .01; Day 2, Group: obtained ES = .35, counternull ES .70, (3, 79) =10.2, p < .01) indicating that it should not be concluded that the training has no effect onresponse levels on Day 1 and Day 2. While the lack of a significant difference between groupswas expected on Day 1, the failure to see differences between the groups on Day 2 was not, aslong-term habituation might be apparent as a difference between the Day 2 level of respondingamong groups. These results supply no evidence of long-term retention of habituation aftertraining at a 10-s 1ST.A particular concern in the ANOVA on the Day 1 response levels of the trained groupsand the Day 2 response level of the single stimulus control group was whether the Day 2response level of the single stimulus control group was the same as the average response levelof the trained groups on Day 1. A planned comparison contrasted the average Day 1 responselevel of the trained groups (distributed and massed training groups and the twenty stimulicontrol group) with the Day 2 response level of the single stimulus control group (alpha =.025). It was found that the difference between the trained groups and the single stimuluscontrol was not significant (E(1, 79) = 4.34, i&). although the counternull value of the effectsize was significant (obtained ES: .46; counternull ES: .91, F(1, 79) = 17.18, < .01). Aninspection of Figure 5 suggests that the Day 2 response level of the single stimulus group maybe depressed when compared to the average of the trained groups on Day 1. Based on theseresults, it cannot be concluded that age and handling did depress response levels on Day 245Figure 5. Habituation by block at a 10-s ISI. Fig. 5 shows the mean block magnitude(mm) for Day 1 and Day 2 of each of the three groups that received training on Day 1: thedistributed training group (a = 21), massed training group ( = 21), and the twenty stimulicontrol group (a = 21), and the Day 2 mean block magnitude (mm) of the single stimuluscontrol group ( = 20; error bars show +I SE).U.’(oC’1C.)0-J4632.521.510.50JDAY1B DAY 2DISTRIB MASSED TWENTY SINGLE.47independent of training, but on the other hand, it cannot be concluded that age and handling didnQi affect Day 2 response levels of the single stimulus group.From the results, it might be concluded that habituation training at a 10-s 1ST did notinduce long-term habituation. However, an inspection of Figure 5, in which the Day 1 andDay 2 mean block magnitudes for the trained groups are compared with the Day 2 mean blockmagnitude of the single stimulus control group, suggests two possible confounding factors: a)an age and handling effect, or b) a floor effect. Age and handling effects may have depressedresponse levels on Day 2 in all groups, preventing the expression of long-term habituation.This would explain the failure to find a significant difference between the trained and untrainedDay 2 levels of responding. Alternatively, the 10-s 151 used in these procedures may producesuch strong response decrement that it is not possible to exhibit a further decrease. Both ofthese factors may contribute to produce the pattern of results seen here. Therefore, at a 10-s151, there was no conclusive evidence that habituation training with 60 stimuli producedsignificant long-term habituation even though short-term habituation in all trained groups andretention of habituation between blocks of training on Day 1 in the distributed training groupwas evident.Experiment 2Short- and long-term habituation at a long 151 (60-s 151)As discussed earlier, Rankin and Broster (1992) showed that habituation training at along 1ST (60 s) resulted in slower spontaneous recovery from habituation than training at ashort ISI (10 s). This observation led to the hypothesis that habituation training at a longer 1STsuch as 60 s might also produce better long-term retention of the training than habituation at a10-s 1ST. In addition, the shallower habituation resulting from training at a 60-s ISI may helpprevent a floor effect. Here, this hypothesis is tested by examining short and long-termhabituation at a 60-s 1ST.48MethodsSubjects and Materials. Worms were maintained as described in the general methods.Twenty-one worms were run in each of four conditions for a total of 84 worms (4 worms weredropped because of a variety of technical errors so that the final total was 80 worms). The useof the apparatus, behavioral observations and scoring of behavioral responses were performedas described in the general methods.Procedure. The Day 1 training procedures for the four groups described in Experiment1 were employed in Experiment 2 as well (distributed training, j = 21; massed training, = 20;twenty stimuli control training, n = 20; and single stimulus control, n = 19). Testing was atleast 24 hours later on Day 2. The only procedural difference between Experiments 1 and 2was that in Experiment 2 habituation training and testing were given at a 60-s ISI instead of a10-s 1ST.Statistics. The statistical analysis was performed as described in the general methodsand the statistics section of the methods of Experiment 1. In the analysis of short-termhabituation, two planned comparisons (alpha = .01) were employed to examine recovery fromhabituation during the two intervals of the distributed procedure between Blocks 1 and 2 andBlocks 2 and 3 by contrasting the mean of the last five responses of one block with the mean ofthe first five responses of the next block.ResultsHabituation 60-s 151. At the 60-s 1ST, short-term habituation was again produced by allthree training procedures. Comparisons of the response magnitudes of the first and last stimulion Day 1 in the trained groups show that the distributed (Fig. 6A), massed (Fig. 6B) andtwenty stimuli control (Fig. 6C) procedures produced a significant decrease in mean responsemagnitude during training on Day 1 (mixed-design Group x Training ANOVA; alpha = .025;Training: F(l, 56) = 81.64, <.01). While the group did not affect the overall level of49Figure 6. Habituation curves at a 60 s 1ST. The mean response magnitude (mm) to eachstimulus given during training on Day 1 and testing on Day 2 is shown for subjectsreceiving distributed training (Fig. 6A; n = 21; 60 stimuli in three blocks of 20 stimuli withone hour rests between blocks), massed training (Fig. 6B; n = 20; 60 stimuliconsecutively), and twenty stimuli control training (Fig. 6C;n = 20; 20 stimuliconsecutively).50DAY1 —DAY2—A3.cn 2 2.5’O 2’J1’LU< 0.5’0’ I0 20 21 40 41 60 0 20STIMULIB3.cnEZ 2 2.5o0.50• I I0 20 40 60 0 20STIMULIC3.cn 2 2.5ZE1’<011.1< 0.5’0’ I0 20 0 20STIMULI51responding (Group: (2, 56) = 1.96, n), the groups did differ in the way they exhibitedhabituation (Group x Training: E(2, 56) = 5.46, < .01). In Figure 6, it seems clear thathabituation with fewer stimuli (the twenty stimuli control group Fig. 3C) led to less profoundhabituation than massed or distributed training with 60 stimuli (X +1- SE in mm; distributedtraining HAB: .68 +1- .11; massed training HAB: .599 +1- .09; twenty stimuli control HAB:1.25 +1- .0 1). The habituation during distributed training was considered separately. Overall,training produced habituation during the blocks of training (two repeated-measures Block xTraining ANOVA, alpha = .025; Training: E(1, 19) = 40.79, < .01). There were nosignificant differences in the way habituation was expressed within blocks (Block x Training:E(2, 38) = .7, j. However, there was evidence of short-term retention of habituationbetween the blocks (Block: E(2, 38) = 9.99, <.01).Spontaneous recoveiy between the blocks of training, a significant increase in responsemagnitude from the end of one block to the beginning of the next, was examined by plannedcomparisons of Block 1 HAB with Block 2 NIT, and Block 2 HAB with Block 3 INIT (alpha= .025 I 2 = .01). The recovery between Block 1 and Block 2 was not significant (E(1, 38) =3.39, n) although the value of the counternull effect size suggests that some recovery mayhave occurred (obtained ES = .40, counternull ES = .80, E(1, 38) = 13.57, < .01).However, there was evidence of recovery from habituation between Blocks 2 and 3 (E( 1, 38)= 14.89, < .01).Comparisons of habituation on Day 1: lOs- and 60-s 151. The analyses of short-termhabituation during training at 10-s and 60-s ISIs showed many similarities, though differenceswere also apparent. At both ISIs, there was habituation on Day 1 in all trained groups;however, at a 60-s ISI, the depth of habituation was affected by the type of training, while at a10-s 151, it was not. This difference between ISIs is likely to be related to the rapid andprofound habituation exhibited during habituation training at a 10-s 151.52As noted in Rankin and Broster (1992), the depth of habituation after training at a short1ST is greater that after a long 1ST. In the present experiments, it is clear from a comparison ofFigures 4 (short-term habituation at a 10-s 1ST) and 6 (short-term habituation at a 60-s 1ST) thathabituation training at a short 1ST produced greater response decrement than did habituationtraining at a long 1ST with the same number of stimuli and the same procedures. This rapiddecrement seen during training at a 10-s 1ST may make training at a 10-s ISI insensitive toparticular training procedures. However, at both the 10-s and 60-s ISIs, there was evidence ofretention of habituation between blocks training in the distributed, which would not beexpected if the habituation produced by training at a 10-s 151 were completely insensitive totraining.Long-term habituation 60-s 151. As can be seen in Figure 7, the single stimulus controlgroup had a level of responding on Day 2 comparable to the levels of responding on Day 1 ofthe trained subjects (alpha = .025; E(3, 76) = 1.46, j. In addition, when the levels ofresponding on Day 2 were compared, the training received on Day 1 significantly affectedresponding (alpha = .025; E(3, 76) = 3.45, = .02). From an examination of Figure 7, it isapparent that distributed or massed training groups may have expressed long-term habituation,while training with twenty stimuli did not affect response levels on Day 2. In a plannedcomparison, the average of the Day 2 response levels of the two groups that received 60 stimuliduring training, the distributed and massed training groups, was contrasted with the Day 2response level of the single stimulus group. Subjects receiving distributed and massed trainingshowed evidence of long-term habituation (alpha = .025 /2 = .01; (1, 76) = 9,195,.005). An inspection of Figure 7 indicates that distributed training may be slightly moreeffective than massed training at inducing long-term habituation. However, as seen in a secondplanned comparison, the response level on Day 2 after distributed training was not significantly53Figure 7. Habituation by block at a 60-s ISI. The mean block magnitude on Day 1 and Day2 of the three trained groups, distributed (= 21; 60 stimuli in three blocks of 20 stimuliwith one hour rests between blocks), massed (n = 20; 60 stimuli consecutively) and twentystimuli (11=20) training and on Day 2 of the single stimulus control group (fl= 19; 20stimuli consecutively) are shown (error bars show +1- SE).Cl) -I Cl) C,) m -I m z -II43L20STIMULIBLOCKMAGNITUDE(mm)p CII1%)j43CflC)C,) z G) r mL155different from the Day 2 response level on Day 2 after massed training (alpha = .01; E(1,76).72, n).Age and handling effects did not appear to contribute to the decrease in responding seenin the distributed training group on Day 2. Therefore, the decrease in responding in the 60-s1ST seen in the distributed and massed training groups can be attributed to the long-termretention of habituation training received on Day 1.Comparisons of lone-term habituation 10 s- and 60-s ISIs. At a 10-s 1ST thereappeared to be age and handling effects that depressed Day 2 response levels regardless oftraining; in addition with habituation at a 10-s 1ST, the depth of the habituation may result in afloor effect thus preventing the expression of long-term habituation (Figure 5). These effectswere not evident at a 60-s 1ST (Figure 7). Long-term habituation was expressed after training ata 60-s 1ST; no evidence for long-term habituation was seen after training at a 10-s 1ST.Although this failure to express long-term habituation after training at a 10-s 151 cannot beinterpreted as a failure to remember training, it is clear that habituation training at 60-s 151would be a better procedure to use in the examination the processes underlying long-termhabituation.In conclusion, these experiments have shown that number of stimuli, 1ST, anddistribution of training all influence the expression of long-term memory in.elegans. Thedistributed and massed habituation training procedures at a 60-s 1ST, taken together, wereeffective at producing LTM for habituation. As the selection of one procedure was required forthe subsequent experiments, the distributed training procedure, which appeared from aninspection of Figure 7 to be slightly more effective than the massed training procedure atinducing long-term habituation, was chosen to be used in the experiments on the effects of heatshock on LTM that follow.56Experiment 3The effects of heat shock on hspl6 induction and rate of egg-layingIn the series of experiments that follow (Experiments 4 through 8), the behavioralconsequences of heat shock treatments at different times in relation to training will beexamined. Here, in Experiment 3, the effects of the heat-shock treatments used in thosebehavioral experiments on the induction of a heat-shock protein gene, hspl6, and on the rate ofegg-laying, a sensitive measure of organism-wide stress, will be characterized.Experiment 3A: The effect of the heat shock treatments on hsp 16 inductionIt is possible to insert a segment of DNA into an organism’s genome such that when aspecific gene is induced, the inserted strand is also transcribed and the resulting protein issynthesized. This type of transformation was performed to mark the transcription of a gene fora low molecular weight HSP (HSPI6) in . elegans (Stringham, Dixon, Jones & Candido,1992). The reporter gene in this strain is iZ, the induction of which leads to the expressionof B-galactosidase. The induction of hspl6 is tightly controlled by heat shock; induction doesnot occur constitutively, making this gene an excellent marker for the effects of heat shock.After heat shock, the worms may be fixed and stained for B-galactosidase expression. Theexpression of hsp 16 as reported by the expression of 8-galactosidase may then be examined insitu.While the expression of 8-galactosidase after induction of hspl6 cannot be used as astrict quantitative assay of heat shock protein production, there are still relative differences instaining that relate to treatment intensity (Stringham, Dixon, Jones & Candido, 1992). Thisstrain was used to examine the relative effectiveness of the different heat shock treatments usedin the behavioral experiments that follow.57MethodsSubjects. This transgenic strain, hsp16-1Z (48.1C), was obtained from the Candidolaboratory (Stringham et al., 1992). At the time of treatment, worms were 3 to 4 days old (atthe peak of egg-laying). Twenty to thirty worms were placed on plates together. Five plateswere prepared.Apparatus. Heat shock was delivered by immersing the plates (sealed with Paraflim) ina temperature controlled water bath at 32 ° C. Fixing and staining of the worms was done asdescribed by Stringham et al. (1992) with a histochemical stain containing X-gal (Fire, White-Harrison & Dixon, 1990).Procedure. The effects of four heat shock treatments on the expression of Bgalactosidase were examined as well as a no-heat shock control. The four heat shock treatmentswere: a) single heat shock (45 mm, 32°C), b) three heat shocks (45 mm, 32 °C) given at thesame intervals as are used in Experiments 5 and 6, starting 1 h, 20 mm apart, c) single heatshock (15 mm, 32°C), d) three heat shocks (15 mm, 32° C) given at the same intervals as the45 mm heat shocks. After the end of heat shock treatment, the worms were allowed to rest for30 mm to permit the 8-galactosidase to develop (Stringham, et al., 1992). The animals werethen fixed and stained as described by Stringham et al. (1992).Analysis. The analysis was a qualitative judgment of staining intensity. Photographs ofrepresentative staining after each of the treatments are presented here. A positive heat shockcontrol (same strain, heat shock at 33°C, for 30 min) performed by the Candido laboratory isalso presented for comparison.Results. As can be seen in Figure 8, there was staining after all heat shock treatments,but none in the no-heat shock control group. Overall, it appears there was a dose-effect instaining intensity expressed between the 15 mm and 45 mm heat shock. Interestingly, it seems58Figure 8. The expression of JZ in transformed C. elegans. In this strain, JZ has beeninserted behind the promotor of hsp 16. The heat shock treatments were the ones used inthe behavioral experiments (Experiments 4 through 8). No-heat shock controls. Positiveheat shock controls, 33°C, 30 mm (Candido laboratory). Single heat shock (15 mm, 32°C).Three heat shocks (15 mm, 32°C) at 1 h 20 mm intervals. Single heat shock (45 mm,32°C). Three heat shocks (45 mm, 32°C) at I h 20 mm intervals.597../NEGATIV (30 MIN, 33°C)SINGLEHS (15 MIN, 32°C)rHS (15 MIN, 32°C)SINGLEHS (45 MIN, 32°C)I THREEHS (45 MIN, 32°C)4r.60that staining after three heat shocks (45 mm, 32°C) was no more intense than the staining aftera single heat shock (45 mm, 32°C; see Fig. 8E and F).These results clearly indicate that the heat shock treatments used in Experiments 4through 8 caused cellular stress marked by the induction of hspl6, a gene for HSP16, a lowmolecular weight heat shock protein. The effects of these heat shock treatments on the rate ofegg-laying was examined next.Experiment 3B. The effects of the heat shock treatments on the rate of egg-layingThe rate of egg-laying in . elegans is under optimal conditions, fairly regular: oneworm at the peak of egg-laying, 4 days of age, lays about 8 to 12 eggs per hour. This rate issensitive to conditions in the environment; particularly to the absence of food, but also toovercrowding, tactile stimulation and the presence of contaminants. The sensitivity of egg-laying to heat shock was examined by heat shocking worms then counting the number of eggsfive worms laid in a 1/2 h interval, 15 to 45 min after the end of the heat shock.MethodsSubjects. The strain used in the behavioral experiments, N2 (Bristol) was maintained asdescribed in the general methods. The age of the worms was synchronized by laying gravidadults on a plate with a lawn of. iii for 2.5 to 3 h, then removing the adults, but leaving theeggs laid during that time. The worms were handled in the same way regardless of the heatshock treatment received; all worms received the same number of transfers from plate to plate atthe same times relative to the time of the egg test.Apparatus. Heat shock was delivered as described above in the general methodssection. The apparatus described in the general methods used for behavioral observations wasused to count the eggs laid during the test interval, after the adults had been removed. Theassay for the number of eggs was done on plates streaked with 0.1 ml of an . jj culture 24 hpreviously and incubated at 20°C.61Procedure. Three heat shock treatments and a no-heat shock control were tested. Theheat shock treatments were: a) single heat shock (45 mm, 32°C), b) three heat shocks (45 mm,32°C) at the intervals described above, and c) three heat shocks (15 mm, 32°C) at the intervalsdescribed above. Egg-laying was assayed by placing five worms on a plate 15 to 45 mm afterthe end of the heat shock treatment. The worms were left on the plate for 30 mm and thenremoved. The number of eggs laid was then counted. The egg-laying after each treatment wasassessed with five plates ( (plates) = 20).Analysis. The mean number of eggs laid by five worms in 30 mm was comparedacross groups using nonparametric statistics (Kruskal-Wallis one-way ANOVA forindependent samples; Siegel, 1956).Results. As can be seen in Figure 9, there was a significant difference in the number ofeggs laid between groups receiving different heat shock treatments. It appears as though thesingle 45 mm heat shock produced an intermediate depression in egg-laying, while the tripleheat shock (both the 15 mm and the 45 mm) caused a more severe decrease in egg-laying. Thisis interesting in light of the results from the staining which failed to show much change inintensity between single and triple heat shocks of either 45 mm or 15 mm duration. It is notsurprising that heat shock affects different aspects of the physiology of the worm differently.However, these results do reinforce the necessity of examining a number of aspects ofhabituation in the experiments that follow; based on what was seen here, it is certainly possibleto imagine that one aspect of learning or memory would be affected by one parameter of heatshock while another aspect of learning might be more sensitive to a different parameter of heatshock.Taken together, the results of Experiments 3A and B indicate that the heat shocktreatments used in these experiments have consequences for the worm; while these treatments62Figure 9. The number of eggs laid by five worms in a 30 mm test period, 15 to 45 mmafter the end of the heat shock treatment (5 worms per plate, 5 plates per group). Thecontrol group received no heat shock.63zCcø0>-0w50-40-30-20-10-0-CONTROL SINGLE TRIPLE TRIPLE45M, 32°C 45M, 32°C 15M, 32°CHEAT SHOCK HEAT SHOCK HEAT SHOCK64do not affect its ability to move or respond to tactile stimuli (Beck, unpublished observations),they do cause the induction of hspl6 and they do alter the rate of egg-laying.Experiment 4Short- and long-term habituation with pre-exposure to heat shockIn this experiment, the effects the pre-exposure to heat shock on the expression ofhabituation and long-term habituation are examined. The presence of the HSPs themselvesmay affect the later expression of short- or long-term habituation. In addition, a history ofcellular stress may have consequences for learning and memory processes. To test thesepossibilities, heat shock was administered before training, and the habituation expressed duringtraining and testing was compared with that of similarly trained controls that did not receiveheat shock.MethodsSubjects and Materials. Worms were maintained as described in the general methods.Twenty subjects were run in each of two groups for a total of 40 subjects. The administrationof heat shock (45 mm, 32°C), the use of apparatus, behavioral observations and scoring ofbehavioral responses were performed as described in the general methods.Procedure. The first group, LTH (n = 20), received habituation training on Day 1 andwas tested on Day 2 as described in the general methods. The second group, PRE HS I LTH(n = 20), received a heat shock (45 mm, 32°C) that ended 2 hr before the beginning of thebehavioral training, and then received the same training and testing as the LTH group.Statistics. Short- and long-term habituation was analyzed as outlined in the generalmethods. In the analysis of short-term habituation, habituation during distributed training wascompared across groups with a three-way, mixed design ANOVA (Block x Training x Group;alpha = .05). Mauchly’s test of sphericity indicated that the assumption of homogeneity ofvariances and covariances was violated (Mauchly’s sphericity test, chi-square approximation =657.54, p = .02), so planned comparisons examining recovery from habituation were notperformed, and the degrees of freedom of the relevant E tests were adjusted with the HuynhFeldt epsilon.Long-term habituation was assessed with a mixed-design ANOVA (Day x Group; alpha= .05). Expression of long-term habituation may be seen as a significantly lower level ofresponding on Day 2 than on Day 1 across both groups, or as a significant interaction betweenDay and Group. To examine whether any interaction was the product of a difference the Day 2level of responding, a planned comparison contrasting the Day 1 level of responding of the twogroups was performed (alpha = .05).ResultsHabituation. Overall, the LTH (Fig. 1OA) and PRE HS / LTH (Fig. lOB) groupsshowed significant habituation during blocks of training on Day 1 (mixed-design Training xBlock x Group ANOVA; alpha = .05; Training: F(0.9, 34.2) = 125.56, p < .01; Mauchly’stest of sphericity for the effect of Block: chi-square approximation = 7.54, p = .02; HuynhFeldt epsilon = .90).There were no overall differences in the level of responding on Day 1 between thegroup that received heat shock (45 mm, 32°C) before training (PRE HS I LTH) and the groupthat did not (LTH) (Group: f(1, 38) = .64, n&). Taken together, subjects in both groupsshowed short-term retention of habituation between the blocks of training (Block: E(1.8, 68.4)= 48.5, p < .01) and there was no evidence that pre-exposure to heat shock affected theexpression of short-term habituation (Group x Block: (1.8, 68.4) = .06, j, In addition,the pre-exposure to heat shock did not appear to affect the expression of habituation withinblocks of training (Group x Training: E(1, 38) = .01, n..&). Interestingly, when the groupswere taken together, subjects habituated differently within the three blocks of training on Day 1(Block x Training: E(2, 76) = 4.89, p = .01). The pre-exposure to heat shock did not appear66Figure 10. Habituation curves: distributed habituation training with pre-exposure to heatshock (45 mm, 32°C). The mean response magnitude (mm) to each stimulus given duringtraining on Day 1 and testing on Day 2, 24 hr after the end of training, is shown forsubjects receiving distributed habituation training only (LTH; Fig. 1OA; n 20; 60 stimuliin three blocks of 20 stimuli at a 60 s ISI with one hour rests between blocks), and subjectsreceiving distributed habituation training 2 hr after the end of a 45 mm heat shock (PRE HS/LTH; Fig. 10B;=20).67DAY1 —DAY2—A 3, 3——. 2.5fç 2.5IZE0—uJZZ40uJ41 20 21 40 41 60 1 20B 3-,w 2.54QnI LI,g. 1 2______ ______ ______ZZ IW40 L91-I_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _051 20 21 40 41 60 1 20STIMULI68to affect this interaction either (Group x Block x Training: (2, 76) = 1.26, j&) although thecountemull effect size was significantly different from zero, so it cannot be assumed that preexposure to heat shock had no effect on the Block x Traiiiing interaction (obtained ES = .25;counternull ES = .50, (2, 76) = 5.05, p < .0 1). An inspection of Figure 10 suggests that theamount of habituation may be diminishing with each successive block of training. This may bea result of a floor effect of response decrement or it may reflect a limitation of the learningprocedure.Recovery from habituation could not be evaluated with planned comparisons asMauchly’s test of sphericity indicated that the assumption of the homogeneity of variances andcovariances was violated). However, an inspection of Figure 10, comparing the amounts ofresponse increment during the 1-h rest period following habituation, suggests that recoverymight be observed between blocks 2 and 3 but not between blocks 1 and 2.Long-term retention of habituation after pre-exposure to heat shock. The Day 1response levels of the two groups were contrasted with a planned comparison to determinewhether the pre-exposure to heat-shock had affected initial response levels; there was nosignificant difference between the groups (see Fig. 10; alpha = .05; F(l, 38) = 2.49, jj.)although the significance of the counternull effect size indicates that there may be an effect ofpre-exposure to heat shock on initial response levels (obtained ES: .35, counternull ES: .70,E(1, 38) = 9.96, p < .0 1). Retention of habituation training should be evident as a lower levelof responding on Day 2 than on Day 1. Taken together, the worms in the LTH and PRE HS ILTH groups showed significant long-term habituation (see Fig. 11; mixed design Day x GroupANOVA; alpha = .05; Day: E(1, 38) = 37.65, p < .0 1). Pre-exposure to heat shock did notaffect the expression of long-term habituation (Day x Group: F(1, 38) = .16, t).69Figure 11. Habituation by block: distributed habituation training with pre-exposure to heatshock (45 mm, 32°C). The mean block magnitudes on Day 1 and Day 2 of the two groups(LTH, distributed training only, = 20, and PRE HS I LTH, distributed training with preexposure to heat shock, j = 20) are shown (error bars show +1- SE).20STIMULIBLOCKMAGNITUDE(mm)PMTC1-I= -D•EJr3 -rn cj)r4)-‘-L.(1lII71Thus, it seems that neither short-term nor long-term habituation were affected by thepre-exposure to heat shock (45 mm, 32°C) that ended 2 h before training. It may be that heatshock given during training would be more effective at blocking the consolidation of memory.Experiment 5Short- and long-term habituation with heat shock during trainingPrevious work with agents that disrupt LTM has showed that memory formation wasvulnerable to disruption during and immediately after training (Castdllucci et aL, 1989; Frey etal., 1988). Here, the effects of heat shock delivered during each of the 1-h rest periodsimmediately after each training block on the expression of habituation, and the retention ofshort- and long-term habituation are examined.MethodsSubjects and Materials. Worms were maintained as described in the general methods.Twenty subjects were run in each of two groups. The data from 4 subjects were not includedbecause of technical errors in stimulus administration; the fmal total was N = 36 subjects. Theadministration of heat shock (45 min, 32°C), the use of apparatus, behavioral observations andscoring of behavioral responses were performed as described in the general methods.Procedure. The first group, LTH-HS (ii = 18), received habituation training asdescribed in the general methods. Immediately after each of the three training blocks during the1-hr rest periods, worms were exposed to a heat shock. A minimum of 10 mm passed after theheat shock exposure was finished before training began again. Testing was given on Day 2 asdescribed in the general methods. The second group, HS ONLY (jj = 18), received only asingle stimulus on Day 1, but received the heat-shock exposures at the same intervals as theLTH-HS group. Testing was given on Day 2 as described in the general methods. Thiscontrol group was included to examine the effects of the heat-shock treatment on performanceon Day 2.72ResultsHabituation. As can be seen in Figure 12A by comparing the response magnitudes tothe initial and fmal stimuli of training, the LTH-HS group exhibited habituation during theblocks of training (two-way repeated-measures Training x Block ANOVA; alpha = .05;Training: F(.64, 9.6) 49.41, < .01; Mauchly’s test of sphericity chi-square approximation= 13.84, < .01; Huynh-Feldt epsilon = .64).The worms receiving heat shock during training showed a significant decrease in thelevel of responding between blocks (Blocks: E(1.28, 19.2) = 18.23, < .0 1). However, therewere no differences between blocks in the habituation exhibited within blocks (Training xBlocks: E(2, 30) = .38, n&).Because Maucffly’s test of sphericity indicated that the assumption of the homogeneityof variances and covariances was violated, the planned comparisons to describe recovery fromhabituation between blocks of training cannot be performed. However, an inspection of Figure12A suggests that spontaneous recovery might be evident between Blocks 2 and 3 but notbetween Blocks 1 and 2.The effects of heat shock (45 mm, 32°C) during training on responding on Day 1 arenot yet clear. It may be, for example, that heat shock has caused an accumulating depression inthe level of responding which has been masked by the expected development of habituation.Thus we cannot be sure that heat shock is not having an effect on responding; however,because the dynamics of habituation on Day 1 are similar to that seen on Day 1 in thedistributed training group of Experiment 2 (see Figure 6A), it is clear that short-termhabituation has not been blocked by heat shock.Long-term retention of habituation. In this experimental design, retention ofhabituation is seen as a significantly lower Day 2 level of responding of a trained group relativeto an untrained control. Two contrasts using two-tailed, unpaired I tests were made: the first73Figure 12. Habituation curves: distributed habituation training with heat shock (45 mm,32°C) during training. The mean response magnitude (mm) to each stimulus given duringtraining on Day 1 and testing on Day 2, 24 hr after the end of training, is shown forsubjects receiving distributed habituation training (60 stimuli in three blocks of 20 stimuli ata 60 s 1ST with one hour rests between blocks) with heat shock (45 mm, 32°C) after each ofthe three training blocks (LTH-HS; Fig. 12A; n = 18), and subjects receiving heat shock(45 mm, 32°C) at the same intervals as the LTH-HS subjects, but without habituationtraining (HS ONLY; Fig. 12B; = 18).74DAY1 —DAY2A3.2.5zEO 2o.uJ1.5ZZ 140w4 05•0•B0 E 1I12.5z Eo bI N2t4CIW’Ll 0 1.5uJ DI°L piovI<0 I- I-. I’iDuJ420 21 40= F-41 60 1 201STIMULI75was between Day 2 of the heat shock alone group and Day 1 of the trained, heat shocked groupto examine the effects of heat shock on responding on Day 2. The second contrast wasbetween Day 2 of the heat shock alone group and the Day 2 of the trained group to examine theeffect of heat shock during training on long-term habituation. As there were two contrasts,alpha was reduced: .05/2 = .025.The effects of heat shock on habituation on Day 2 were tested by the heat-shock alonegroup. The comparison of the heat-shock alone group Day 2 level of responding with the Day1 responding of the trained group (before heat shock) will demonstrate if heat shock mayfacilitate or depress responding by itself. An unpaired, two-tailed test was used to contrast theDay 2 level of responding of the heat shock alone group with the Day 1 level of responding ofthe LTh-HS group. It was found that the response level on Day 2 of the HS ONLY group wasnot significantly different from the initial response level on Day 1 in the LTH-HS group (i(34)= 1.44, n) although the significance of the counternull effect size indicates that it cannot beconcluded that heat shock alone has no effect on responding on Day 2 (counternull effect size: I(34) = 2.88, < .01). Heat shock during training appeared to block the expression ofhabituation training (LTH-HS Day 2 vs. HS ONLY Day 2; unpaired t test, t(34) = .96,An inspection of Figure 13 indicates that heat shock during training did not affect responseslevels on Day 2 when given alone on Day 1, but that the induction of long-term habituation wasblocked by heat shock given during training on Day 1.While the expression of habituation during training and the short-term retention ofhabituation was not blocked by heat shock during training, long-term retention for habituationwas lost. One possible explanation for the failure to exhibit long-term habituation after heatshock during training is that the LTH-HS happened to have an unusually low initial level ofresponding on Day 1, making it unlikely that significant long-term habituation would beexpressed. As can be seen by comparing Figures 7 and 13, in Experiment 2 (habituation at a76Figure 13. Habituation by block: distributed habituation training with heat shock (45 mm,32°C) during training. The mean block magnitudes on Day 1 and Day 2 of the two groupsare shown (LTH-HS, heat shock during distributed training, j = 18, and HS ONLY, heatshock without training on Day 1, = 18, error bars show +1- SE).20STIMULIBLOCKMAGNITUDE(mm)p-‘0Cli..L%);.nIIIIC*)r -I Cl) Cl) 0 z rh)-‘7860-s ISI) the Day 1 response level of the distributed training group was comparable to the Day1 response level of the LTh-HS in this experiment (Exp. 2, distributed training group, Day 1(X +1- SE; in mm): 1.422 +1- .132; Exp. 5, LTH-HS, Day 1: 1.5437 +1- .149). Thus, thisdoes not seem like a plausible explanation.Experiment 6Short- and long-term habituation with pre-exposure to heat shockand heat shock during trainingOne of the earliest and best documented phenomena related to the heat shock responseis the thermal tolerance effect (Lindquist, 1986). When given a severe heat shock, cells inculture died. However, if the cell culture was given a pre-treatment of a mild heat shock, thengiven the severe heat shock, the cells survived. In some systems, the accumulation of HSPsfrom the early mild heat shock is responsible for the thermal tolerance of the severe heat shock(Lindquist, 1986; Parsell, Taulien & Lindquist, 1993). In this context, the accumulation ofHSPs from pre-exposure to heat shock may protect the formation of LTM for habituationtraining from disruption of heat shock during training. In this experiment, the effects of preexposure to heat shock on the disruption of LTM formation by heat shock during training wereexamined.MethodsSubjects and Materials. Worms were maintained as described in the general methods.Twenty subjects were run in each of two groups for a total of 40 subjects. The administrationof heat shock (45 min, 32°C), the use of apparatus, behavioral observations and scoring ofbehavioral responses were performed as described in the general methods.Procedure. The first group, PRE HS I LTH-HS (n = 20), received a heat shock thatended 2 h before the beginning of habituation training on Day 1. During habituation training,heat shock was presented after each of the three training blocks during the 1-h rest periods.79Testing on Day 2 was as described in the general methods. The second group, LTH-HS (jj =20), was a replication of the disrupting effects of heat shock during treatment observed inExperiment 5. As in Experiment 5, the LTH-HS group received habituation training with heatshock (45 mm, 32°C) after each of the three training blocks during the 1-hr rest periods.Statistics. The analysis of short- and long-term habituation was performed as outlinedin the general methods. Habituation on Day 1 was assessed with a three-way, mixed designANOVA (Block x Training x Group; alpha = .05). Recovery from habituation was examinedby comparing the mean of the last five responses at the end of one block with the mean of thefirst five responses of the next block.Three planned comparisons were used to assess recovery from habituation betweenblocks, and to examine the effects of pre-exposure to heat shock and heat shock during trainingon the degree of recovery expressed (alpha = .05 / 3 .0 1). In the first planned comparison,the average of the habituated response levels (Blocks 1 and 2) across groups and the twointervals was contrasted with the average of the initial response levels (Blocks 2 and 3) toconfirm that recovery from habituation did, overall, occur. The second planned comparisonwas used to examine the effects of pre-exposure to heat shock on the degree of recoveryexpressed by contrasting the average recovery (Block 1 to 2 and Block 2 to 3) of the groupsthat did not receive pre-exposure to heat shock (LTH and LTH-HS) with the average recoveryexpressed by the PRE HS / LTH-HS group. The third planned comparison was used toexamine the effects of heat shock during training on the degree of recovery expressed. Here,the average recovery (Block 1 to 2 and Block 2 to 3) of the LTH group was compared to theaverage recovery exhibited by the two groups that received heat shock during training (LTHHS and PRE HS I LTH-HS).Long-term habituation was assessed with a two-way, mixed-design ANOVA (Day xGroup; alpha = .05). The expression of long-term habituation in all groups taken together80would be seen as a significantly lower response level on Day 2 than on Day 1. A significantinteraction between Day and Group might indicate a difference in the expression of long-termhabituation.Three planned comparisons were used to examine these data. To examine the effects ofpre-exposure to heat shock on initial response levels on Day 1, the average of the Day 1response levels of the two groups that did not receive pre-exposure to heat shock (LTH andLTH-HS) were contrasted with the Day 1 response level of the PRE HS / LTH-HS group. Thecapacity of pre-exposure to heat shock to induce thermal tolerance for the disruptive effects ofheat shock during training on long-term habituation was tested by contrasting the change fromDay 1 to Day 2 of the LTH-HS group and the PRE HS / LTH-HS group. To examine whetherthe PRE HS I LTH-HS group exhibited the disruption of long-term habituation previously seenin Experiment 5 and whether the LTH group exhibited long-term habituation, the average of thechange from Day 1 to Day 2 of the LTH group was contrasted with the average of the changefrom Day 1 to Day 2 of LTH-HS group and the PRE HS / LTH-HS group.ResultsHabituation with pre-exposure to heat shock and heat during training. Short-termhabituation occurred in all three groups (see Figures 14A, B and C; mixed design ANOVA;alpha = .05; Training: E(1, 51) = 155.81, < .01). There was significant retention ofhabituation between blocks of training W(2, 102) = 15.61, < .01). The heat shock treatmentsdid not significantly affect the overall level of responding (Group: F(2, 51) = 1.92, n.&)although the significance of the countemull effect size indicates that it cannot be assumed thatthe heat shock treatments had no effect on response levels (obtained ES: .31, countemull ES:61, F(2, 51) = 7.68, p < .01). Furthermore, habituation within blocks did not vary betweenblocks (Block x Training: E(2, 102) = .27, n&) or across groups (Training x Group: F(2, 51)= .05, n.s.). However, there was a significant interaction between Block, Training and Group81Figure 14. Habituation curves: distributed habituation training with pre-exposure to heatshock (45 mm, 32°C) and heat shock during training (45 mm, 32°C). The mean responsemagnitude (mm) to each stimulus given during training on Day 1 and testing on Day 2,24hr after the end of training, is shown for subjects receiving distributed habituation training(60 stimuli in three blocks of 20 stimuli at a 60 s 1ST with one hour rests between blocks)with pre-exposure to heat shock and heat shock after each of the three training blocks(PRE HS / LTH-HS; Fig. 14A; =20), subjects receiving distributed training with heatshock (45 min, 32°C) after each of the three training blocks (LTH-HS; Fig. 14B; fl = 20),and subjects receiving only training (LTH, Fig. 14C, n= 19).82DAY1 —DAY2—A= 3.z z2.5 2.52 q 2(n C)wZE u in cin.0%1.5:. LU0’. ç 1.5’uj 0.5x x oxI I I I I O’j1 20 21 40 41 60 1 203B 3 x2.5 2.5z zCl)2 eC)1.5’i.! IJt3J”ZE in r inZZ<0 05 Ui Ui w 0.5_____L_I I I 0’,1 20 21 40 41 60 1 20C 3 3.2.5’ 2.5ZE0’LU(n1%kTUJZZ0.5’<0UJ< 0.50, I I I 0-,1 20 21 40 41 60 1 20STIMULI83(E(4, 102) = 2.66, p .04). An inspection of Figure 14 does not suggest any obvious pattern.It may be that there is a subtle difference that wifi be evident in an analysis of recovery fromhabituation.The presence of recovery from habituation was assayed by contrasting the average ofthe habituated response levels of one block with the initial response levels of the followingblock (alpha = .0 1). Overall, there was recovery from habituation between blocks of training(E(1, 51) = 87.45, <.0 1). The effect of pre-exposure to heat shock on the expression ofrecovery was tested by contrasting the average difference between the habituation responselevels of one block with the initial response level of the next of the LTH and LTH-HS groupswith the average difference of the PRE HS I LTH-HS group (alpha = .0 1). Pre-exposure toheat shock did not appear to affect recovery from habituation ((1, 51) = .97, ni). Theeffects of heat shock during training on the recovery from habituation expressed was examinedby contrasting the averaged mean of the differences between habituated and initial responselevels of the LTH group with the averaged mean of the differences between habituated andinitial response levels of the LTH-HS group and the PRE HS / LTH-HS group. No effect ofheat shock on recovery from habituation was observed (E(1, 51) = .10, n.&).The analysis of short-term habituation has determined that the habituation exhibited bythe three groups, LTH, LTH-HS and PRE HS I LTH-HS has much in common, such asretention of habituation between blocks of training, the degree of habituation during trainingand the degree of recovery from habituation. The interaction between Block, Training andGroup has not yet been accounted for. Without a context for this effect or any discerniblepattern in the data, it may be that this interaction, while statistically significant, is not of anytheoretical importance.Long-term retention of habituation training. Response levels on Day 1 and 2 in thethree groups were compared with an overall mixed design ANOVA (see Fig. 15; Day x Group;84alpha = .05). While the overall response level did not vary with either group or day (Group:E(2, 56) = .14, n; Day: E(1, 56) = .04, n). the change from Day 1 to Day 2 did varyacross the groups (Day x Groups: F(2, 56) = 3.7, = .03).One effect that may have contributed to such an interaction is the effect of the preexposure to heat shock on the initial response levels on Day 1. The effect of pre-exposure ofheat shock on the initial response levels was considered in a planned comparison thatcontrasted the average Day 1 response level of the LTH and LTh-HS groups with the Day 1response level of the PRE HS I LTH-HS group (alpha .0 1). No effect of pre-exposure toheat shock was observed (see Figure 10; (1, 56) .024, n.&).Although it cannot be explained by an effect of the pre-exposure to heat shock on theinitial level of responding, the significant interaction between Day and Group from the overallANOVA is not surprising, as the LTH group should exhibit long-habituation (Experiments 2and 4) while the LTH-HS group should exhibit a block in long-term habituation (Experiment5). From an inspection of Figure 14, it is evident that in the LTH group, the level ofresponding decreased from Day 1 to Day 2, while in the LTH-HS group, the level ofresponding increased slightly from Day 1 to Day 2.Two planned comparisons were used to define the effects of pre-exposure to heat shockon the disruption of long-term habituation by heat shock during training and to contrast thechanges from Day 1 to Day 2 in response levels between the LTH-HS and LTH groups. In thefirst comparison, the difference between Day 1 and Day 2 response levels of the PRE HS ILTH-HS group was contrasted with the differences in Day 1 and 2 response levels. If preexposure to heat shock prevented the disruption by heat shock during training, the differencebetween Day 1 and Day 2 would significantly greater in the PRE HS I LTH-HS group than theLTH-HS group (alpha = .01). There was no effect of pre-exposure to heat shock on the85Figure 15. Habituation by block: distributed habituation training with pre-exposure to heatshock (45 mm, 32°C) and heat shock during training (45 mm, 32°C). The mean blockmagnitudes on Day 1 and Day 2 of the three groups that received training on Day 1: thePRE HS I LTH-HS group (jj, = 20) which received heat shock before and during training,the LTH-HS group (ij, = 20) that received heat shock during training only, and the LTHgroup (n = 19) that received training only are shown (error bars show -i-I- SE).86_LUcoc100-J1.J DAY1•LTH PRE HS I LTH-HS LTH-HS87change from Day 1 to Day 2 expressed by the PRE HS / LTH-HS; the two groups changed thesame amount between Day 1 and Day 2 (F(l, 38) = .275, n).A fmal planned comparison was used to compare the change in response level fromDay 1 to Day 2 exhibited by the group that received only training on Day 1 with the averagedchange in response level from Day 1 to Day 2 in the groups that received heat shock duringtraining (alpha .01). It was found that the change exhibited by the LTH group wassignificantly different from the change exhibited by the LTH-HS and PRE HS / LTH-HSgroups ((1, 56) = 7.09, p <.0 1). One interpretation of these results is that the LTH groupexhibited long-term habituation, the LTH-HS group did not, and the pre-exposure to heatshock failed to prevent the block of LTH by heat shock during training. However, this is notthe only possible interpretation. As can be seen in Figure 15, the Day 2 response level of theLTH group is similar to the Day 2 response level of the LTH-HS group, and that the Day 1response levels of these two groups are not as similar. It may be that the significant contrastbetween the differences in Day 1 to Day 2 response levels have been affected by differing initialresponse levels. This possibility is explored in the synthesis of results after Experiment 8.Experiment 7The effects of heat shock just prior to testing on the retention of long-term habituationIn Experiment 4, pre-exposure to heat shock did not prevent the formation of long-termmemory; in Experiment 5, heat shock during training blocked the formation of LTM. In thisexperiment, the effect of heat shock on the retention of memory, is tested. Heat shock beforethe Day 2 retention test long after training may affect the processes necessary for the retentionof long-term habituation.MethodsSubjects and Materials. Worms were maintained as described in the general methods.Twenty worms were run in each of two groups for a total of 40 worms. The administration of88heat shock (45 mm, 32°C), the use of apparatus, behavioral observations and scoring ofbehavioral responses were performed as described in the general methods.Procedure. The first group, LTH / DAY 2 HS (n = 20), received habituation trainingon Day 1 as described in the general methods and received a heat shock ending 2 hr beforetesting on Day 2. The second group, DAY 2 HS ONLY (n = 20), received only a singlestimulus on Day 1, but like the LTH I DAY 2 HS group, received a heat shock ending 2 hrbefore testing on Day 2. The DAY 2 HS ONLY group was included to control for the directeffects of Day 2 heat shock on the level of responding during testing. Only the long-termretention of habituation will be analyzed in the results as the heat shock was presented aftertraining on Day 1 was complete.Statistics. The analysis of long-term habituation was performed as outlined in thegeneral methods. Two factorial ANOVAs (alpha = .05 / 2 = .025) were performed, one oneach level of Day (Day 1 and Day 2), comparing the performance of the different groups. In aplanned comparison on the response levels of Day 1, the average of the Day 1 response levelsof the trained groups (LTH and LTH I D2HS) was contrasted with the Day 2 level ofresponding of the untrained Day 2 heat shock only group to examine the effects of training(alpha = .025).Two planned comparisons were performed on the data from Day 2 (alpha = .025 I 2 =.0 1). The Day 2 response levels of the trained groups, LTH and LTH / DAY 2 HS, werecontrasted to examine the effects of Day 2 heat shock on Day 2 responding (alpha = .01). Todetermine whether both trained groups, overall, exhibited long-term habituation, the average ofDay 2 response levels of these two groups was contrasted with the Day 2 response level of theDay 2 response level of the DAY 2 HS group (alpha = .01).89Figure 16. Habituation by block: distributed habituation training with heat shock (45 mm,32°C) on Day 2, 2 hr before testing. The mean block magnitudes on Day 1 and Day 2 ofthe group which received only training, LTH (n = 19), the group that received training onDay 1 and heat shock before the retention test on Day 2, LTH I D2 HS (ii = 20), and theDAY 2 HS ONLY (11=20) group are shown (error bars show +1- SE). The DAY 2 HSONLY group received only a single stimulus on Day 1, no habituation training, and a heatshock on Day 2 that ended 2 hr before the retention test.90LU—z0<c1C.)0-J0 DAY 11•ITLTH LTH I D2 HS D2 HS ONLY91ResultsLong-term retention of habituation after Day 2 heat shock. The effects of training andheat shock on Day 2 on the levels of responding on Day 1 and Day 2 were compared with twofactorial ANOVAs, one on each level of Day (alpha = .025). The ANOVA on Day 1 includedthe Day 2 response levels of the D2 HS control group. There was no difference between thegroups (see Fig. 16; Day 1, Group: (2, 56) = .400, n). A planned comparison thatcontrasted the average of the Day 1 response levels of the trained groups with the Day 2response level of the DAY 2 HS control group confirmed that heat shock on Day 2 did notdepress or elevate responding from baseline (E(1, 56) = .01, nj. On Day 2, there was asignificant difference between the response levels of three groups (Day 2, Group: E(2, 56) =.93, n). The effects of heat shock on Day 2 on the retention of long-term habituation wasexamined with a planned comparison contrasting the Day 2 response levels of the LTh andLTH / DAY 2 HS groups. There was no effect of heat shock on Day 2 on the response levelsafter training on Day 2 (E(1, 56) = 3.44, The average of the Day 2 response levels of thetrained groups was contrasted with the Day 2 response levels of the D2 HS control group.Taken together, the LTH and LTH I DAY 2 HS groups express long-term habituation, as seenby a significantly lower response level than the controls (E(1, 56) 9.54, < .01). Thus, itappears that heat shock on Day 2 does not affect responding on Day 2 or the retention of long-term habituation.Experiment 8Titration of the effects of heat shock during the rest periodWork with Aplysia and other systems has emphasized the dynamics of the cellularprocesses that support LTM (Byrne et al., 1993). In Experiments 5 and 6, we have seen thatin . elegans, heat shock delivered during the rest periods of distributed training disrupts theretention of LTM for habituation training. Unlike the protein synthesis inhibitors used in work92with Aplysia, heat shock can be delivered in brief, discrete pulses. Thus, it may be possible totitrate the timing of heat shock to determine whether there is a shorter interval within the 1-hrest period during which LTM formation is particularly vulnerable to disruption. In thisexperiment, the effects of brief heat shock (15 mm, 32°C) given either early, mid-way or late inthe 1-h rest periods on the retention of habituation training were examined.MethodsSubjects and Materials. Worms were maintained as described in the general methods.Twenty subjects were run in each of four groups for a total of 80 subjects. The administrationof heat shock (15 mm, 32°C), the use of apparatus, behavioral observations and scoring ofbehavioral responses were performed as described in the general methods.Procedure. All groups received habituation training on Day 1 and were tested on Day 2as described in the general methods. The first group, LTH (n = 20), received no heat shock inthe 1-hr rest periods during training. The second group, LTH-EARLY HS (j = 20), received15 mm heat shock from 0 to 15 min in the 1-hr rest periods during training. The third group,LTH-MID HS (n = 20), received 15 mm heat shock from 15 to 30 mm in the 1-hr rest periodsduring training. The fourth group, LTH-LATE HS ( = 20), received 15 mm heat shock from30 to 45 min in the 1-hr rest periods during training. Only long-term habituation wasconsidered in the analysis because the longer duration 45 mm heat shock used in Experiments3, 4 and 5 did not affect the expression of habituation during training on Day 1 nor the short-term retention of habituation.Statistics. The levels of responding on Day 1 and Day 2 were compared across groupswith a mixed-design ANOVA (Day x Group; alpha = .05). Although there was no significantinteraction between Day and Group, five comparisons were made to examine the data forpatterns that may help direct future efforts to narrow the critical period of memoryconsolidation during long-term habituation. To compare the initial response levels, a factorial93ANOVA was performed across groups on the Day 1 response levels. With a significantinteraction between Day and Group in the overall ANOVA, the long-term habituation expressedin each group would be assessed with a planned comparison of that group’s Day 1 and Day 2response levels. The alpha level for these comparisons was reduced: alpha = .05 / 5 = .01.ResultsLong-term retention of habituation training. The overall ANOVA (Day x Group)compared levels of responding on Day 1 and Day 2 across the groups tested. Overall, the levelof responding on Day 1 was significantly higher than the level of responding on Day 2 (seeFig. 17; (1, 76) = 30.10, p <.01). The level of responding overall did not vary significantlywith the heat shock treatment (Group: F(3, 76) = 2.13, nor was there a significantdifference in the change in response levels between Day 1 and Day 2 across groups (Day xGroup: E(3, 76) = 1.29, n). However, the countemull values of the effect sizes of Groupand the interaction of Day and Group were significant, so it cannot be assumed that Group hadno effect on the overall level of responding or the change in response levels from Day 1 to Day2 (Group: obtained ES: .32, counternull ES: .65, f(3, 76) = 8.52, p < .01; Day x Group:obtained ES: .25, counternull ES: .50, E(3, 76) = 5.16, p < .01). The initial response levelson Day 1 were compared across groups with a factorial ANOVA; there was no significantvariation between groups in the initial response level (alpha = .01; Day 1, Group: E(3, 76) =.449, nj. Therefore any apparent differences between groups in the change from Day 1 toDay 2 in response levels may be attributed to a difference in the Day 2 level of responding.Because no significant interaction between Day and Group was found, the plannedcomparisons between Day 1 and Day 2 within each group were not carried out. However, ascan be seen in Figure 17, it is possible that there was a weakening of the long-term habituationexpressed by the LTH-early HS and LTH-mid HS groups perhaps representing a timedependent vulnerability to disruption in memory consolidation. However, these effects are far94Figure 17. Habituation by block: distributed habituation training alone or with brief heatshock (15 mm, 32°C) during training either early, mid or late in the 1-h rest periodfollowing each training block. The mean block magnitudes on Day 1 and Day 2 of all fourof the groups: the LTH group (n = 20; no heat shock during 1-h rest periods), the LTHEARLY HS group (11=20; heat shock from 0 to 15 min in the 1-h rest periods), the LTHMID HS (ja = 20: heat shock from 15 to 30 mm in the 1-h rest periods, and the LTH-LATEHS group (= 20; heat shock from 30 to 45 min in the 1-h rest periods), are shown (errorbars show +1- SE).m3:-I-<IDO.>-<-<20STIMULIBLOCKMAGNITUDE(mm)0 01-LII1%)a’IIC’,Ix-IrrLit96too small to interpret these results as an indication of defined critical period. In future studies,an effort should be made to sharpen the possible disruption of long-term habituation by briefheat shock by altering the training procedure, the intensity of heat shock, or both.Synthesis of results: An analysis including Experiment 2and Experiments 4 through 8There are several issues in this work which may be addressed by comparing the resultsof the experiments with each other, as has been done throughout the results, and by theperformance of direct comparisons across experiments. Differences between experiments inthe overall level of responding may be examined. The variability in initial response levelsobserved between groups receiving the same treatment but in different experiments, or differenttreatments in the same experiments may be examined, and the impact on the interpretation ofthe results of individual experiments discussed. The effect size for those phenomena over all ofthe experiments may be calculated. The objective here is to synthesize the results from thedifferent experiments, examine the data for systematic differences by experiment and treatment,and discuss the pattern of the results and the appropriate interpretation of this pattern. Thisdiscussion will be limited to the analyses of long-term habituation, as this is where the moststriking and theoretically important differences with training and treatment were observed.One possible problem with the interpretation of these experiments may be differencesbetween experiments in overall response levels; such differences may affect the outcome ofindividual experiments, making it difficult to interpret the results of each experiment in thecontext of the others presented here. To examine this possibility, a comparison of the overalllevel of responding across experiments was done with the groups that received only distributedhabituation training (60 stimuli at a 60-s ISI) on Day 1 and the retention test of 20 stimuli onDay 2 with a mixed-design ANOVA (Day x Experiment; alpha = .05). The five LTh groupsfrom Experiments 2 (j = 21), 4 (n= 20), 6 (ii = 19), 7 (a = 19) and 8 (a = 20) were included97Figure 18. The mean Day 1 and Day 2 response levels across experiments: LTH groups.A. The effects of training by Day and Group. B The effects of training by Day.98DTEXP. 2EXP. 4EXP. 6EXP. 7EXP. 8UDAY1A—Ui-JoI-zCoeC.,0BE_LUI-zC1C)0-JDAY 23-2.5-2-1.5-1—0.5-0-3—2.5-21 .510.50DAY1 DAY299Figure 19. A. The effects of training with heat shock by Day and Group. B. The effects oftraining with heat shock by Day.100AUiz00-JBUiz00-JQ EXP. 5• EXP. 6DAYI DAY2DAYI DAY2101(see Figures 18A and B). This analysis showed that the level of responding on Day 2 waslower than on Day 1 across experiments (Day: (1, 94) = 33.875, < .0 1). There was nooverall significant difference in the level of responding between experiments (Experiment: F(4,94) = 1.374, Furthermore, the change between Day 1 and Day 2 did not varysignificantly between experiments (Day x Experiment: E(4, 94) = .538, This last resultis particularly reassuring, as it suggests that the expression of long-term habituation did notvary between the experiments. The countemull value of the effect size of Experiment wassignificant, while the countemull value of the effect size of the interaction between Experimentand Day was not (Experiment: obtained ES: .25; counternull ES: .50, E(4, 94) = 5.51,.01; Day x Experiment: obtained ES: .16; countemull ES: .32, E(4, 94) = 2.15,. While itis possible that there are differences in the overall response level between experiments, it can besaid with some confidence that differences between the experiments does not affect the changein the level of responding from Day 1 to Day 2 (see Figures 18A and B). A mixed-designANOVA (Day x Experiment) comparing the two groups that received heat shock duringtraining was also performed (alpha = .05). The two LTH-HS groups from Experiment 5 (=18) and 6 (j, = 20) were included (see Figures 19A and B). There was no significant changewith Day overall (Day: E(1, 36) 3.243, ii.&). In addition, response levels did not varyoverall with the Experiment (Experiment: E(1, 36) = 3.395, j). Finally, the response levelsof Day 1 and Day 2 did not interact significant with experiment (Day x Experiment: f(1, 36) =.452, It is worth noting that the effects of Day and Experiments came close tosignificance; the counternull values of the effect size (ES) of each are statistically significant:Day: observed ES = .41; counternull ES = 2(.41) - 0 = .82, E(1, 36) = 12.97, < .01;Experiment: observed ES = .42; counternull ES = .84, E(1, 36) = 13.58, < .01). Thuswhile Day and Experiment did not have a significant effect on the level of responding in these102LTH-HS experiments, it cannot be concluded that there is no effect of Day and Experiment inthis type of training.Overall, while there was no significant effect of Experiment or interaction between Dayand Experiment on the response levels observed on Day 1 and Day 2 in the LTH and LTH-HSgroups, the possibility that differences between experiments may contribute to overall levels ofresponding in both LTH and LTH-HS groups and to the changes in the level of respondingfrom Day 1 to Day 2 in the LTH-HS groups cannot be discounted. However, as can be seenfrom Figures 18 and 19, the effect of training and heat shock during training is apparent whenthe results from different groups and experiments are taken together.To quantify the effects of training and heat shock during training, the effect size for dayof training across all LTH groups and across the two LTH-HS groups may be calculated. Theeffect size associated with the Day main effect in the ANOVA comparing the LTH groups was1.30, while the effect size associated with the Day main effect in the ANOVA comparing theLTH-HS groups was 0.40. As can be seen in Figures 18 and 19, these effects go in theopposite direction.In studies from other laboratories examining long-term habituation, expression of LTHwas sometimes observed as a decrease in the initial response magnitude from training to test(Leaton & Supple, 1991). Here, the change in the initial response magnitude from training totest was examined across the LTH groups (Experiments 2, 4, 6, 7 and 8) and the LTH-HSgroups (Experiments 5 and 6) using a factorial design ANOVAs (Day x Experiment; a factorialANOVA was used because the number of missing values made a repeated-measures analysis ofDay of training inappropriate). Across the LTH groups, there was a significant decrease fromDay 1 to Day 2 in the mean magnitude of the initial response (E(1, 150) = 8.42 1, < .0 1),while there was no difference between experiments in the overall initial response magnitude orthe difference between Day 1 and Day 2 in initial response magnitude (Experiment: f(4, 150) =103.72, n; Experiment x Day: E(4, 150) = .61, Across the LTH-HS groups, there wasno significant difference between Day 1 and Day 2 in the initial response level ((2, 59)1.326, n); in addition there was no significant difference between experiments in the overallinitial response magnitude or the initial response level on Day 1 and Day 2 (Experiment: E(1,59) = .16,; Experiment x Day: F(1, 59) = .40, ii.aJ. Overall, these results indicate thatthe change in initial response magnitude from Day 1 to Day 2 reflects both the retention ofhabituation in the LTH groups and the disruption of habituation after heat shock duringtraining.One issue that is of concern in the interpretation of the results from the LTH-HSgroups of Experiments 5 and 6 is the possibility that the failure to observe long-termhabituation on Day 2 in these groups was not caused by the heat shock treatment duringtraining, but rather, to an unusually low initial level of responding on Day 1. If so, themean initial response level of the LTH-HS groups should be consistently lower than theDay 1 response levels and comparable to Day 2 response levels of the LTH groups.While the LTH-HS group from Experiment 6 did exhibit a lower level of responding onDay 1 than any of the LTH groups (Day 1 mean +1- SE (in mm): 1.245 +1- .075), theLTH-HS group of Experiment 5 had a level of responding on Day 1 higher than two ofthe five LTH groups (LTH-HS, Exp. 5 Day 1: 1.544 +1- .146; LTH, Exp. 2, 1.422 ÷1-.132; LTH, Exp. 7: 1.429 +1- .113).However, such comparisons only go so far in answering such concerns. Thefactors that resulted in a low level of responding in the LTH-HS group of Experiment 6are not known (note that during the first block of training, the worms have not yetreceived the first heat shock treatment). An overall ANOVA comparing the Day 1response levels across all groups that received distributed training and did not receivepre-exposure to heat shock showed a significant effect of group despite the fact that all104groups received the same treatment up to that point (eleven groups included; E( 10, 206)= 2.463, < .01). Fisher’s PLSD post-hoc comparisons showed significantdifferences between a number of groups (see Table 1). The variability in initialresponse levels is consistent with the findings in other studies of habituation of thereversal response to tap from this laboratory; in these studies, as with this one, thecharacteristics of habituation are consistent despite the variability of the initial responselevel (Rankin & Broster, 1992; Broster & Rankin, 1994). Therefore, the failure to seelong-term habituation in the groups of the present experiments that received heat shockduring training need not be attributed to a depressed initial level of responding.In Experiment 4, the effects of pre-exposure to heat shock on levels ofresponding on Day 1 and Day 2 were explored. To examine the effects of pre-exposureto heat shock on Day 1 response levels more closely, the data from Experiments 4 and6 may be considered together. Both the PRE HS I LTH (Exp. 4) and the PRE HS ILTH-HS (Exp. 6) groups received heat shock 2 h before the first 20 stimuli block onDay 1, while the LTH (Exp. 4) and the LTH-HS (Exp. 6) groups did not. A two-wayfactorial ANOVA (PRE HS x Experiment) was used to compare the effects of preexposure to heat shock on response levels to the first twenty stimuli on Day 1 acrossexperiments (alpha = .05). The results showed a significant effect of Experiment, withsubjects in Experiment 4 responding at a significantly higher level than subjects inExperiment 6 (Experiment: E(1, 95) = 18.07, < .0 1). There was no significant effectof the pre-exposure to heat shock, nor any interaction between the effect of experimentand the presence of the pre-exposure to heat shock (PRE HS: F(1, 95) = .76, n.&;Experiment x PRE HS: E(1, 95) = 1.07, n). Clearly, while there were experimentalfactors that affected the initial response levels, pre-exposure to heat shock did not have105an impact on initial response levels measured during the first twenty stimuli of trainingon Day 1.In summary, the pattern of results of the experiments using the distributedtraining procedure at a 60-s ISI indicate that while there were uncontrolled factors thataffected response levels, the effects of training on the change in responding from Day 1to Day 2 and the disruption of that change by heat shock during training were robust.Factors which may have contributed to the differences observed between experimentsusing the same treatments include differences in experimenter, temperature of testingenvironment (19° - 24°C), variations in the worm strain, scorer differences and so on.It is important to consider the results of each experiment in the context of the others. Inthis manner, by comparing the performance over all of the groups that received trainingonly, that received heat shock during training, or pre-exposure to heat shock beforetraining, the factors which affect the development of retention for habituation maybecome apparent.DiscussionThe experiments in this dissertation examine factors that affect the developmentof long-term habituation in . elegans. The objective was to determine trainingparameters that affect the development of long-term habituation, establish a procedurethat produces long-term habituation, and then use interference with memoryconsolidation as a tool to explore the dynamics of long-term habituation.In Experiment 1 (10-s 1ST), neither distributed nor massed habituation trainingat a 10-s 1ST produced unambiguous LTH. The possibility that a longer interstimulusinterval would lead to the expression of long-term memory was tested in Experiment 2(60-s 1ST). At a 60-s 1ST, distributed and massed habituation training taken together ledto long-term memory for habituation.Fishers PLSD for dlEffect: Day 1 GroupsSIgnIflcarn Level: 5 %LTH (E2), LTH(E4)LW (E2), LTH-HS (ES)LTH (E2). LTH (E6)LTH (E2). LTH (E7)LW (E2), LTH (ES)LTH (E2). LTH-HS (E6)LTH (E2), LTH / 1)2 MS (El)LW (E2), LTH-eHS (ES)LTH (E2). LTH-mHS (ES)LW (E2), LTH-LHS (ES)LTH(E4). LTH-HS (E5)LTH(E4), LTH (E6)LTH(E4). LTH (El)LTH(E4), LTH (ES)LTI-1(E4), LTH-HS (E6)LTH(E4), LTH / 1)2 MS (E7)LTh(E4), LTH-eHS (E8)LTH(E4), LTH-mHS (E8)LTH(E4), LTH-IHS (ES)LTM-HS (ES). LTH (E6)LTH-HS (E5). LTH (El)LTH-HS (E5). LTH (E8)LW-MS (E5). LTH-HS (E6)LTH-HS (E5). LTH / 1)2 MS (E7)LW-MS (E5), LTH-eHS (ES)LTH-HS (ES), LTH-znHS (ES)LTh-HS (ES), LTH-IHS (ES)LTH (E6), LTH (El)LW (E6). LTH (ES)LTH (E6). LTH-HS (E6)LW (E6), LW I D2 MS (E7)LW (E6), LW-eHS (ES)LW (E6). LTH-mHS (E8)LW (E6), LTH-114S (ES)LW (El), LW (E8)LW (El), LW-MS (E6)LW (El), LW /1)2 MS (El)LW (E7), LW-eMS (ES)LW (E7), LW-mHS (ES)LW (E7), LW-hiS (E8)LW (ES), LW-HS (E6)LW (ES), LW / 1)2 HS (E7)LW (ES), LW-eNS (ES)LW (ES), LW-mHS (ES)LW (ES), LTh-IHS (ES)LW-MS (E6), LW /1)2 HS (E7)LW-HS (E6). LW-eHS (ES)LW-MS (E6). LW-mHS (ES)LW-MS (E6), LW-INS (ES)LW / 1)2 MS (El). LW-eMS (E8)LW / 1)2 HS (E7), LTh-miIS (ES)LW / 1)2 MS (E7), .LTH-IHS (ES)LW-eMS (ES), LW-inNS (ES)LTH-eHS (Es). LW-INS (ES)LW-mNS (ES), LW-iNS (ES)SSSSMean Diff. Crit. Diff P-Va...106SSSS-.320 .330 0571-.122 .339 .4787-.106— .334 .5333• 007 .334 .9655-.258 .330 .1249.176 .330 .2929-216 .330 .1980-.386 .334 0239-.443 .326 .0079-.304 .330 .0706.198 .343 2562.214 .338 .2129.313 .33S 0697.062 .334 .7130.496 .334 .0037.104 .334 .5394-.066 .338 .7021-.123 .330 .4620.016 .334 .9244.016 .341 .9265.115 .347 .5155-.136 .343 .4362.298 .343 .0877-.094 .343 .5896-.264 .347 .1351-.321 .339 .0631-.182 .343 .2968.098 .342 .57 15-.152 .338 .3761.282 .338 .1015-.110 .338 .5212-.280 .342 .1085-.33 8 .334 .0418-.198 .338 .2492-.250 .338 .1458.184 .338 .2853-.209 .338 .2251-.378 .342 .0305-.436 .334 .0108-.297 .338 .0852.434 .334 .0111.042 .334 .8056-.128 .338 .4562-.186 .330 .2684-.046 .334 .7849-.392 .334 .0214-.562 .338 .0012-.620 .330 .0003-.480 .334 .0050-.170 .338 .3234-.227 .330 .1756-.088 .334 .6038-.058 .334 .7344.082 .338 .6340.139 .330 .4057SSS107It is interesting that the age and handling dependent effect was seen with testingat a 10-sec ISI but not at a 60-s ISI. This pattern was observed in an additionalexperiment in two untrained groups (a single stimulus control group and a zerostimulus control group) at a 10-s 1ST (Marion, Beck & Rankin, 1992). However, intheir work on the effects of age on habituation, Beck and Rankin (1993) did not findany change in the depth of response decrement after habituation training at a 10-s 1STwhen four day old worms were compared to seven day old worms using handlingprocedures similar to those used in the present study. The present experimentscompare the habituation of four day old worms with five day old worms; it is possiblethat the difference in habituation observed here over 24 h may have been missed inBeck and Rankin’s (1993) earlier comparison of four and seven day old worms. Morework on the nature of the effects of age and prior stimulation on habituation and long-term retention of habituation is needed to clarify this issue.Although the short 1ST (10-s) produced a greater response decrement duringtraining on Day 1 than the long TSI (60-s), LTH was evident after training with the long151. In addition, on Day 1, the expression of habituation was affected by the trainingprocedure (distributed, massed or twenty stimuli control training) when training wasgiven at a 60-s 151, but not at a 10-s 1ST. Whether habituation training at a 10-s 1ST iscapable of producing long-term habituation in . elegans is not known. Retention ofhabituation across blocks of training, or short-term retention of habituation, wasevident with training at a 60-s ISI and at a 10-s ISI. It should be noted that the failureto find evidence for long-term habituation at a 10-s 1ST may be due to a floor effectresulting from the very rapid and deep habituation to training at a 10-s 1ST. It seemsparadoxical that greater response decrement on Day 1 should lead to less retention ofhabituation; however, this same pattern is apparent in the rate of spontaneous recovery108from short-term habituation. As discussed above, the rate of recovery from habituationis faster after habituation at a 10-s ISI than after habituation at a 60-s 1ST, despite thegreater depth of habituation during training at a 10-s ISI. These findings suggest thatthe rate of spontaneous recovery from habituation may be a better predictor of retentionfor habituation training than the amount of response decrement expressed duringtraining.Broster and Rankin (1994) hypothesized that 10-s and 60-s ISIs may recruitdifferent cellular processes during habituation training. If this is the case, the cellularprocesses recruited by habituation training at a 60-s 1ST may be necessary for theformation of memory for habituation. Testing the effects on habituation of treatmentsthat alter specific cellular processes, such as ablations of subclasses of neurons in thetouch circuit (Wicks and Rankin, unpublished observations), may help characterize thecontributions of these processes to habituation at different ISIs.The effects of the distribution of training at a 60-s 1ST on LTH are similar to theresults seen in experiments on long-term habituation with Aplvsia. and support the ideathat memory formation benefits from training that is distributed over time rather thantraining that is condensed (Carew et al., 1972; Carew and Kandel, 1973). There aretwo explanations that could account for the effect of distributed training on retention.The first is that memory for training is formed during training itself. During distributedtraining, a subject habituates three times. Breaking the habituation training into threedistinct blocks permits recovery from short-term habituation. Each further block oftraining leads to rehabituation and thus greater accumulation of memory. According tothis hypothesis, available time for an appropriate amount of recovery from habituationbefore the next block of training would be critical, so that the rehabituation would havea maximal impact on the long-term memory for habituation training.109The second possible explanation of the benefit of distributed training overmassed training is that the long-term memory for training forms after training hasfinished. In the distributed procedure, the 1-h rest periods would permit the activeencoding of the habituation into long-term memory. A number of training blocks withrest periods after each permit greater accumulation of memory for the training.According to this hypothesis, short-term habituation and recovery processes would notgovern the formation of long-term memory for habituation. Rather, other independentprocesses (e.g. up- or down-regulation of specific proteins) that control memoryformation would determine the optimal rest period between blocks of training indistributed procedure.Interference as a tool to examine memory consolidationTo test these possibilities, the role of the rest period of distributed training inmemory formation was examined. In this approach, conditions were introduced duringthe rest period that would perturb the cellular processes which may be necessary formemory formation without interrupting spontaneous recovery from habituation in anattempt to determine whether there is an interval of the rest period critical for memoryformation.Heat shock (45 min, 32°C) induces heat shock protein production as evidencedby the B-galactosidase staining in transgenic animals after this heat-shock treatment (seeFigure 8; Stringham et al., 1992). The presence of heat shock proteins is a marker ofcellular stress (Stringham et al., 1992; Lindquist, 1986); therefore worms that havereceived heat shock (45 mm, 32°C) show evidence of a history of cellular stress. Thedamage resulting from the stress or the active cellular response to stress, whichincludes a reduction in protein synthesis and the production of heat shock proteins, mayalter the processes which mediate learning or consolidation of memory. Thus, heat110shock may be used to probe the dynamics of long-term habituation for an interval orintervals critical to memory formation.In Experiment 4, heat shock (45 mm, 32°C) was delivered so that it ended 2 hbefore the beginning of training. Heat shock before training might affect responding onDay 1 or it might affect the change in the level of responding from Day 1 to Day 2.This could occur due to residual cell damage, altered cellular processes or theaccumulation of heat shock proteins. However, the findings indicated that a history ofcellular stress produced by this heat shock treatment did not affect either the respondingon Day 1 or the expression of long-term habituation as seen by the change ofresponding from Day 1 to Day 2 (see Figures 10 and 11).The failure of a single heat shock (45 mm, 32°C) to produce a change inbehavior is interesting, in light of the demonstration that a single heat shock (45 min,32°C) caused the production of heat shock proteins (from Experiment 3A; see Figure8), and depressed egg-laying measured 15 mm after the end of the heat shock treatment(Experiment 3B; see Fig. 9). Overall, these findings suggest that this heat shocktreatment was indeed a significant cellular stressor, and that its lack of effect onbehavior is a reflection of the time it was given: two hours before training.In Experiment 5, the effects of heat shock (45 mm, 32°C) given during trainingon habituation, short-term retention of habituation, and long-term retention ofhabituation were examined. Heat shock given without training had no effect onresponding on Day 2. In addition, heat shock during training did not affect theexpression of habituation or short-term retention of habituation training; however, it didprevent the expression of long-term habituation (see Figures 12 and 13). Heat shockgiven in this procedure was certainly sufficient to cause the induction of heat shockproteins (Experiment 3A; see Figure 8); however, the induction of 8-galactosidase did111not appear to be any more intense after three heat shocks for 45 mm each than after asingle heat shock of 45 mm. On the other hand, the triple heat shock (45 mm, 32°C)definitely appeared to produce a greater depression in egg-laying than a single heatshock (45 mm, 32°C). Thus, while the results of Experiment 3A on the induction ofhsp 16 support an interpretation of the effect of heat shock on training being a result ofthe time the heat shock treatment was given, the possibility of a dose-effect of the singlevs. triple heat shock cannot be set aside.In many systems, the effects of severe heat shock can be alleviated by an earliermilder heat shock; this phenomenon is called induced thermal tolerance (Lindquist,1986; Parsell, Taulien & Lindquist, 1993). In Experiment 6, to examine whetherthermal tolerance could be induced to the effects of heat shock during training on long-term habituation, a pre-treatment of a heat shock (45 mm, 32°) was given as inExperiment 4 ending 2 h before training, followed by training with heat shock as inExperiment 5. If thermal tolerance for the effects of heat shock (45 min, 32°C) duringtraining were induced by the exposure to heat shock before training, long-termhabituation should not be blocked. However, the findings indicated that thermaltolerance for the effects of heat shock during training on long-term habituation was notinduced; in the group that received heat shock before training as well as during training,long-term habituation was blocked as it was in the group that received only heat shockduring training (see Figures 14 and 15).In Experiment 7, retention of long-term habituation was examined byintroducing a heat shock treatment (45 mm, 32°C) that ended 2 h before the retentiontest on Day 2. Studies examining the effects of interference on retention of memoryhave generally found that interference is not effective at disrupting long-term memorylong after training has finished (Squire & Davis, 1985). The results of Experiment 7112showed that heat shock (45 mm, 32°C) on Day 2 without training on Day 1 did notaffect responding on Day 2. In addition, heat shock on Day 2 had no effect on theexpression of long-term habituation. The caveats that were noted in the discussion ofExperiment 4 and 5 apply here as well; the single heat shock treatment (45 mm, 32°C)may not be of a comparable magnitude to the triple heat shock during training (45 mm,32°C) that disrupted long-term habituation. While this possibility must be considered,these findings still suggest that memory is not vulnerable to disruption long aftertraining has been completed.In Experiment 8, the 1-h periods following blocks of training in the distributedprocedure during which long-term habituation was vulnerable to disruption wereexamined further for short intervals within the 1 h periods during which memoryconsolidation was particularly vulnerable to disruption by heat shock. Three brief heatshocks were used (15 mm, 32°C), which in Experiment 3A, caused the expression of8-galactosidase, and in Experiment 3B, depressed the rate of egg-laying as much as thetriple 45 min heat shocks. Heat shock (15 mm, 32°C) was given either early (0 - 15mm), mid-way (15- 30 mm) or late (30 - 45 mm) in the 1-h periods following trainingblocks. Although the overall ANOVA showed no significant interaction between Dayand Group, an inspection of Figure 17 suggests that there may be a weakening of long-term habituation when brief heat shock is given early or mid-way in the 1 h rest period.Little weight can be placed on these results as the overall ANOVA failed to detect anydifferences between groups in the change in the level of responding from Day 1 to Day2. However, these data do suggest that the first 30 mm following training blocks maybe particularly important to memory consolidation. In future studies, the definition of acritical period for the consolidation of long-term memory may be resolved by using a113brief heat shock at a higher temperature (e.g. 34°C) to emphasize and clarify the effectof brief heat shock at different intervals on long-term habituation.In addition, it would fascinating to examine whether the 1-h intervals after eachblock were equally important to memory consolidation. This could be done by treatinggroups with heat shock after different combinations of two blocks, or after only oneblock. If, for example, there was a differential effect of heat shock after the first blockcompared with the last block, that would have implications for the dynamics of long-term habituation and the processes that support the distributed training effect.The behavioral experiments, considered separately, each describe some aspectof the factors affecting the development of long-term habituation and the effects ofvarious heat shock treatments on learning and memory in . elegans. However, it isnot so much in the results of any one experiment that the emphasis should be placed buton the overall pattern of the results. With this in mind, an analysis of the effects offactors varying between experiments on the overall response levels, the initial responselevels and the expression of the main effects observed: the effects of training, and theblock of the effects of training with heat shock during training. The results indicate thatthere are factors which vary between experiments which must affect at least the initiallevel of responding.Knowing this, it is interesting to note that the effects of training and of heatshock during training on response level were consistent across experiments (seeFigures 18 and 19). The expression of these effects was robust across experiments inwhich differences in the worms strains, ambient temperature of the testingenvironment, experimenters and scorers may have all contributed to produce differentlevels of responding. In addition, when the groups were considered together, longterm habituation in the LTH groups was evident in a second measure, the change in the114magnitude of the response to the initial stimulus from Day 1 to Day 2; the disruption ofLTH was evident in this measure of initial response magnitude in the groups thatreceived heat shock during training.The distributed training effect: a psychological perspectiveUp to this point in this work, the implications of the effect of distributedtraining have been discussed from a biological perspective. However, the effect ofdistributed training is frequently discussed in psychological literature, and is considereda robust effect, particularly in verbal learning tasks (Melton, 1970; Hintzman, Block &Summers, 1973; Shaughnessy, 1976; Dellarosa & Bourne, 1986). While most of theseexperimental paradigms have focused on tasks which involve a conscious effort toretrieve information (either free recall or recognition tasks), a small effect of distributedtraining on implicit memory has also been demonstrated (Pemichet, 1994).There is no consensus as yet on the psychological processes that produce theadvantage for memory formation associated with distributed training (Perruchet, 1994;Dellarosa & Bourne, 1986; Toppino, Kasserman & Mracek, 1991; Ross & Lindauer,1978; Cornell, 1980, Hintzman, 1974). In the psychological literature, there are atleast three classes of theories explaining the effect of distributed training.The first theory explaining the distributed training effect is the encodingvariability hypothesis, which attributes greater recall of repeated items to the variety ofsubjective contexts in which the target is encountered during training, thus increasingthe number of retrieval routes (Dellarosa & Bourne, 1986). There has been relativelylittle empirical support for this theory; in one study there was a demonstration thatvariable contexts of encoding actually impeded memory rather than improving it(Postman & Knecht, 1983).115The second theory is the processing-inhibition hypothesis, which suggests thatmassed training inhibits processing of stimuli through habituation, while distributedtraining permits encoding of each item before the next is presented (Cornell, 1980;Hintzman, 1974). However, Dellarosa and Boume (1986) suggest that the failure todemonstrate the distributed training effect in infants (Toppino & DiGeorge, 1984)indicates that habituation is unlikely to be the mechanism behind the distributed trainingeffect, as infants are capable of habituation.The third is the previous encoding-accessibility hypothesis, which argues thatthe likelihood of encoding processes being engaged upon presentation of a stimulusdepends on how accessible the last encoding of that stimulus is. Thus the likelihood ofengaging the encoding processes should vary directly with the length of the intervalbetween stimuli (Rose, 1980; Rose & Rowe, 1976). It has been found that the effect ofdistributed training on recall is lost when other manipulations that increase thelikelihood of encoding were performed (Cuddy & Jacob, 1982; Dellarosa & Bourne,1986).Although in this literature, habituation is viewed as a hard-wired process(Dellarosa & Bourne, 1986), the encoding-accessibility hypothesis is in some sensescloser to our present understanding of retention for habituation training. As seen in theearlier discussion of habituation and long-term retention of habituation, habituation isnot a unitary, hard-wired process, but rather a form of plasticity which integratesinformation about stimulus familiarity and the organism’s internal state in sophisticatedways. The encoding-accessibility hypothesis permits, through the flexibility of theprocesses that control the engagement of encoding, the influence of organism-widestates on the process of memory consolidation.116This hypothesis predicts that anything that increases the likelihood of engagingmemory processes would tend to mask the effect of distributed training on memory,while manipulations that decrease the likelihood of engaging memory processes shouldemphasize the effect of distributed training (Dellarosa & Bourne, 1986). If so, it maybe possible to manipulate the expression of the distributed training effect by increasingor decreasing the likelihood of encoding.One possible way to increase the likelihood of encoding during habituation inC. elegans may be to alter the interstimulus interval. Habituation in . elegans ishighly sensitive to interstimulus interval; in recent work, there is evidence thatinterstimulus interval is encoded and that anticipatory responding may be observed inthe worm’s behavior (Rankin & Broster, 1992, Broster & Rankin, 1994; Wicks &Rankin, unpublished observations). If so, changing the interstimulus interval for shortperiods during massed training and distributed training may stimulate encoding andmask the distributed training effect on retention of habituation.Encoding of the stimuli may be depressed by an agent such as magnesiumchloride that causes reversible paralysis by affecting motor neuron synapses ininvertebrates. Touch neurons in the head are stimulated as the worm moves forwardthrough its environment (Kaplan & Horvitz, 1993). It is possible that blocking themovement of the worm would prevent such stimulation from triggering encoding. Sucha treatment maintained during training with monotonous stimulation should exaggeratethe distributed training effect on memory, although it may have an overall effect onmemory as well. No attempts have been made yet to habituate . elegans while it isunable to respond. This psychological perspective would provide a theoretical contextfor doing so.117While the long-term habituation expressed is not a large effect in the presentexperiments, the procedure does appear to consistently produce significant retention ofhabituation (see Experiments 2,4, 6,7, and 8). Building an understanding of thefactors affecting the processing of stimuli in . elegans by testing predictions based ontheories of memory may help us conceptualize the behavioral plasticity observed in .elegans as learning and memory. This theoretical approach makes the work on thecellular processes underlying these forms of behavioral plasticity in C. elegans morerelevant to our understanding of learning and memory phenomena.Differentiation of types of long-term memory through interference treatmentsStudies on the effects of protein synthesis inhibition and other interferencetreatments on memory consolidation have been used to differentiate between differentforms of memory in Drosophila. Tully et al. (1994) found that mid-term memory(lasting 7 h) for olfactory conditioning was vulnerable to anesthesia produced by coldshock (flies are immobilized by cold temperatures), while the longer lasting anesthesia-resistant memory (lasting 4 days) was attenuated in the radish memory mutant. Long-term memory (lasting over 7 days) induced by distributed training was disrupted bycycloheximide while long-term memory induced by massed training was not (Yin et al.,1994).The finding that LTM resulting from distributed training was blocked bycycloheximide while LTM resulting from massed training was not, suggests aninteresting experiment in C. elegans: the effects of the application of a protein synthesisinhibitor on memory resulting from distributed and massed training could be studied.If LTM for distributed training were blocked while LTM for massed training werespared, the contention that distributed and massed habituation training producequalitatively different types of memory would be substantiated. Contrasting the effects118of heat shock on the retention of massed and distributed training with the effectscycloheximide on the retention of the same might suggest whether heat shock has itseffect on long-term habituation through protein synthesis or through interference withsome other cellular process.If protein synthesis is induced during memory consolidation, one likelypathway for this induction is the cAMP second-messenger pathway, which has beenimplicated in learning in Aplysia (Schacher et aL, 1988; Dash et aL, 1990). Throughreverse genetics, the role of cAMP-responsive transcription in memory was examined.There is evidence that the transcriptional factor CREB may play a role in mediating thiscAMP-induced transcription during long-term facilitation in Aplysia, as an antagonist ofCREB prevented development of long-term facilitation (Dash et aL, 1990). In addition,work with rats has demonstrated that long-term potentiation (LTP), a form of long-lasting synaptic plasticity induced by tetanic stimulation, may be mediated by a cAMP-responsive transcriptional factor (Frey et al., 1993; Huang & Kandel, 1994;Bourtchuladze et al., 1994).The role of CREB in Drosophila was tested using the gene called dCREB2 thatexists in several naturally-occurring Drosophila CREB isoforms (Yin et al., 1994). Oneisoform, dCREB2-a, is a cAMP-dependent protein kinase A (PKA)-responsivetranscriptional activator, while another isoform, dCREB2-b, blocks PKA-responsivetranscription (Yin et al., 1994). Transcription is the process by which the DNA of thegenome is read and the corresponding RNA is generated, while translation is theprocess by which a protein molecule is constructed from the template of the RNAstrand (Watson, Hopkins, Roberts, Steitz & Weiner, 1987). The antagonistic isoformdCREB2-b was placed under the control of a heat-shock promoter, so that theexpression of the antagonistic dCREB2-b isoform could be induced by heat shock.119When heat shock was given to these transgenic flies three hours before training,the long-term memory for distributed olfactory conditioning was blocked, while it wasunimpaired in the transgenic strain without heat shock, and after heat shock in thematched control strain without the antagonistic dCREB2-b isoform (Yin et al., 1994).This experiment is an elegant demonstration of the role of the PKA-responsivetranscription pathway in memory. Significantly, the effects of the block of PKAresponsive transcription were measured in the behavior of intact animals. This factlends validity to their results.It is interesting to note that the heat shock treatment given 3 h before training inthe work on the role of CREB in memory consolidation to induce isoforms of theprotein did not affect learning and memory by itself (Yin et al., 1994). This finding fitsin with the present results, where it was observed that heat shock given 2 h beforetraining did not affect learning and memory in C. elegans.The approach of reverse genetics is one to which C. elegans is particularly wellsuited. The transgenic strain used in Experiment 3a to examine the expression of hspl6after various heat shock treatments is an example of a gene (in this case, iZ) insertedafter a heat shock promoter (the hspl6 promoter), so the jZ product, 8-galactosidase,is produced whenever hspl6 is induced by heat shock (Stringham et al., 1992). Aninteresting twist on the experiment performed by Tully and his colleagues (Yin et al.,1994) described above could be done employing the laser to heat shock individualneurons in adult worms (Stringham & Candido, 1993). With this technique, theantagonistic isoform of the CREB gene could be induced only in an individual neuronor subset of neurons thought to be the locus for structural changes responsible forlong-term memory. If induction of the antagonistic CREB gene blocked long-term memory120while heat shock of the same cells in genetically matched controls did not, this wouldbe strong evidence for the locus of structural changes due to learning in those neurons.Work with Aplysia has also revealed more than one type of retention fortraining based on interference studies using transcriptional and translational inhibitors(Ghirardi, Montarolo & Kandel, 1995). Long-term facilitation, a cellular analogue forlong-term sensitization, can be separated into two components; the weaker intermediateform (lasting 3 to 6 h) is blocked by anisomycin, a translational inhibitor, but notactinomycin, a transcriptional inhibitor. This memory process must require thesynthesis of proteins from RNA stored in the cell but not new manufacture of RNAfrom DNA. On the other hand, the long-term form (lasting over 24 h) is blocked byeither translational or transcriptional inhibitors, and thus requires both the synthesis ofproteins from RNA and the synthesis of RNA from DNA to form LTM (Ghirardi et al.,1995).This result opens the interesting possibility that other systems may also haveforms of memory that rely differentially on translation and transcription. Thispossibility could be tested in.elegans by comparing the effects of exposure to atranslational inhibitor with the effects of exposure to a transcriptional inhibitor duringand after training on the development of long-term memory for habituation.One of the common themes in the survey of literature in the introduction of thepresent work was the importance of examining factors affecting long-term memory inwhole intact animals, in the behavior as well as cellular analogues of the behavior, andin more than one form of learning. From the Drosophila work, it is clear that thecharacterization of the impact of the learning mutants on learning and memoryprocesses will require a synthetic approach to the results from a great variety of learningparadigms. This principle also applies to the investigation of the cellular processes121underlying LTM in . elegans. If we wish to characterize long-term memory and thecellular processes that mediate it in this organism, we must examine memory in as greata range of behavioral paradigms as possible. This is true because it is difficult topredict what value can be gained by studying a specific form of learning; aninvestigation that begins simply as a description of a specific form of learning mayproduce lines of research which lead to interesting insights or the development of newmethodologies that may be applied to many systems. But even more importantly, justas the study of learning and memory in every organism has its unique advantages,every form of learning offers the opportunity to examine a unique aspect of the animal’sbehavior. To achieve an understanding of the principles that guide learning and memoryand the cellular processes that mediate learning and memory, the results of studiesfocusing on a great range of behaviors must be intelligently synthesized.What other forms of behavioral plasticity does . elegans offer as choices forfuture research on the cellular processes underlying memory? There are manypossibilities, but a logical choice would be an examination of sensitization and theretention of sensitization in the tap withdrawal reflex. Sensitization to tap has beendemonstrated in C. elegans (Rankin et al., 1990). Generally speaking, sensitization isthought to play a critical role in habituation (Groves & Thompson, 1970). The cellularprocesses underlying sensitization have been implicated in classical conditioning inAplysia (Byrne et al., 1993). The results from an exploration of factors affectingsensitization and memory for sensitization in the tap withdrawal circuit may help tocharacterize the role of sensitization in the habituation and retention of habituationobserved in the same circuit.Interestingly, work examining habituation to tap in worms in which the anteriortouch cells were ablated through laser microsurgery so that the animals consistently122accelerate forward to tap has indicated that the acceleration response appears to be moresensitive to sensitization of the tap withdrawal reflex than the reversal response (Wicks,unpublished observations); while, as discussed earlier, the reversal response is moresensitive to the habituation of the tap withdrawal reflex. Already the comparison ofthese two forms of learning, habituation and sensitization, has shed some light on thenature of the relationship of the two competitive responses of the tap withdrawal reflexand perhaps on the circuit that supports it.The use of heat shock as an interference treatmentAs a tool to investigate the dynamics of memory consolidation, heat shock has somecritical advantages. First, it can be administered easily, without disturbing the organism, orintroducing any exogenous substances. Second, unlike cold shock and other anesthetics, itdoes not cause paralysis. C. elegans continues to move and forage during moderate heatshock, and is able to respond to tactile stimuli normally immediately after treatment. Third,unlike chemical agents that affect protein synthesis, heat shock can be administered for adefined, brief period of time. Fourth, heat shock evokes an active cellular response, which isinteresting in and of itself; but more importantly, the interaction of the cellular response to heatshock and the plasticity expressed in the tap withdrawal reflex may be considered a modelsystem for the examination of the interaction between organism-wide responses to changes inthe environment and plasticity specifically by the nervous system. Finally, the fmding that heatshock disrupts the formation of long-term memory means that we now have a fme-grained toolfor the investigation of the temporal parameters of memory consolidation and for the expansionof our understanding of the processes underlying memory.123List of AbbreviationsANOVA analysis of varianceATP adenylyl triphosphate°C degrees centigradecAMP cyclic adenylyl monophosphateC. elegans Caenorhabditis elegansCREB cAMP response element binding proteinDa DaltonE. jj Escherichia ijES effect sizeratio, a statistical test in the analysis of varianceh hourHAB mean of the responses to the last five stimuliHSPs heat shock proteinsHS group receiving heat shock treatmentINJT mean of the responses to the first five stimuliISI interstimulus intervalLTH group receiving long-term habituation trainingLTH-HS group receiving heat shock during trainingLTM long-term memoryLW long-term potentiationmm minuteMS mean sum of squaresnumber of subjects in groupN number of subjects in experiment124N Newtonnot statistically significantp probabilityPKA protein kinase APRE-HS group receiving heat shock that ended 2 h before trainingR responses secondS stimulusSE standard error of the meanSPS S Statistical Package for the Social Sciencestest, a statistical test comparing two meansx mean5-HT serotonin125BibliographyAceves-Pifla, E. 0., Booker, R., Duerr, J.S., Livingstone, M. S., Quinn, W. G.,Smith, R. F., Sziber, P. P., Tempel, B. L., & Tully, T. P. (1983). Learning and memory inDrosophila, studies with mutants. Cold Spring Harbor Symposia in Ouantitative Biology. 48,831 - 839.Bailey, C. H., & Chen, M. (1983). Morphological basis of long-term habituation andsensitization in Aplvsia. Science. 220. 91 - 93.Bailey, C. H., & Chen, M. (1988a). Long-term memory in Aplysia modulates the totalnumber of varicosities of single identified sensory neurons. Proceedings of the NationalAcademy of Sciences. USA, 85, 9356 - 9359.Bailey, C. H., & Chen, M. (1988b). Long-term sensitization in Aplysia increases thenumber of presynaptic contacts onto the identified gill motor neuron L7. Proceedings of theNational Academy of Sciences, USA. 85. 9356 - 9359.Bargmann, C. I. (1993). Genetic and cellular analysis of behavior in . elegans.Annual Review of Neuroscience. 16. 47 - 71.Barry, R. J., & Sokolov, E. N. (1993). Habituation of phasic and tonic components ofthe orienting reflex. International Journal of Psychophysiology. 15, 39 - 42.Barzilai, A., Kennedy, T. E., Sweatt, J. D., & Kandel, E. R. (1989). 5-HTmodulates protein synthesis and the expression of specific proteins during long-term facilitationin Aplsia sensory neurons. Neuron. 2, 1577 - 1586.Beck, C. D. 0., & Rankin, C. H. (1993). Effects of aging on non-associative learningin the nematode Caenorhabditis elegans. Behavioral Processes. 28. 145 - 164.Bicker, G., & Hahnlein, I. (1994). Long-term habituation of an appetitive reflex inthe honeybee. NeuroReport, , 54-56.Bourtchuladze, R., Frenguelli, B. Cioffi, D., Biendy, J., Schutz, G., & Silva, A. J.(1994). Normal short-term, but deficient long-term memory in mice with a targeted mutationof the cAMP Responsive Element Binding (CREB) protein. Cell. 79, 59 - 68.Braun, G. & Bicker, G. (1992), Habituation of an appetitive reflex in the honeybee.Journal of Neurophysiology. 67. 588 - 598.Brenner, S. (1974). The genetics of the nematode Caenorhabditis elegans. Genetics.- 94.Broster, B. S., & Rankin, C. H. (1994). Effects of changing interstimulus intervalduring habituation in Caenorhabditis elegans. Behavioral Neuroscience. 108. 1019 - 1029.126Byers, D., Davis, R. L., & Kiger, J. A. (1981). Defect in cyclic AMPphosphodiesterase due to the dunce mutations of learning in Drosophila melanogaster. Nature.289, 79 - 81.Byrne, J. H. (1982). Analysis of synaptic depression contribution to habituation ofgill-withdrawal reflex in Ap1sia californica. Journal of Neurophvsiologv. 48, 431 - 438.Byrne, J. H., Zwartjes, R., Homayouni, R., Critz, S. D., & Eskin, A. (1993). Rolesof second messenger pathways in neuronal plasticity and in learning and memory. In S.Shenolikar & A. C. Nairn (Eds.), Advances in Second Messenger and PhosphoproteinResearch. Vol. 27 (pp. 47 - 108). New York: Raven Press, Ltd.Candido, E. P., Jones, D., Dixon, D. K., Graham, R. W., Russnak, R. H., & Kay,R. J. (1989). Structure, organization and expression of the 16-kDa heat shock gene family ofCaenorhabditis elegans. Genome. 3, 690 - 697.Carew, T. J., Pinsker, H. M., & Kandel, E. R. (1972). Long-term habituation of adefensive withdrawal reflex in Aplysia. Science. 175. 451 - 454.Carew, T. J., & Kandel, E. R. (1973). Acquisition and retention of long-termhabituation in Aplysia: Correlation of behavioral and cellular processes. Science. 182. 1158 -1160.Castellucci, V. F., Blumenfeld, H., Goelet, P., & Kandel, E. R. (1989). Inhibitor ofprotein synthesis blocks long-term behavioral sensitization in the isolated gffl-withdrawal reflexof Aplysia. Journal of Neurobiology. 20, 1 - 9.Castellucci, V. F., & Kandel, E. R. (1974). A quantal analysis of syanptic depressionunderlying habituation of the gill-withdrawal reflex in Aplysia. Proceedings of the NationalAcademy of Sciences. USA. 71, 5004- 5008.Castellucci, V. F., Kennedy, T. E., Kandel, E. R., & Goelet, P. (1988). Aquantitative analysis of 2-D gels identifies proteins in which labeling is increased followinglong-term sensitization in Aplysia. Neuron. 1, 321- 328.Castellucci, V., Carew, T. J., & Kandel, E. R. (1978). Cellular analysis of long-termhabituation of the gill-withdrawal reflex of Aplysia californica. Science. 202, 1306 - 1308.Castelluci, V., Pinsker, H., Kupfermann, I., & Kandel, E. R. (1970). Neuronalmechanisms of habituation and dishabituation of the gill-withdrawal reflex in Aplvsia.Science. 167, 1745 - 1748.Cerbone, A., & Sadile, A. 0. (1994). Behavioral habituation of spatial novelty:interference and noninterference studies. Neuroscience and Biobehavioral Reviews. 18. 497 -518.Chalfie, M., & Sulston, J. (1981). Developmental genetics of the mechanosensoryneurons of Caenorhabditis elegans. Developmental Biology. 82, 358- 370.127Cheever, L. & Koshland, D. E. (1992). Retention of habituation in PC12 cells.Proceedings of the National Academy of Science. U.S.A.. 89. 10084 - 10088.Corfas, G., & Dudai, Y. (1989). Habituation and dishabituation of a cleaning reflex innormal and mutant Drosophila. Journal of Neuroscience. 9, 56 - 62.Corfas, G., & Dudai, Y. (1990). Adaptation and fatigue of a mechanosensory neuronin wild-type Drosophila and in memory mutants. Journal of Neuroscience. 10, 491 - 499.Cornell, E. H. (1980). Distributed study facilitates infants’ delayed recognitionmemory. Memory & Cognition. 8. 539 - 542.Cuddy, L. 3. & Jacoby, L. L. (1982). When forgetting helps memory: an analysis ofrepetition effects. Journal of Verbal Learning and Verbal Behavior. 21. 451 - 467.Dalley, B. K., & Golomb, M. (1992). Gene expression in the Caenorhabditis elegansdauer larva: developmental regulation of hsp9o and other genes. Developmental Biology. 151,80 - 90.Dash, P. K., Hochner, B., & Kandel, E. R. (1990). Injection of the cAMP-responsive element into the nucleus of Aplysia sensory neurons blocks long-term facilitation.Nature. 345. 718 - 721.Davis, H. P., Rosenzweig, M. R., Bennett, E. L., & Squire, L. R. (1980). Inhibitionof cerebral protein synthesis: Dissociation of nonspecific effects and amnesic effects.Behavioral and Neural Biology. 28, 99 - 104.Davis, H. P. & Squire, L. R. (1984). Protein synthesis and memory: A review.Psychological Review. 96, 518 - 559.Del Castillo, 3., & Katz, B. (1954). Statistical factors involved in neuromuscularfacilitation and depression. Journal of Physiology. 124. 574 - 585.Dellarosa, D. & Bourne, L. E. Jr. (1986). Surface form and the spacing effect.Memory & Cognition. 13. 529 - 537.Dudai, Y. (1988). Neurogenetic dissection of learning and short-term memory inDrosophila. Annual Review of Neuroscience. 11, 537 - 563.Dudai, Y., & Jan, Y.-N., Byers, D., Quinn, W. G., Benzer, S. (1976). dunce, amutant of Drosophila deficient in learning. Proceedings of the National Academy of Sciences.USA. 73, 1684 - 1688.Dudai, Y., & Zvi, S. (1983). Adenylate cyclase in the Drosophila memory mutantrutabaga displays an altered Ca2 sensitivity. Neuroscience Letters. 47, 119 - 124.Duerr, J. S., & Quinn, W. G. (1982). Three Drosophila mutations that blockassociative learning also affect habituation and sensitization. Proceedings of the NationalAcademy of Sciences. USA. 79, 3646 - 3650.128Dworkin, B. R., & Dworkin, S. (1990). Learning of physiological responses: I.Habituation, sensitization, and classical conditioning. Behavioral Neuroscience. 104, 298 -319.Eskin, A., Garcia, K. S., & Byrne, J. H. (1989). Information storage in the nervoussystem of Aplysia: specific proteins affected by serotonin and cAMP. Proceedings of theNational Academy of Sciences. USA. 86, 2458- 2462.Falk, C. X., Wu, C.-Y., Cohen, L. B., & Tang, A. C. (1993). Nonuniformexpression of habituation in the activity of distinct classes of neurons in the Aplysia abdominalganglion. Journal of Neuroscience. 13,4072-4081.Fire, A., White-Harrison, S., & Dixon, D. (1990). A modular set of !3-galactosidasefusion vectors for studying gene expression in Caenorhabditis elegans. Gene. 93, 189 - 198.Feige, U., & Polla, B. S. (1994). Heat shock proteins: the hsp70 family. Experientia., 979 - 986.Flexner, L. B., Flexner, J. B., & Roberts, R. B. (1966). Stages of memory in micetreated with acetoxycycloheximide before or immediately after learning. Proceedings of theNational Academy of Sciences. USA. 52, 1165 - 1169.Flood, J. F., Bennett, E. L., Orme, A. E., & Jarvik, M. E. (1977). Protein synthesisgradient of ECS retrograde amnesia. Behavioral Biology. 21, 307 - 328.Flood, J. F., Jarvik, M. E., Bennett, E. L., Orme, A. E., & Rosenzweig, M. R.(1977). The effects of stimulants, depressants, and protein synthesis inhibition on retention.Behavioral Biology. 20. 168 - 183.Frey, U., Krug, M., Reymann, K. G., & Matthies, H. (1988). Anisomycin, aninhibitor of protein synthesis, blocks late phases of LTP phenomena in the hippocampal CAlregion in vitro. Brain Research. 452, 57 - 65.Frost, W. N., Clark, G. A., Kandel, E. R. (1988). Parallel processing of short-termmemory for sensitization in Aplysia. Journal of Neurobiology. 19, 297 - 334.Ghirardi, M., Montarolo, P. G., & Kandel, E. R. (1995). A novel intermediate stagein the transition between short- and long-term facilitation in the sensory to motor neuronsynapse of Aplvsia. Neuron. 14. 413- 420.Gingrich, K. J., & Byrne, J. H. (1985). Simulation of synaptic depression, posttetanicpotentiation, and presynaptic facilitation of synaptic potentials from sensory neurons mdiatinggill-withdrawal reflex in Aplysia. Journal of Neurophvsiologv. 53. 652- 669.Gold, P. E. (1989). Neurobiological features common to memory modulation by manytreatments. Animal Learning and Behavior. 17, 94 - 100.Groves, P. M., & Thompson, R. F. (1970). Habituation: A dual-process theory.Psychological Review. 77. 419 - 450.129Groves, P. M., De Marco, R., & Thompson, R. F. (1969). Habituation andsensitization of spinal interneuron activity in acute spinal cat. Brain Research. 14, 521 - 525.Groves, P. M., Glanzman, D. L., Patterson, M. M., & Thompson, R. F. (1970).Excitability of cutaneous afferent terminals during habituation and sensitization in acute spinalcat. Brain Research. 18, 388 - 392.Groves, P. M., Lee, D. & Thompson, R. F. (1969). Effects of stimulus frequency andintensity on habituation and sensitization in acute spinal cat. Physiology and Behavior. 4, 383- 388.Hall, D. H., & Russell, R. L. (1991). The posterior nervous system of the nematodeCaenorhabditis elegans: serial reconstruction of identified neurons and complete pattern ofsynaptic interactions. Journal of Neuroscience. 11. 1 - 22.Hall, G. (1991). Perceptual and associative learning. Oxford: Clarendon Press.Harris, J. D. (1943). Habituatory response decrement in the intact organism.Psychological Bulletin. 40, 385 - 422.Hawkins, R. D., Castellucci, V. F., & Kandel, E. R. (1981). Interneurons involved inmediation and modulation of the gill-withdrawal reflex in Aplysia. I. Identification andcharacterization. Journal of Neurophvsiologv. 45, 304 - 314.Heisenberg, M. (1989). Genetic approach to learning and memory (mnemogenetics) inDrosophila melanogaster. In R. Rahman (Ed.), Fundamentals of memory formation: Neuronalplasticity and brain function (pp. 3 - 45). New York: Gustav Fischer Verlag.Hernández-Peón, R. (1960). Neurophysiological correlates of habituation and othermanifestations of plastic inhibition (internal inhibition). Electroencephalographv and ClinicalNeurophvsiologv. Supplement 13, 101 - 114.Heschl, M. F. P., & Baillie, D. L. (1990). The HSP7O family of Caenorhabditiselegans. Comparative Biochemistry and Physiology. 96B, 633 - 637.Hintzman, D. L. (1974). Theoretical implications of the spacing effect. In R. L.Solso, (ed.), Theories in cognitive psychology: the Loyola symposium. Hillsdale, NJ:Erlbaum.Hintzman, D. L., Block, R. A., & Summers, 3. 1. (1973). Modality tags and memoryfor repetitions: locus of the spacing effect. Journal of Verbal Learning and Verbal Behavior.fl, 229 - 239.Huang, Y. Y., & Kandel, E. R. (1994). Recruitment of long-lasting and proteinkinase A-dependent long-term potentiation in the CAl region of hippocampus requires repeatedtetanization. Learning & Memory (Cold Spring Harbor). 1. 74 - 82.130Huston, J. P., Mueller, C., & Mondadori, C. (1977). Memory facilitation by posttrialhypothalamic stimulation and other reinforcement. Neuroscience and Biobehavioral Review. 1143 - 150.Jordan, W. P. (1989). Mesencephalic reticular formation lesions made after habituationtraining abolish long-term habituation of the acoustic startle response in rats. BehavioralNeuroscience. 103, 805 - 815.Kaplan, J. M., & Horvitz, H. R. (1993). A dual mechanosensory and chemosensoryneuron in Caenorhabditis elegans. Proceedings of the National Academy of Sciences. USA.2fl 2227 - 2231.Klein, M., Shapiro, E., & Kandel, E. R. (1980). Synaptic plasticity and themodulation of the Ca current. Journal of Experimental Biology. 89, 117 - 157.Koroshetz, W. J. & Bonventre, J. V. (1994). Heat shock response in the centralnervous system. Experientia. 50, 1085 - 1091.Kuhl, D., Kennedy, T. E., Barzilai, A., & Kandel, E. R. (1992) Long-termsensitization training in Aplysia leads to an increase in the expression of BiP, the major proteinchaperone of the ER. Journal of Cell Biology. 119 (5). 1069 - 1076.Laming, P. R., & McKinney, S. J. (1990). Habituation in goldfish (Carassius auratus)is impaired by increased interstimulus interval, interval variability, and telencephalic ablation.Behavioral Neuroscience. 104, 869 - 875.Leaton, R. N., & Supple, W. F. (1991) Medial cerebellum and long-term habituationof acoustic startle in rats. Behavioral Neuroscience. 105 (6), 804 - 816.Lindman, H. R. (1991). Analysis of variance in experimental design. New York:Springer-Verlag.Lindquist, S. (1986). The heat-shock response. Annual Review of Biochemistry. 55,1151 - 1191.Livingstone, M. S., Sziber, P. P., & Quinn, W. G. (1984). Loss ofcalcium/calmodulin responsiveness in adenylate cyclase of rutabaga. a Drosophila learningmutant. Cell. 37. 205 - 215.Lozado, M., Romano, A., & Maldonado, H. (1990). Long-term habituation to adanger stimulus in the crab Chasmagnathus granulatus. Physiology and Behavior. 47, 35 - 41.Lukowiak, K. (1978). L9 modulation of gill-withdrawal reflex habituation in Aplysia.Journal of Neurobiology. 10. 255 - 271.Marion, T. M., Beck, C. D. 0. & Rankin, C. H. (1992). An exploration of long-termmemory in. elegans. Society for Neuroscience Abstracts. 18, 532.131McFadden, P. N., & Koshland, D. E. Jr., (1990a) Habituation in the single cell:diminished secretion of norepinephrine with repetitive depolarization of Pci2 cells.Proceedings of the National Academy of Science. U.S.A.. 87. 2031 - 2035.Mcfadden, P. N., & Koshland, D. E. Jr. (1990b) Parallel pathways for habituationin repetitively stimulated PC12 cells. Neuron. 4. 615 - 621.Melton, A. W. (1970). The situation with respect to the spacing of repetitions andmemory. Journal of Verbal Learning and Verbal Behavior. 9, 596- 606.Menzel, R., Hammer, M., & Sugawa, M. (1989). Non-associative components ofconditioning in honeybees. In J. Erber, R. Menzel, H.-J. Pflueger, & D. Todt (Eds.), Neuralmechanisms of behavior (p. 221). FRG: Thieme.Menzel, R., Sugawa, M., Wittstock, S., & Braun, G. (1989). chemical codes oflearning and memory in honey bees. In J. Erber, R. Menzel, H. Pfluger, & D. Todt (Eds.),Neural mechanisms of behavior (pp. 139 - 140). Stuttgart: Thieme.Mizumori, S. J. Y., Rosenzweig, M. R., & Bennett, E. L. (1985). Long-term workingmemory in the rat: effects of hippocampally applied anisomycin. Behavioral Neuroscience. 99,220 - 232.Mondadori, C., Huston, J. P., Ornstein, K., & Waser, P. G. (1976). Posttrialreinforcing hypothalamic stimulation can facilitate avoidance learning. Neuroscience Letters. 2,183 - 187.Montarolo, P. G., Goelet, P., Castellucci, V. F., Morgan, J., Kandel, E. R., &Schacher, S. (1986). A critical period for macromolecular synthesis in long-termheterosynaptic facilitation in Aplysia. Science. 234. 1249- 1254.Montarolo, P. G., Kandel, E. R., & Schacher, S. (1987). Long-term heterosynapticdepression of the sensorimotor synapses in culture is produced by the peptide FMRF-amide.Society for Neuroscienee Abstracts. 13, 391.Morimoto, R. I., Tissieres, A., & Georgopoulos, C. (Eds.) (1990). Stress proteins inbiology and medicine. Cold Spring Harbor: Cold Spring Harbor Press.Nakashima, M., Yamada, S., Shiono, S., Maeda, M., & Satoh, F. (1992). 448-detector optical recording system: development and application to Aplysia gill-withdrawalreflex. IEEE Transactions of Biomedical Engineering. 39, 26- 36.Nowak, T. 5. (1993). Synthesis of heat shock/stress proteins during cellular injury.Annals of the New York Academy of Sciences. 679, 142 - 156.O’Dell, K. M. C. (1994). The inactive mutation leads to abnormal experience-dependent courtship modification in male Drosophila melanogaster. Behavioral Genetics. 24,381 - 384.132Papa, M., Pellicano, M. P., Weizi, H., & Sadile, A. G. (1993). Distributed changesin c-Fos and c-Jun immunoreactivity in the rat brain associated with arousal and habituation tonovelty. Brain Research Bulletin. 32. 509 - 515.Parsell, D. A., Taulien, J., & Lindquist, S. (1993). The role of heat-shock proteins inthermotolerance. Philosophical Transactions of the Royal Society of London. Series B. 339,279 - 286.Peeke, H. V. S., & Petrinovich, L. (Eds.). (1984). Habituation. sensitization andbehavior. New York: Academic Press.Perruchet, p. (1989). The effect of spaced practice on explicit and implicit memory.British Journal of Psychology. 80. 113 - 130.Petrinovich, L. (1984). A theory of habituation and sensitization. In H. V. S. Peeke &L. Petrinovich (Eds.), Habituation. sensitization and behavior (pp. 17 - 55). New York:Academic Press.Pietsch, P., & Schneider, C. W. (1990). Anterior decerebration blocks visualhabituation in the larval salamander (Arnbvstoma punctatum). Brain Research Bulletin. 25, 613- 615.Pilcher, C. W. T., & Booth, D. A. (1979). Effect of cycloheximide on the long-termretention of reversed paw-preference in the rat. Experimental Brain Research, 23, 161.Postman, L. & Knecht, K. (1983). Encoding variability and retention. Journal ofVerbal Learning and Verbal Behavior. 8. 133 - 152.Rankin, C. H. (1991). Interaction between two antagonistic reflexes in the nematodeCaenorhabditis elegans. Journal of Comparative Physiology A. 169, 59- 67.Rankin, C. H., Beck, C. D. 0., & Chiba, C. M. (1990). Caenorhabditis elegans: Anew model system for the study of learning and memory. Behavior and Brain Research. 37.89 - 92.Rankin, C. H., & Broster, B. S. (1992). Factors affecting habituation and recoveryfrom habituation in the nematode Caenorhabditis elegans. Behavioral Neuroscience. 106. 239 -242.Ratikin, A., Tomsic, D., & Maldonado, H. (1991). Habituation and sensitization to anelectrical shock in the crab Chasmagnathus: Effect of background illumination. Physiology &Behavior. 50, 477 - 487.Rose, R. J. (1980). Encoding variability, levels of processing, and the effects ofspacing upon judgments of frequency. Memory & Cognition. 18. 84 - 93.Rose, R. J. & Rowe, E. J. (1976). Effects of orienting task, spacing of repetitions,and list context upon judgments of frequency. Journal of Experimental Psychology: HumanLearning & Memory. 2. 142 - 152.133Rosenthal, R., & Rubin, D. B. (1994). The counternull value of an effect size: A newstatistic. Psychological Science. 5. 329 - 334.Ross, B. H., & Lindauer, T. K. (1978). Memory for at least one of two items: testand failure of retrieval theories of spacing effects. Journal of Verbal Learning and VerbalBehavior. 17. 669 - 680.Russnak, R. H., & Candido, P. M. (1985). Locus encoding a family of small heatshock genes in Caenorhabditis elegans: two genes duplicated to form a 3.8-kilobase invertedrepeat. Molecular and Cellular Biology. 5(6). 1268 - 1278.Sadile, A. G., Cerbone, A., & Cioffi, L. A., (1981). Vasopressinergic modulationof behavioral habituation. Dose-response curves in rats of the NSD-HE and NSD-LE strains.Neuroscience Letters. 7 (Supplement). 51.Sadile, A. G., Cerbone, A., Lamberti-D’Mello, Amoroso, S., Annunziato, L., Menna,T., Buono, C., Rafti, F., & Giuditta, A. (1995). The dorsal noradrenergic bundle modulatesDNA remodeling in the rat brain upon exposure to a spatial novelty. Brain Research Bulletin.32, 9 - 16.Schacher, S., Montarolo, P. G., Kandel, E. R., Chen, M., & Bailey, C. H. (1991).Inhibitors of protein and RNA synthesis block structural changes accompanying long-termpresynaptic facilitation and inhibition in Aplysia sensory neurons. Society for NeuroscienceAbstracts. 17, 1301.Schlesinger, M. J., Tissieres, A., & Ashburner, M. (Eds.). (1982). Heat shockproteins: From bacteria to man. Cold Spring Harbor, NY: Cold Spring Harbor Press.Segal, D. S., Squire, L. R., & Barondes, S. H. (1971). Cycloheximide: Its effects onactivity are dissociable from its effects on memory. Science. 172, 82 - 84.Siegel, S. (1956). Nonparametric statistics for the behavioral sciences. New York:McGraw-Hill.Shaughnessy, J. 3. (1976). Persistence of the spacing effect in free recall under varyingincidental learning conditions. Memory and Cognition. 4. 369 - 377.Sharp, F. R. & Sagar, S. M. (1994). Alterations in gene expression as an index ofneuronal injury: heat shock and the immediate early gene response. NeuroToxicitv. 15, 51 -60.Snutch, T. P. & Baillie, D. L. (1983). Alterations in the pattern of gene expressionfollowing heat shock in the nematode Caenorhabditis elegans. Canadian Journal of CellBiology. 61, 480 - 487.Squire, L. R., & Barondes, S. H. (1974). Anisomycin, like other inhibitors of cerebralprotein synthesis, impairs “long-term” memory of a discrimination task. Brain Research. 66,301- 308.134Squire, L. R., Emanuel, C. A., Davis, H. P., & Deutch, J. A. (1975). Inhibitors ofcerebral protein synthesis: Dissociation of aversive and amnesic effects. Behavioral Biology..1.4, 335 - 341.Squire, L. R., Geller, A., & Jarvik, M. E. (1970). Habituation and activity as affectedby cycloheximide. Communications in Behavioral Biology. 5, 249 - 254.Staddon, J. E. R. (1993). On rate-sensitive habituation. Adaptive Behavior. 1, 123 -131.Staubli, U., Faraday, R., & Lynch, G. (1985). Pharmacological dissociation ofmemory: anisomycin, a protein synthesis inhibitor, and leupeptin, a protease inhibitor, blockdifferent learning tasks. Behavioral and Neural Biology. 43, 287 - 297.Stopfer, M., & Carew, T. J. (1994). Homosynaptic depression in tail sensory neuronsis not the mechanism of habituation of tail-induced tail or siphon withdrawal. Society forNeuroscience, 20, 1073.Stringham, E. P., Dixon, D. K., Jones, D., & Candido, E. P. (1992). Temporal andspatial expression patterns of the small heat shock (Hspl6) genes in transgenic Caenorhabditiselegans. Molecular Biology of the Cell. 3, 221 - 233.Stringharn, E. G., & Candido, E. P. (1993). Targeted single-cell induction of geneproducts in Caenorhabditis elegans: a new tool for developmental studies. Journal ofExperimental Zoology. 266, 227 - 233.Sulston, J. E., & White, J. 0. (1980). Regulation and cell autonomy duringpostembryonic development of Caenorhabditis elegans. Developmental Biology. 78, 542 -576.Tempel, B. L., Bonini, N., Dawson, D. R., & Quinn, W. G. (1983). Reward learningin normal and mutant Drosophila. Proceedings of the National Academy of Sciences. USA,80, 1482 - 1486.Thompson, R. F., & Spencer, W. A. (1966). Habituation: A model phenomenon forthe study of neuronal substrates of behavior. Psychological Review. 73, 16 - 43.Toppino, T. C., & DiGeorge, W. (1984). The spacing effect in free recall emergeswith development. Memory & Cognition. 12. 118 - 122.Toppino, T. C., Kasserman, J. E., & Mracek, W. A. (1991). The effect of spacingrepetitions on the recognition memory of young children and adults. Journal of ExperimentalChild Psychology. 51. 123- 138.Trudeau, L. E., & Castellucci, V. F. (1992). Contribution of polysynaptic pathways inthe mediation and plasticity of Aplysia gill and siphon withdrawal reflex: evidence fordifferential modulation. Journal of Neuroscience. 12, 3838 - 3848.135Tully, T., Preat, T., Boynton, S. C., Del Vecchio, M. (1994). Genetic dissection ofconsolidated memory in Drosophila. Cell. 79, 35 - 47.Watson, J. D., Hopkins, N. H., Roberts, J. W., Steitz, I. A., & Weiner, A. M.(1987). Molecular biology of the gene. Memo Park, CA: Benjamin I Cummings PublishingCo.Weinberger, N. M., Goodman, D. A., & Kitzes, L. M. (1969). Is behavioralhabituation a function of peripheral auditory system blockade? Communications in BehavioralBiology. 3 (Part A), 111- 116.White, J. G., Southgate, E., Thomson, J. N., & Brenner, S. (1986). The structure ofthe nervous system of Caenorhabditis elegans. Philosophical Transactions of the Royal Societyof London, B. Biological Sciences. 314. 1-340.Wicks, S. R., & Rankin, C. H. (1995). Integration of mechanosensory stimuli inCaenorhabditis elegans. Journal of Neuroscience. 15. 2434 - 2444.Wittekind, W. C., & Spatz, H.-C. (1988). Habituation of the landing response ofDrosophila. In G. Hertting & H.-C. Spatz (Eds.), Modulation of sYnaptic transmission andplasticity in the nervous system (pp. 351 - 368). New York: Springer Verlag.Wittstock, S., Kaatz, H. H., Menzel, R. (1993). Inhibition of brain synthesis bycycloheximide does not affect formation of long-term memory in honeybees after olfactoryconditioning. Journal of Neuroscience. 13, 1379 - 1386.Wood, D. C. (1988). Habituation in Stentor: a response-dependent process. Journal ofNeuroscience. 8, 2248 - 2253.Wood, W. B. (Ed.). (1988). The nematode Caenorhabditis elegans. Cold SpringHarbor: Cold Spring Harbor Laboratory.Yin., 3. C. P., Wallach, J. S., Del Vecchio, E. L., Wilder, E. L., Zhou, H., Quinn,W. G., & Tully, T. (1995). Induction of a dominant-negative CREB transgene specificallyblocks long-term memory in Drosophila. Cell. 79, 49 - 58.Zecevic, D., Wu, J.-Y., Cohen, L. B., London, J. A., Hopp, H.-P., & Fallc, C. X.(1989). Hundreds of neurons in the Aplysia abdominal ganglion are active during the gill-withdrawal reflex. Journal of Neuroscience. 9, 3681 - 3689.Zucker, R. S. (1989). Short-term synaptic plasticity. Annual Review of Neuroscience.U, 13 - 31.136Appendix 1Analysis of the Distribution of Missing Values in the Habituation DataBecause there are a sizable number of missing values due to the presence of theacceleration response to tap in the habituation data, it is appropriate to consider how thedistribution of missing values affects the interpretation of reversal response data. Therelationship between the habituation of acceleration responses to tap and the habituation of thereversal responses to tap, which was discussed in the introduction of the thesis, bears on thisissue. There, it was argued that because of the relationship between the two responses,reversals and accelerations, the habituation of the magnitude of the reversal response reflectsthe habituation of both responses (Wicks and Rankin, unpublished observations). In thisanalysis, the distribution of missing values will be examined for characteristics that may help tointerpret this relationship between the habituation of the acceleration and reversal responses.In the present experiments, the overall percentage of missing data values was 26.5%.Fewer than 1% of the responses to stimuli were not scored due to technical errors in stimulusadministration or video recording, so the percentage of missing data values represents theproportion of acceleration responses to stimuli. The proportion of accelerations expressed mayvary across training and between groups and experiments. Wicks and Rankin (unpublishedobservations) found that the proportion of accelerations varied significantly during habituationtraining in normal worms and that the proportion of accelerations was significantly higheroverall during habituation training at a 10-s ISI than at a 60-s 1ST; these findings support thepossibility that the distribution of missing values may vary significantly over habituationtraining in the present experiments. Therefore, the consequences of uneven distributions ofaccelerations must be explored.One concern is that the distribution of accelerations within blocks will vary duringtraining on Day 1 or between training and testing days, affecting the expression of reversal137responses. For example, if in the first block, accelerations occurred late in the habituation run,displacing relatively short reversal responses; but on Day 2, accelerations occurred early in thehabituation run, displacing relatively long reversal responses, this may affect the mean blockmagnitudes of Day 1 and Day 2. Another possibility is that the number of accelerations willdiffer significantly between groups or experiments, thus affecting the data. The analyses thatfollow address these concerns.This analysis is intended to investigate overall patterns in the distribution of missingvalues. Consequently, only a little emphasis will be placed on individual significant results. Inaddition, the alpha level will not be adjusted when multiple comparisons are done; multiplecomparisons are limited as no strong hypotheses as to what patterns may be evident wereformed a priori.Experiment 1: Short- and lone-term habituation with a short interstimulus interval (10-s 1SF)The distribution of missing values was uneven between the training groups— overall,the massed training group had a greater proportion of missing values than the other groups,while the single stimulus control group had fewer (see Fig. 20A; distributed vs. massedtraining groups, number of missing values across all 80 stimuli, unpaired, two-tailed test:(40) = 2.911, < .01; Fig. 20B; distributed, massed training groups, twenty stimuli control,number of missing values across the first 20 stimuli on Day 1 and the 20 stimuli on Day 2,factorial ANOVA: f(2, 60) = 7.043, < .01; Fig. 20C; distributed, massed training groups,twenty stimuli and single stimulus controls, number of missing values from the twenty Day 2stimuli, factorial ANOVA: E(3, 79) = 11.998, p < .01). The unusually high proportion ofaccelerations in the first twenty stimuli of the massed training group is difficult to interpret, asthere were no differences in the treatment that worms in the distributed and massed traininggroups, as well as the twenty stimuli control group received in the first twenty stimuli.138Figure 20. The number of missing values in Experiments I (10-s ISI) and 2 (60-s ISI). A.The number of missing values across all training and testing in the massed (60 stimulitogether on Day 1) and distributed (60 stimuli, in 3 blocks of 20 stimuli on Day 1)habituation training. B. The number of missing values totaled over the first 20 stimuli givenon Day 1 and the 20 stimuli on Day 2 in the twenty stimuli control group (20 stimuli onDay 1), massed and distributed habituation training groups. C. The number of missingvalues in the 20 stimuli given on Day 2 in the single stimulus control group (one stimuluson Day 1), the twenty stimuli control group, and the massed and distributed habituationtraining groups.139MISSING VALUES DAY 1 AND DAY 2EXP. 1 AND EXP. 210-s ISIK 60-s ISIMASSED DISTA80-Cl)LU0.J 60->cJ0 c 40-20-0-B 40Cr,Ui. 30> 20-10-41:C 20-Cl)LUD-J 0 15->..—a.., 10CI)41: 0LJ 1 0-s ‘SIK 60-s 151MASSED DIST TWENTY10-s ‘SI60-s (SIMASSED DIST TWENTY SINGLE140When the proportion of missing values were considered across three blocks of 20stimuli on Day 1 in the distributed and massed training groups, there was no significant changein the proportion of missing values across the blocks; however, there was a difference betweenthe distributed and massed training groups, and a significant interaction between block andgroup (see Fig. 21A; mixed design ANOVA, Block: f(2, 80) = .49, n&; group: E(1, 40) =8.993, p < .01; Block x Group: E(2, 80) = 5.63, p < .01). Interestingly, none of the threetrained groups, distributed, massed or twenty stimuli control, showed a change in the numberof accelerations when the number of missing values in the data from the first 20 stimuli on Day1 were compared to those from the 20 stimuli on Day 2 (Fig. 22A; paired, two-tailed tests;distributed: 1(20) = 1.671, massed: (20) = .677, n.&; twenty stimuli control: 1(20) =.681, u.).Within blocks of training, the distribution of missing values during habituation wasconsidered by separating the responses to the 20 stimuli into four sets of five stimuli (stimuli 1- 5, 6 - 10, 11 - 15, 16 - 20). The number of missing values to the first 20 stimuli on Day 1and the 20 stimuli on Day 2 were analyzed in the distributed and massed training groups usingtwo-way repeated measures ANOVAs. In the massed training group, the number of missingvalues changed significantly during habituation, while in the distributed training group, thenumber of missing values did not vary significantly during training (Fig. 23A and 24A;massed: (3, 60) = 7.65, p < .01; distributed: E(3, 60) = 2.08, i&). The day of training didnot significantly affect the number of missing values in either group (massed: E(1, 20) = .46,n&; distributed: f(l, 20) = 2.79, nj, nor was there an interaction between habituationstimuli and day of training (massed: E(3, 60) = .57, n..&; distributed: E(3, 60) = 1.01, nJ.In summary, there were significant differences between groups in the total number ofmissing values, but no differences in the number of missing values between the two days oftraining. The differences between the groups may affect the interpretation of the results from141Figure 21. The number of missing values in the three blocks of training (20 stimuli each)on Day 1 in the massed habituation training group (60 stimuli given together on Day 1) andthe distributed habituation training group (60 stimuli given in three blocks of 20 stimulieach). A. Experiment 1: 10-s ISI. B. Experiment 2: 60-s ISI. Error bars indicate ÷1-SEMs.0Cl)UiD-J>0zCl)Cl)4$:MISSING VALUES ON DAY 1MASSED AND DISTRIBUTEDEXP. 110-s ISI142TRAININGU MASSEDI DISTA0csJCl)UiD-J>0zCl)Cl)4$:B20-15-10-5-0-20—1 5-1050BLOCK 1 BLOCK 2 BLOCK 3EXP. 260-s ISIMASSEDI DISTBLOCK 1 BLOCK 2 BLOCK 3143Figure 22. The number of missing values on Day 1 (first 20 stimuli of training) and on Day2 of the massed (60 stimuli given together on Day 1), distributed (60 stimuli given in 3blocks of 20 stimuli on Day 1) and the twenty stimuli control group (20 stimuli on Day 1).A. Experiment 1: 10-s 1ST. B. Experiment 2: 60-s 1ST. Error bars indicate +1- SEMs.A0CJCl)UiD-J4>C’,zCl)Co#4:0c..JCoLUD-j4>C!,zCoCl)44:MISSING VALUES DAY 1 AND 2MASSED, DISTRIBUTED TRAINING,TWENTY STIMULI CONTROL144EXP. 110-s ISIEJ DAY1DAY2MASSED 01ST TWENTYBEXP. 260-s ISI1 J DAY1I DAY250MASSED DIST TWENTY145Figure 23. The distribution of missing values during distributed habituation training (60stimuli in three blocks of 20 stimuli) on Day 1 (first 20 stimuli of training) and on Day 2(20 stimuli retention test). The number of missing values in five stimuli bins (S 1-5, s6-10,si 1-15, and s16-20) was considered. A. Experiment 1: 10-s 1ST. B. Experiment 2: 60-s1ST. Error bars indicate +1- SEMs.U)Cl)UiD-J>zCl)Cl)0zCl)Cl)DISTRIBUTION OF MISSING VALUESDURING DISTRIBUTED TRAINING ONDAY 1 AND DAY 2•DAY2146AEXP. 110-s ISIDAY1DAY2FSi - 55-4-3.21•0•5.4.3.21-0S6-1O S11-15 S16-20EXP. 260-s ISIBU)Cl)UiD-J>IJ DAY ISi - 5 S6 - 10 Sli - 15 S16 - 20147Figure 24. The distribution of missing values during massed habituation training (60stimuli given together) on Day 1 (first 20 stimuli of training) and on Day 2(20 stimuliretention test). The number of missing values in five stimuli bins (sl-5, s6-10, sl 1-15,and s16-20) was considered. A. Experiment 1: 10-s 1ST. B. Experiment 2: 60-s 1ST. Errorbars indicate +1- SEMs.LI)Cl)UiD-J>C.,zCl)Cl)LI)Cl)UiD-J>C.,zCl)Cl)DISTRIBUTION OF MISSING VALUESDURING MASSED TRAINING ONDAY 1 AND DAY 2148A EXP. 110-s ISI5—4-3-2-•1—0-DAY1DAY2BS1-5 S6-10 S11-15 S16-20EXP.60-s2ISILI DAY 1DAY2Si - 5 S6 - 10 Sil - 15 S16 - 20149direct comparisons between the groups. The findings that the day of training did not affect thetotal number of missing values or the distribution of missing values during habituation withineach block, taken together, indicate that whatever differences are apparent in the levels ofresponding on Day 1 and Day 2 in this experiment cannot be attributed to changes in thefrequency of acceleration responses. Thus, the presence of accelerations in the data does notlimit the interpretation of differences within groups in the reversal response magnitude data.Experiment 2: Short- and long-teim habituation with a long interstimulus interval (60 sThe distribution of missing values was found to be even across groups (Fig. 20B;distributed vs. massed, number of missing values in all 80 stimuli: unpaired, two-tailed i test,(39) = 1.75 1, j; Fig. 20B; distributed, massed and twenty stimuli control, number ofmissing values in the first 20 stimuli on Day 1 and the 20 Day 2 stimuli: factorial ANOVA,E(2, 58) = 1.941, n.&; Fig. 20C; distributed, massed, twenty stimuli control, and singlestimulus control, number of missing values in the 20 Day 2 stimuli: factorial ANOVA, E(3, 76)= 2.544,When the distributions of missing values were considered across three blocks of 20stimuli on Day 1 in the distributed and massed training groups, a significant change across theblocks was observed, but no difference between the groups and no interactions were evident(see Fig. 21B; mixed design ANOVA, Blocks 1 to 3: E(2, 78) = 5.374, < .01; groups[distributed vs. massed]: E(1, 39) = 1.026, The massed training group and twentystimuli control group both showed a significantly higher number of missing values on Day 2than in the first 20 stimuli on Day 1 (Fig. 22B; paired, two-tailed tests; massed: (19) =3.337, < .01; twenty stimuli control: 1(19) = 3.15 1, < .01), while the distributed group didnot (distributed: (20) = .81, n.&).Within blocks of training, the distribution of missing values during habituationwas considered by separating the 20 stimuli in a training block into four sets of five150stimuli (stimuli 1 - 5, 6 - 10, 11 - 15, and 16 - 20). The first 20 stimuli on Day 1 andthe 20 stimuli on Day 2 were analyzed in the distributed and massed training groupsusing mixed design ANOVAs. In the distributed group, there was no significant effectof training during habituation while in the massed training group, the number ofmissing values changed significantly during habituation (Fig. 23B and 24B; distributed:E(3, 60) = 2.68, j; massed: (3, 57) = 6.06, < .01). In the distributed group,there was no significant change in the number of missing values from Day 1 to Day 2,while in the massed group there was a significant increase in the number of missingvalues on Day 2 (distributed: E(1, 20) = .66, n; massed: E(1, 19) = 11.14, < .01).Most importantly, however, in neither group was there a significant interaction betweenhabituation stimuli and day of training (distributed: f(3, 120) = .40, n; massed: E(3,114) = .18, jj). As can be seen in Figures 22 and 23, the distribution of missingvalues during habituation follows the same pattern on Day 1 and 2 in both thedistributed and massed training groups.In summary, there were no differences between groups in the proportion ofmissing values overall; however, there were differences between Day 1 and Day 2 intwo groups, and there was a significant pattern in the distribution of missing valuesduring habituation within a block in the massed training group. To counter thisfinding, the absence of a significant interaction between day of training and habituationstimuli indicates that the pattern of missing value distribution during habituation withinblocks is the same on Day 1 and 2 in both groups. This is a critical point, as itdemonstrates that any difference in the response levels between Day 1 and 2 exhibitedby these groups cannot be attributed to a change in the distribution of accelerationresponses during habituation within blocks. In addition, as all trained groups exhibitedan increase in the proportion of accelerations on Day 2 over Day 1, so any difference151between the groups in the change in level of responding from Day 1 to Day 2 cannot beattributed to the change in the proportion of accelerations. Thus, although theproportion of accelerations did vary during training, the presence of missing values inthe data does not limit the interpretation of the reversal response magnitude data in Exp.2.The effects of three types of heat-shock treatments on the distribution ofmissing values during habituation on Day 1 and Day 2 require examination: heat shockbefore training on Day 1, as used in Experiments 4 and 6; heat shock during training,as used in Experiments 4, 6 and 8; and heat shock before testing on Day 2, as used inExperiment 7.Experiment 4: the effects of heat shock (45 mm. 32°C) before training on thedistribution of missing valuesAn examination of the distribution of missing values between groups LTH andPRE HS / LTH, the days of training within those groups, the blocks of training on Day1, and within the first block on Day 1 and the test block on Day 2 showed that heatshock (45 mm, 32°C) before training had no effect on the propensity to accelerate totap. The distribution of missing values was not affected by the day of training in eithergroup (see Fig. 25A; mixed design ANOVA; Group: E(1, 38) = .120, n&; Day: E(1,38) = .996, n; Group x Day: E(1, 38) = .008, n.&). The mean number of missingvalues did not change across the three blocks of training on Day 1 for either group (seeFig. 25B; mixed design ANOVA; Group: E(1, 76) = 1.599, Block: (2, 38) =.252, n; Group x Block: E(2, 76) = .658, n).Finally, the distribution of missing values within blocks of training wasconsidered by separating the responses to the 20 stimuli within blocks of training intofour sets of five stimuli (stimuli 1 - 5, 6 - 10, 11 - 15, and 16 - 20). When the first152Figure 25. The number of missing values in Experiment 4: pre-exposure to heat shock. A.the number of missing values on Day 1 (first 20 stimuli of training) and on Day 2 (the 20stimuli of the retention test) in the LTH group (distributed training at a 60-s 1ST) and thePRE HS I LTH group (distributed training with pre-exposure to heat shock (45 mm, 32°C)ending 2 h before training begins). B. The number of missing values for the three blocks oftraining on Day 1 (each 20 stimuli) in the LTH and PRE HS I LTH groups. Error barsindicate +1- SEMs.0C4CoLiiD-j>CDzCOCoB 200csJCoLiiD-j>CDzCoCl)4*:A153EXPERIMENT 3: PRE-EXPOSURETO HEAT SHOCKDAY 1 AND DAY 2DAY2LTH PREHS/LTHBLOCKS ON DAY 11510S0D BLK 1BLK 2BLK 3LTH PREHS/LTH154Figure 26. The distribution of missing values during training in the LTH and PRE HS /LTH groups on Day 1 (first 20 stimuli of training) and on Day 2 (20 stimuli retention test).The number of missing values in five stimuli bins (sl-5, s6-l0, si 1-15, and s16-20) wasconsidered. A. LTH group (distributed training at a 60-s ISI). B. PRE HS / LTH group(distributed training at a 60-s ISI preceded by a heat shock (45 mm, 32°C). Error barsindicate +1- SEMs.U)Cl)uJD-J>C.,zC,,Cl)AMISSING VALUES AFTERPRE-EXPOSURE TO HEAT SHOCKEXPERIMENT 4155LTHE:rDAY2S1-5 S6-iO Sii-15 S16-20PRE HS I LTH4—3-2-1—0-. DAY1DAY21Si-5 S6-i0 Sii-15 Si 6-20156block of training on Day 1 and the test block on Day 2 were examined, the distributionof missing values was found to be even within each block; furthermore, no significantdifferences were seen between days in either group (see Fig. 26A & B; mixed designANOVA; Day: E(1,38) = 1.00, n.&; Mauchly’s sphericity test, chi-squareapproximation = 11.85, <.05, Huynh-Feldt epsilon = .94; Training: E(2.82, 107.16)= .945, n.; Day x Training: E(3, 114) = .18, Group: E(1,38) = .12, Groupx Day: E(1,38) = .01, ; Group x Day x Training: E(3, 114) = 2.00, Thecomparisons of the two groups of Experiment 4, LTH and PRE HS I LTH provideevidence that heat shock before training does not affect the propensity to accelerate totap.ExperimentS: the effects of heat shock (45 mm. 32°C) during training on thedistribution of missing valuesIt is possible that the distribution of missing values during habituation changeswith the presentation of heat shock during training and that this change affects theexpression of habituation on Day 1 or on Day 2. This possibility was examined by ananalysis of the distribution of missing values on Day 1 across blocks of training andbetween Day 1 (Block 1) and Day 2 of the LTh-HS group in Experiment 5 and theLTH of Experiment 2 (distributed training group, 60-s ISI) with mixed designANOVAs (see Fig. 27A & B). In the analysis of the distribution of missing values onDay 1, there was no significant overall difference in the number of missing valuesbetween the LTH-HS and the LTH groups (Group: E(1, 37) = .59, Across thegroups, the number of missing values between blocks changed significantly (Block:E(2, 74) = 3.96, = .02); however, the distribution of missing values across blockswas not significantly different between the groups (Block x Group: E(2, 74) = 1.35,jj. The distribution of missing values was uneven with the blocks of training157Figure 27. LTH-HS (Exp. 5) and LTH (Exp. 2; distributed training) and the distribution ofmissing values during habituation training on Day 1. The number of missing values in fivestimuli bins (sl-5, s6-10, sll-15, and s16-20) was considered. A. LTH group(Experiment 2, distributed training at a 60-s 1ST) across three blocks of training on Day 1.B. LTH-HS group (Experiment 5, distributed training at a 60-s 1ST, with heat shock (45mm, 32°C) between blocks of training) across three blocks of training on Day 1. Errorbars indicate +1- SEMs.158DISTRIBUTION OF MISSING VALUESDURING TRAINING ON DAY 1•A 5 LTH (EXP. 2)JBLK1b-..ØBLK2D3- •BLK3z 2-Cl)Cl)1-0-Si - 5B LTH-HS (EXP. 5)DBLK1. 4-t3BLK2LU••J 3•BLK3C_I, -ziTr._TH0HL1 1 IS6-i0 Sii-i5 S16-20Si - 5 S6 - 10 Sii - 15 Si6 - 20159Figure 28. The distribution of missing values during training in the LTH (Experiment 2)and LTH-HS (ExperimentS) groups on Day 1 (first 20 stimuli of training) and on Day 2(20 stimuli retention test). The number of missing values in five stimuli bins (s 1-5, s6- 10,si 1-15, and s16-20) was considered. A. LTH group (distributed training at a 60-s 1ST). B.LTH-HS group (distributed training at a 60-s ISI with heat shock (45 mm, 32°C) betweenblocks of training). Error bars indicate +1- SEMs.U)CoLU-J‘IzCl)Cl)U)Cl)LUD-J>1%zCl)U)A160DISTRIBUTION OF MISSINGVALUES ON DAY 1 AND DAY 2LTH (EXP. 2)EJ DAY1• DAY2-ISi - 55-4.3.21•0-5-4-3-2-1•0•S6-iO Sli-15 S16-20B LTH-HS (EXP. 5)EJ DAY1•1Si - 5 S6-i0 Sli -15 S16 - 20161(Training: E(3, 111) = 10.71, p < .01); again, however, there was no significantdifference between the groups in this distribution (Training x Group: E(3, 111) = 2.21,n&). Finally, there was no significant difference in the distribution of missing valueswithin blocks of training across blocks (Training x Block: E(6, 222) = .36, n). or inthe expression of this between the groups (Training x Block x Group: E(6, 222) = .71,The introduction of heat shock treatments (45 mm, 32°C) during training did notseem to affect the number or the distribution of missing values during training on Day1.The distribution of missing values on Day 1 and Day 2 was considered in theLTH-HS group of Experiment 5 and the LTH group of Experiment 2 (distributedtraining, 60-s ISI; see Fig. 28A & B). There was no significant overall difference inthe number of missing values between these groups (Group: (1, 37) = .52, n.s.).The number of missing values did not differ significantly between Day 1 and Day 2(Day: E(1, 37) 2.60, j, nor was it different between groups (Group: E(1, 37) =.12, jj. The distribution of missing values was uneven within blocks of training(Training: E(3, 111) 5.07, p < .01). However, this distribution of missing valuesdid not change significantly between Day 1 and Day 2 (Training x Day: E(3, 111) =.92, nj; nor did this pattern alter between the groups (Training x Group: f(3, 111) =1.59,Overall, these analyses suggest that heat shock (45 min, 32°C) did not affect thenumber of missing values. Therefore any difference observed in habituation on Day 1or long-term habituation on Day 2 between these groups cannot be attributed to achange in the number or distribution of missing values seen here.162Experiment 6: the effects of heat shock (45 mm. 32°C) before training and heat shock(45 mm. 32°C) during training on the distribution of missing values between Day 1 andDay 2.In Experiment 6, the number of missing values seen in the LTH, LTH-HS andPRE HS / LTH-HS groups may be compared across days of training. The overallnumber of missing values expressed by the groups did not differ significantly (see Fig.29A, B, & C; Group: E(2, 56) = .64, n&). There was no significant differencebetween days of training in the number of missing values (Day: E(1, 56) = 3.73,nor was this distribution altered by heat shock treatments (Day x Group: F(2, 56) =1.68, The distribution of missing values was uneven across training withinblocks (Training: (3, 168) = 16.47, p < .0 1); however, this distribution was notdifferent across groups (Training x Group: E(6, 168) = 1.64, n), or between thedays of training (Training x Day: E(3, 168) = .84, n). Finally, the distribution ofmissing values within blocks of training across days of training was not significantlyaltered by the heat shock treatments (Training x Day x Group: E(6, 168) = 1.35,Overall, it appears that these heat shock treatments do not alter the distribution ofmissing values seen in these data. Therefore, any differences between the groups ordays of training in the reversal magnitudes observed cannot be attributed to the numberor distribution of missing values.Experiment 7: the effects of heat shock (45 mm. 32°C) on Day 2 on the number anddistribution of missing valuesThe possibility that the heat shock on Day 2 affected the distribution or numberof missing values on Day 2 was examined with a mixed design ANOVA (Group x Dayx Training). There was no significant difference overall in the number of missingvalues between groups (see Fig. 30A & B; LTH vs. LTH / D2 HS; Group: E(1, 37) =163Figure 29. The distribution of missing values during training in the LTH, LTH-HS and thePRE HS I LTH-HS groups on Day 1 (first 20 stimuli of training) and on Day 2 (20 stimuliretention test). The number of missing values in five stimuli bins (sl-5, s6-10, sl 1-15,and s16-20) was considered. A. LTH group (distributed training at a 60-s 1ST). B. LTHHS group (distributed training at a 60-s 1ST with heat shock (45 mm, 32°C) between blocksof training). C. PRE HS I LTH-HS group (distributed training at a 60-s 1ST preceded byheat shock (45 mm, 32°C) and with heat shock (45 mm, 32°C) during training. Error barsindicate +1- SEMs.U)z—BU)zwC5-‘4) 4-z C1) 3-—2-1—0-SI -5PRE HS I LTH-HSD DAY1DAY2AMISSING VALUES INEXPERIMENT 6LTH164E1 DAY1• DAY25.-4-3-2-I—0-5-4-3-2-1—0-S1-5 S6-lO S11-15 S16-20LTH-HSD DAY1•SI-5 S6-10 SI1-15 S16-20S6-10 S1I-15 S16-20165Figure 30. The distribution of missing values during training in the LTH and LTH I D2HSgroups on Day 1 (first 20 stimuli of training) and on Day 2 (20 stimuli retention test). Thenumber of missing values in five stimuli bins (s 1-5, s6- 10, si 1-15, and s 16-20) wasconsidered. A. LTH group (distributed training at a 60-s 1ST). B. LTH I D2HS group(distributed training at a 60-s 1ST with heat shock (45 mm, 32°C) on Day 2 ending 2 hbefore the retention test). Error bars indicate +1- SEMs.166MISSING VALUES INEXPERIMENT 7A LTHU)C,)LUD-I>C!,zCl)Cl)BU)Cl)LiiD-J>C!,zCl)Cl)*DAY1DAY2S1-5 S6-1O S1i-15 S16-20LTH I D2HSE:J DAY1DAY2Si -5 S6-1O Si1-15 Si6-20167.64, n). The number of missing values did not differ overall with the day of training(Day: E(1, 37) = 1.88, n), nor was there a significant effect of the Day 2 heat shockon this distribution of missing values across days of training (Group x Day: E(1, 37) =1.88,The distribution of missing values within blocks of training varied significantly(Training: E(3, 111) = 3.60, p = .02). However, this distribution was not affected bythe heat shock on Day 2 (Training x Day: f(3, 111) = .57, n). In addition thedistribution of missing values within blocks of training did not differ significantlybetween days of training (Training x Day: F(3, 111) = .89, nor was this affectedby the heat shock on Day 2 (Training x Group x Day: f(3, 111) = .32, n.&).These results indicate that the number and distribution of missing values wasnot affected by heat shock on Day 2. This implies that any differences in performanceobserved in this experiment cannot be attributed to the number or distribution ofmissing values.Experiment 8: the effects of brief heat shock on the number and distribution of missingvaluesThe effects of brief heat shock (15 mm, 32°C) during training may be differentthan that of the longer heat shock (45 min, 32°C) given during training in Experiments5 and 6. To examine this, a mixed design ANOVA (Group x Day x Training) wasperformed. It was found that the overall number of missing values differedsignificantly between the groups (see Figures 31 and 32; Group: E(3, 76) = 3.00, p =.04). In addition, the overall number of missing values differed significantly betweenthe days of training (Days: E(1, 76) = 9.95, p < .01), and the difference between Day 1and Day 2 was affected by the group (Group x Day: E(3, 76) = .3.34, p =.02).168Figure 31. The distribution of missing values during training in the LTH and LTH-earlyHS groups on Day 1 (first 20 stimuli of training) and on Day 2 (20 stimuli retention test).The number of missing values in five stimuli bins (sl-5, s6-10, sll-15, and s16-20) wasconsidered. A. LTH group (distributed training at a 60-s ISI). B. LTH-early HS group(distributed training at a 60-s 1ST with heat shock (15 mm, 32°C) during training from 0 to15 min in the 1-h interval following each training block). Error bars indicate ÷1- SEMs.U,CoUiD-J>zCoCoLTH-early HSMISSING VALUES INEXPERIMENT 8169A LTHU,b.-. 4Cl)UiD.-J<>( 2-zCl)Cl)1-*. DAY1DAY2Si-S S6-10 S11-15 S16-20B0-5-4-3-2-1•0•E: DAY1DAY2Si-5 S6-i0 S11-15 Si 6-20170Figure 32. The distribution of missing values during training in the LTH-mid HS andLTH-late HS groups on Day 1 (first 20 stimuli of training) and on Day 2 (20 stimuliretention test). The number of missing values in five stimuli bins (sl-5, s6-10, sl 1-15,and s16-20) was considered. A. LTH-mid HS group (distributed training at a 60-s 1ST withheat shock (15 mm, 32°C) during training from 15 to 30 mm in the 1-h interval followingeach training block). B. LTH-late HS group (distributed training at a 60-s 1ST with heatshock (15 min, 32°C) during training from 30 to 45 mm in the 1-h interval following eachtraining block). Error bars indicate +1- SEMs.It)U)Lii-J>C.,zCl)Cl)L()Cl)UID1>C.,zU)U)A.171MISSING VALUES INEXPERIMENT 8LTH-mid HSDAY1DAY24.3.21•0•Si -5LS6-iO S11-15 Si6-20B5.. LTH-Iate HS4-3-2-1-0-J DAY1DAY2Si-5 S6-iO S11-15 S16-20172The distribution of missing values was uneven during habituation (Training:E(3, 228) = 24.86, p < .01); however, this distribution was not affected by the groupor by the day of training (Day x Training: E(3, 228) = 1.43, n; Group x Training:E(9, 228) = 1.77, Day x Training x Group: E(9, 228) = 1.40, n.&).An inspection of Figures 31 and 32 suggests that while there was variation inthe number of missing values in the data between groups, that, overall, this differencebetween groups does not affect the distribution of missing values within trainingblocks. The significant interaction between group and day of training in the number ofmissing values found does not seem to follow any distinct pattern (for example, it doesnot appear to be different between the groups that received heat shock and the groupthat did not).Analysis of the number and distribution of missing values in the LTH and LTH-HSgroupsAs was done for the habituation data (Chapter Nine, synthesis of results), it ispossible to examine the LTH groups together and the LTH-HS groups together, in thiscase, to look for effects of experiment on the number and distribution of missingvalues. A mixed design ANOVA (Experiment x Day x Training) was performed on thefive LTH groups (Exp. 2, 4, 6, 7, and 8) and the two LTH-HS groups (Exp. 5 and 6).In the analysis of the number and distribution of missing values in the data ofthe LTH groups, it was found that there was a significant difference in the number ofmissing values between experiments (see Figure 33, Experiment: E(4, 94) = 3.03, p =.02). The difference in the number of missing values between Day 1 and Day 2 wasvery nearly significant (E(l, 94) = 3.76, The observed difference in the numberof missing values between experiments did not interact with the distribution of missingvalues between days of training (Day x Experiment: F(4, 94) = 1.01, The173Figure 33. The distribution of missing values during training in all the LTH groups (Exp.2, 4, 6, 7 and 8) on Day 1 (first 20 stimuli of training) and on Day 2 (20 stimuli retentiontest). The number of missing values in five stimuli bins (sl-5, s6-10, si 1-15, and s16-20)was considered. A. Day 1. B. Day 2. Error bars indicate +1- SEMs.174MISSING VALUES INALL LTH GROUPSA DAY150 EXP.2 I EXP.4U) 0 EXP. 6 ISI EXP. 73 EXP.6>0zS1-5 S6-1O S11-15 S16-20B DAY20 XP. 2 L EXP. 40 EXP. 6 D EXP. 7LC 4EXP.8Cl)LU3-JS1-5 S6-1O S11-15 S16-20175distribution of missing values across habituation training was uneven (Mauchly’s testof sphericity, chi-square approximation = 16.46, < .01, Hunyh-Feldt epsilon = .95;Training: E(2.85, 267.9) = 12.86, < .01) and this distribution differed betweenexperiments (Training x Experiment: E(11.4, 267.9) = 2.32, < .01).However, the distribution of missing values during habituation training was notdifferent between days of training (Day x Training: F(3, 282) = 1.19, n.s.), whichsuggests that any difference between Day 1 and Day 2 in the level of responding cannotbe attributed to a change in the distribution of missing values during habituation. Inaddition, the distribution of missing values during habituation on Day 1 and on Day 2 isnot affected by Experiment (Experiment x Training x Day: F(12, 282) .65, n.s.).Interestingly, these two experiments (Exp. 4 and 8) were run during the same4-month period, though by different experimenters; worms for both experiments weredrawn from the same worm colonies. Although the strain of worms used throughoutthese experiments remained constant and genetic variation should be slight, apparentdifferences in the propensity to accelerate to tap have been observed in the past in thislaboratory (Rankin, Beck and Chiba, unpublished observations). It is possible thatsome factor affecting behavior such as slight changes in the temperature of the trainingenvironment (20° - 24 °C) or colony population conditions alters the propensity toaccelerate to tap, resulting in a lower mean number of accelerations overall in theseexperiments.Overall, an increase in the number of accelerations observed from Day 1 to Day2 in the LTH groups is not surprising because it has been observed that greaterhabituation is accompanied by a greater proportion of accelerations (Wicks and Rankin,unpublished observations). The interaction between Day and Experiment may beproduced by the loss of this pattern in the experiments such as Exp. 4 and 8 when the176number of accelerations becomes too low to observe this relationship between thehabituation process and acceleration distribution.The number and distribution of missing values was examined in the LTH-HSgroups with a mixed design ANOVA (Day x Training x Experiment). It was found thatthe number of missing values did not vary significantly between experiments (seeFigure 34; Experiment: F(1, 36) = 1.03, n.s.); however, there was a significantincrease overall in the number of missing values between Day 1 and Day 2 (Day: F( 1,36) = 7.97, p < .01). This difference between Day 1 and Day 2 was not affected byexperimental factors (Experiment x Day: F( 1, 36) = .08, n.s.). The significant effect ofDay is interesting as it may have been expected that the number of accelerations wouldremain constant between days of training in the LTH-HS, which showed no long-termhabituation between Day 1 and Day 2. It is clear from these results that the differencebetween the LTH and LTH-HS groups in the expression of long-term habituation isprobably not produced by a difference between the conditions in the distribution ofmissing values between Day 1 and Day 2.There was a significantly uneven distribution of missing values acrosshabituation training (Training: E(3, 108) 7.73, p < .01). This pattern did not differsignificantly across experiments (Training x Experiment: E(3, 108) = .15,) oracross days of training (Day x Training: E(3, 108) = .32, j; Day x Experiment xTraining: E(3, 108) = 2.43, n).Overall, the analysis of missing values across the LTH and LTH-HS groupsindicates that there are variations between experiments in the number and distribution ofmissing values observed that are not related to the training condition. These differencesin the number and distribution of missing values may bear some relationship to thequalities of the expression of habituation and the unknown factors which affect177Figure 34. The distribution of missing values during training in both of the LTH-HSgroups (Exp. 5 and 6) on Day 1 (first 20 stimuli of training) and on Day 2 (20 stimuliretention test). The number of missing values in five stimuli bins (s 1-5, s6- 10, sil- 15,and s16-20) was considered. A. Day 1. B. Day 2. Error bars indicate +1- SEMs.MISSING VALUES INALL LTH-HS GROUPSA DAY1178D EXP.5• EXP. 6TTrITILiLf)Cl)LU-J>zCl)U)BCl)UiD-J>zU)Cl)41:5.4.3.2-1-0-5.4.3.21—0-S 6-10 S11-15DAY2iISi -5I1TSi-5S 16-20D EXP. 5• EXP. 6TITS6-1 0TJS11-15 S 16-20179performance between experiments. However, as seen in the overall analysis of thehabituation data (Chapter Nine, synthesis of results), the expression of long-termhabituation and the effect of heat shock during training on LTH seems to be robust tothese variations.

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088240/manifest

Comment

Related Items