UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effects of changing the interstimulus interval during habituation in Caeorhaditis elegans Broster, Brett S. 1992

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

Item Metadata


831-ubc_1992_fall_broster_brett.pdf [ 2.59MB ]
JSON: 831-1.0098832.json
JSON-LD: 831-1.0098832-ld.json
RDF/XML (Pretty): 831-1.0098832-rdf.xml
RDF/JSON: 831-1.0098832-rdf.json
Turtle: 831-1.0098832-turtle.txt
N-Triples: 831-1.0098832-rdf-ntriples.txt
Original Record: 831-1.0098832-source.json
Full Text

Full Text

THE EFFECTS OF CHANGING THE INTERSTIMULUS INTERVALDURING HABITUATION IN CAENORHABDITIS ELEGANSbyBRETT SHELDON BROSTERB.Sc., The University of British Columbia, 1989A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF ARTSinTHE FACULTY OF GRADUATE STUDIES(Department of Psychology)We accept this thesis as conformingto the the required standardTHE UNIVERSITY OF BRITISH COLUMBIASeptember 1992© Brett S. Broster, 1992In 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.(SignDepartment of ^Ps CAA) (e), The University of British ColumbiaVancouver, CanadaDate (4(101;,e,„^) DE-6 (2/88)iiABSTRACTAlthough habituation is one of the simplest forms ofnon-associative learning, its underlying neural mechanisms arestill not well understood. One factor that plays a key role inhabituation is interstimulus interval (ISI). Understanding, at abehavioural level, the effects that ISI has on habituation mayprovide important insights into the cellular events involved inthis form of learning.The experiments in this thesis further explored the role ofISI in habituation of the reversal response of the nematodeCaenorhabditis elegans by examining the effect of changing theISI during habituation training. The effect of ISI change wasexamined in terms of both its impact on habituation and itsimpact on spontaneous recovery from habituation.One type of ISI change tested was continual variation in theISI used during habituation. When habituation stimuli weredelivered at variable ISIs having an overall average of 10 s therecovery from habituation observed was slower than that seen whenhabituation stimuli were given at regular 10 s intervals. Acomparison of fixed and variable stimulation during habituationwith a 60-s ISI revealed no differences in recovery rate. Thus,the impact of variable ISIs during habituation on recovery fromhabituation was noticeable at a 10-s ISI, but not a 60-s ISI.In a second experiment, the effect of shifting to a differentISI part-way through habituation training was explored. Whetherthe shift was from a 10-s ISI to a 60-s ISI or a 60- to a 10-siiiISI, in both cases the recovery rate (which is typicallydifferent for each ISI on its own) observed after habituation wasprimarily determined by the ISI given in the last half of thehabituation treatment.Examination of the impact on response patterns resulting fromvariation or change in ISI generated a model of how responsepotential may interact with ISI that can be used to furtherunderstand the relationship between ISI and response magnitudeduring habituation.ivTable of ContentsAbstract^.^•^•^•^•^•^•^•^. iiList of Figures^.^.^•^•^•^.^•^•^.^vAcknowledgements^.^•^•^•^•^•^•^. viIntroduction^.^•^• .^•^•^.^1General Methods^.^.^•^•^•^•^•^.^13Subjects^.^. •^•^•^•^•^.^.^13Apparatus^•^•^•^•^•^•^.^. 13Procedure^•^•^•^•^•^•^•^. 17Response Analysis^.^.^•^•^•^•^. 17Statistical Analysis^•^•^•^•^•^•^.^18Experiment 1. Variable vs. fixed interstimulus intervals^. 20Method^.^.^•^•^•^•^•^.^.^. 20Results and Discussion^•^•^•^•^.^.^. 21Experiment 2. The effect of mixed ISIs during habituation . 34Method^•^•^•^•^•^•^•^•^•^.^. 34Results and Discussion^.^.^.^•^•^.^. 35General Discussion^•^•^•^•^•^•^•^49References^.^•^•^•^•^•^•^•^64Appendix I^.^.^•^•^•^•^•^•^70Appendix II^.^.^•^•^•^•^•^.^.^71Appendix III^•^•^•^•^•^•^•^•^. 72vList of Figures1. The observation and stimulus generation equipment used inthe experiments^.^.^.^.^.^.^.^162. A comparison between the 10-s ISI fixed and variabletreatment groups during habituation and recovery .^.^233. A plot of the average response magnitude for differentinterval lengths during habituation for both the 10-s and60-s variable ISI groups^.^.^.^.^.^.^284. A comparison between the 60-s ISI fixed and variabletreatment groups during habituation and recovery .^. 315. A comparison of the 10-s and 60-s ISI control groups^.^376. The effect on response patterns of shifting from a 10-s toa 60-s ISI during habituation.^.^ .^.^417. The effect on response patterns of shifting from a 60-s toa 10-s ISI during habituation^.^.^.^.^. 438. A comparison of the mean habituated response levels foreach of the mixed ISI groups^.^.^.^.^489. An illustration of how the response-potential curve maybe affected by different treatments^.^.^.^5910. A representation of the periodicity existent in the 10-sfixed ISI habituation data^•^•^.^. 75AcknowledgementsWith all the people who advised and supported me throughoutthe completion of this thesis, these acknowledgements are notunlike an academy award acceptance speech. First of all, I wouldlike to thank all the scorers who helped score the data. I amdeeply indebted also to many in the lab who helped in a number ofdifferent ways (data analysis, interpretation and stress relief),especially Chris Beck Marian Buday, Bill Mah, and Steve Wicks.Special thanks go to Don Wilkie and Charlotte Johnston, whom Ireally appreciated having on my thesis committee. Most of all,however, I would like to thank my supervisor, Cathy Rankin, whohas contributed far more to my appreciation of science thananyone else during my university education.vi1IntroductionHabituation, defined as the decrease in an organism'sresponsiveness resulting from repeated stimulation, is probablythe most simple and ubiquitous form of learning known (Groves &Thompson, 1970). It has been studied in a wide variety ofcreatures and preparations, ranging from protozoa (Wood, 1970)and isolated spinal neurons (e.g., Farel, Glanzman, & Thompson,1973) to many different invertebrates and vertebrates, includinghumans (e.g., Sokolov, 1963). Many of these efforts, especiallythose conducted in the past 30 years, have aimed at elucidatingthe neural events that underlie this relatively simple form ofplasticity, but, despite great technological advances inmolecular biology and neurophysiology, habituation is stillpoorly understood.In all organisms habituation is characterized by the formthe response curve takes when response level is graphed overtime. At the beginning of habituation there is a sharp declinein responsiveness, and this is followed by a flattening of theresponse curve, known as the asymptote, beyond which littlefurther decrease in response level takes place. This decrementcan be distinguished from processes not considered learning,such as sensory or motor fatigue, by the ability of a novel ornoxious stimulus to restore responsiveness sooner than anorganism would recover from fatigue. This phenomenon is knownas dishabituation (Groves & Thompson, 1970).2Most of the progress made thus far has involved the use ofa simple systems approach, whereby researchers study learning ineither partially intact nervous systems or organisms withrelatively few neurons, as is the case with many invertebrates.For example, Thompson and Spencer (1966) used data from a spinalcat preparation in their classic characterization of the mainbehavioural features of habituation, while Kandel and colleagueshave attempted to elucidate the cellular mechanisms ofhabituation using the marine mollusc Aplysia californica.Extensive research on this organism by Kandel, Carew and manyothers (for a reviews see Hawkins, 1988, and Sahley & Carew,1983) has indicated that habituation probably involves adecrease in available neurotransmitter (Bailey & Chen, 1988;Castellucci & Kandel, 1974), coupled with a decrease in calciumcurrent (Klein, Shapiro, & Kandel, 1980) and possibly a declinein the efficiency of the neurotransmitter replenishment system(Bailey & Chen, 1988; Kandel, 1976). All of these processeswould contribute to a gradual decrease in the amount oftransmitter released, and, therefore, a drop in responsiveness.While work on Aplysia continues to be fruitful, otherresearchers have chosen to study invertebrates having nervoussystems that are even less complex and, more importantly, moreamenable to genetic and neuroanatomical analysis.One such organism is the soil dwelling nematodeCaenorhabditis elegans. This animal was originally isolated byBrenner (1974) who felt it held great promise for studying thegenetic basic of behaviour. Since that time, researchers haveprimarily concentrated on understanding its genome, which hasnow been almost completely mapped out (Coulson, Sulston,Brenner, & Karn, 1986; Hodgkin, Edgley, Riddle, & Albertson,1988). The genome is only about half the size of that ofDrosophila melangoster, containing only six small haploidchromosomes and 8 X 10 7 nucleotide base pairs (Sulston &Brenner, 1974). As well, a number of techniques have beendeveloped for the purposes of isolating, maintaining, andstudying genetic mutants (Wood, 1988). Most of these mutantsare readily obtainable from a library of mutants located at theUniversity of Missouri, Columbia (Hodgkin et al., 1988).C. elegans has only 302 neurons (Sulston, Schierenberg,White, & Thompson, 1983), all of which have had their putativesynapses (electrical or chemical) located and their cellularlineage defined (Chalfie 1984; White, Southgate, Thomson, &Brenner, 1986). As well, the general function of many of theseneurons in various behaviours has been ascertained, which hasgreatly aided efforts to examine the plasticity of thesebehaviours at the neural level.In addition to being a promising link between behaviour andgenes, C. elegans is also easy to maintain. These free-livingnematodes are generally kept on nematode growth medium agar(Brenner, 1974), though they can also be frozen (often using34liquid nitrogen), a stasis from which they can be revivedwithout ill-effect (Wood, 1988). They have a life cycle thatlasts, on average, 12-14 days, during which time they can besustained on a diet of Escherichia coli. Reproduction isprimarily accomplished through self-fertilization, as C. elegansis typically hermaphroditic. Male C. elegans exist, but theyarise only through chromosomal abnormality, and, therefore, arerare unless specifically bred for (Hodgkin, Horvitz, & Brenner,1979).Many of the behaviours studied in C. elegans involvelocomotion. The worm usually lies on its side and moves in anundulating fashion made possible by alternating contractions ofthe dorsal and ventral muscles. When on agar, C. elegans willmove forward most of the time, but it will occasionally movebackward, either spontaneously, or when stimulated by vibrationor a tactile stimulus delivered to the anterior region of itsbody.Recently, using this reversal response to vibration, Rankinand associates have explored the ability of the worm todemonstrate learning. Rankin, Beck, and Chiba (1990) found thata mechanical tap delivered to the side of an agar-filled Petridish caused a worm on the agar's surface to cease its forwardmovement and swim backward. Furthermore, they found that thistap-withdrawal response decreased in both frequency andmagnitude with repeated stimulation. The pattern of this5response decrement was very similar to classic descriptions ofhabituation (Groves & Thompson, 1970). It was also shown that,once habituated, this reversal response could be dishabituated,which indicated that this behaviour was indeed learning, and notjust fatigue. In addition to habituation, Rankin, Beck, andChiba were also able to demonstrate other forms ofnon-associative learning such as sensitization and long-termhabituation (Rankin, Beck, & Chiba, 1990; Rankin & Chiba, 1988).An important part of the study of plasticity in thetap-withdrawal response has been the elucidation of the neuralcircuit underlying this behaviour. Through the use of lasermicrosurgery, genetic mutation, and electron microscopy, Chalfieet al. (1985) described a subset of neurons known as the touchwithdrawal circuit. This circuit consists of 85 neurons (5sensory cells, 5 pairs of interneurons, and 69 motor neurons)and underlies head-touch induced backward movement andtail-touch induced forward movement. Rankin and Chalfie (1989;see also Rankin & Wicks, 1991) demonstrated that this touchwithdrawal circuit was also responsible for mediating the tapwithdrawal response. The next step in a circuit analysis is touse laser ablation and mutation to localize where and how thelearning is taking place.Before a neural understanding of plasticity in the tapwithdrawal response can truly be accomplished, however, it isessential to have a firm grasp of the characteristics of thelearning at a behavioural level. Knowledge of the descriptivecharacteristics of a form of learning is essential because anytheory involving neural mechanisms must account for the ways inwhich that learning changes under different behaviouralconditions and parameters. This forces theorists to expand orsharpen their theories to match behavioural observations.Thus, Rankin and colleagues have continued theirbehavioural assessment of learning in C. elegans, concentratingparticularly on habituation. Rankin and Broster (1990, 1992)have focussed their attention on the effect of interstimulusinterval (ISI) on habituation of the tap withdrawal response.The importance of the ISI factor in habituation has beenrecognized for many years and observed with a variety ofbehaviours. Yerkes (1906), who studied habituation of theshadow withdrawal response in the serpulid Hydroides dianthus,was one of the first invertebrate researchers to note thathabituation developed at a faster rate and to a greater extentwhen repeated stimulation was more frequent (shorter ISI).Similar results have been obtained with creatures as simple asthe protozoan Stentor coeruleus (Wood, 1970) and the leech(Ratner, 1972). Even complex behaviours, such as territorialityin fish (e.g., Peeke & Peeke, 1973) and the orienting responsein humans (e.g., Geer, 1966) appear to follow the the samepattern. In their study, Rankin and Broster (1992) found thatC. elegans conformed to this trend; shorter interstimulus67intervals resulted in more rapid and complete habituation thanlonger ISIs.One important issue that has been raised, however,concerning the evaluation of habituation is that the test usedto measure the effects of a particular parameter on habituationis often confounded with the training protocol used to producethat habituation (Davis, 1970). For instance, direct comparisonof response levels during the last stimulus of habituation forone animal habituated with a 5-s ISI and another habituated witha 60-s ISI is confounded by unequal periods of rest since theprevious stimulus. Thus, the relative performance of these twoanimals may be more dependent on the amount of time since thelast stimulus, as opposed to the actual amount of habituationthat has taken place. To remedy this, Davis (1970) suggestedthat a better approach would be to test all groups at equaltimes by using a small battery of post-habituation test stimuli.For example, in his experiment, Davis tested all groups with asimilar procedure 24 hours following habituation. Rankin andBroster (1992) adopted a similar protocol in that they testedall of their groups using spontaneous recovery from habituation.This was done by giving all animals probe stimuli at 30 s, 5min, 10 min, and 20 min following habituation at either a 2-s,10-s, 30-s, or 60-s ISI. The responses to these probe stimuliserved as a reflection of the rate of recovery from habituation.Rankin and Broster (1992) found that, although shorter ISIs8produced more rapid and extensive habituation, such treatmentalso resulted, surprisingly, in a recovery rate that was fasterand more complete than with longer ISIs.There are very few other investigations of habituation thatinvolve thorough examination of spontaneous recovery, and evenfewer that have looked at the relationship between recovery andISI. Most relevant research has instead looked at the relatedphenomenon of retention, which is often assessed by eitherrehabituation or brief tests given some time after the initialhabituation. The majority of this research agrees, inprinciple, with the findings of Rankin and Broster in that theeffects of habituation last longer when a long ISI is used. Forinstance, Davis (1970) found greater 24-hr retention ofhabituation (i.e., reduced response) in the acoustic startleresponse in the rat with longer ISIs than with shorter ISIsSimilarly, File (1973) reported better retention of habituationto lick suppression when rats had previously been habituatedwith longer ISIs.One advantage that a recovery protocol might hold over aretention protocol is that, because recovery reflects thereversal of habituation, it may provide more direct clues as tohow habituation develops in the first place. For example, theresults of Rankin and Broster (1992) suggest that habituation isunlikely to be produced purely by depletion of neurotransmitterat key areas in the response circuit. If this were the case,then treatment with shorter ISIs , which results in morecomplete habituation, should require longer recovery time,rather than the shorter recovery time that was observed. Forthese results to be explained, at least one other process inaddition to reduced transmitter release must be involved inhabituation and recovery. It may just be that a short ISIconstrains the amount of recovery that can take place betweenstimuli, resulting in lower response levels during habituation,or it may be that the interplay of cellular mechanisms duringhabituation is different for different ISIs.A model of habituation and recovery in the gill-withdrawalreflex of Aplysia has been proposed by Gingrich and Byrne (1985)that both accounts for the results of Rankin and Broster (1992)and suggests an explanation for the critical role played byinterstimulus interval. Their model, which focuses on theevents that might take place during habituation and recovery ina single neuron, suggested that during short ISIs the cell doesnot have enough time to reduce the high intracellular calciumconcentration resulting from influx during stimulation. As aresult, calcium accumulates in the cell. Some of this excesscalcium activates the neurotransmitter mobilization system.This enhances the ability of the neuron to re-stock transmitterat the terminal, which means that shorter ISIs can producefaster recovery.91 0The accuracy of this model remains to be demonstrated, butit does emphasize a trend that can be seen in many aspects ofhabituation and recovery and that is that there seem to be bothshort- and long-term processes at work. For instance, one thingthat has often been noted about recovery is that up to 85% ofthe initial response amplitude can return within a relativelyshort time while the rest of the recovery process can take muchlonger (e.g., Pakula & Sokolov, 1973). Such a pattern was alsoevident in the C. elegans data reported by Rankin and Broster(1992). While the group treated with the shortest ISI (2 s)reached 100% recovery in the first 10-min post-habituation, allother groups exhibited rapid recovery up to 60 or 70% ofbaseline in the same time, but then improved very little or notat all in the 20-min time period afterward. It is easy toimagine that the first phase of recovery might represent one setof mechanisms that recover rapidly and are associated with shortISIs, which would explain the advantage that shorter ISI groupsappear to have in the early part of recovery. Likewise, thelater stages of recovery may involve the resetting of processesthat are more strongly affected by habituation, particularlywhen longer ISIs are used.A number of researchers (e.g., Davis, 1970; File, 1973)have suggested that the rapid and extensive response decrementseen during habituation with short ISIs may be due more to somesort of refractory phenomenon, rather than a process that is1 1more clearly habituation (such as that seen during long ISIs).Rankin and Broster (1992) have expressed the idea that apossible refractory process might be sensory adaptation. Theysuggested that greater involvement of sensory adaptation duringshort ISIs would substantially reduce response levels whileprotecting the neuron terminal from extensive neurotransmitterdepletion. This savings would then be reflected in superiorrecovery once habituation stimulation had been terminated.Rankin and Broster further highlighted the impact of ISI on theprocesses that determine pattern of habituation and recovery bydemonstrating that the level of habituation at asymptote didlittle to dictate the rate of subsequent recovery. Moreimportantly, they also showed that increasing the number ofstimuli given during habituation once asymptote had been reachedhad very little effect on the recovery that followed. Forexample, they found the recovery rate after 60 stimuli at a 10-sISI to be very similar to the recovery rate after as few as 8stimuli at the same ISI. Both of these findings highlight theapparently critical role played by the ISI used during theinitial part of habituation. They also reinforce the idea thathabituation may consist of two phases. The first phase, whichappears to have the biggest impact on recovery, seems to occurduring the initial sharp decline in response level prior toasymptote. The second phase, which appears to have littleeffect on recovery, would begin once asymptote is reached.The following experiments were designed to further explorethe role played by ISI in habituation and recovery fromhabituation in an attempt to clarify some of the issues thathave been discussed. Given that the time between stimuli hassuch an important impact, I felt it would be interesting toexplore the effects of switching the ISI during habituation.In the first experiment, the importance of precise andregular ISIs was examined by comparing worms habituated withcomparable fixed and variable ISI schedules. Little researchhas been done on the subject of ISI variation, and even fewerstudies have dealt with its effects on recovery fromhabituation. Davis (1970) explored the effect of a variable ISIon retention of habituation in the acoustic startle response ofthe rat. He found that this treatment resulted in lessretention 5 min after habituation than when a fixed interval wasused. While less habituation with a variable ISI has often beenreported (Laming & McKinney, 1990; Mackworth, 1968; Ruchkin,1965) there are few studies that have looked at its effects froma post-habituation frame of reference.The other broad issue investigated in these experiments wasthe apparent importance of the first few stimuli of habituation,and how this position effect may interact with a shift in theISI during habituation. This was explored by examining theeffect on habituation and recovery of shifting from a long to ashort ISI (and vice versa) part way through the habituation1213procedure and observing which ISI, the initial or the final, hadthe most influence in determining the rate of recovery.General Methods$ubjects A total of 160 hermaphroditic adult C. elegans (Bristolstrain N2) were used in these experiments. Until testing, allworms were stored at 200C on 5-cm petri plates filled with 10 mlof NGM agar and streaked with Escherichia coli (strain op50;Brenner, 1974).In these experiments recovery from habituation was assessedby comparing response levels of post-habituation test stimuli tothe initial response in the habituation series. Because of thiscomparison to initial response levels a response criterion wasused to screen subjects. To qualify as a subject each worm hadto have an initial reversal of half a body length or more, andat least one of the two subsequent responses had to be areversal. In these experiments about 90% of worms tested metthis criterion.Apparatus Individual worms were tested and observed on unstreakedpetri plates using a stereomicroscope (Wild Leitz, Canada, Ltd.,model M3Z) and attached videorecording equipment (Panasoniccamera D5000, Panasonic AG-1960 VCR, JVC colour monitor).Stimulus delivery timing was aided by a time-date generator14(Panasonic 814) which superimposed a stop-watch onto the videoimage.For testing, plates with single worms were placed in aholder made from a petri plate lid which was glued to a plasticrod. The other end of the rod was held by a Marzhausermicromanipulator (MM33) so that the plate could be movedsmoothly when keeping the worm within the camera field. Alsomounted on the rod is the mechanical tapper used to stimulatethe worm (refer to Figure 1). The tapper consisted of anL-shaped copper wire arm (1.7 mm thick) attached at one end tothe armature of an electromagnetic relay (6 V). The main arm ofthe tapper was 14 cm in length from the point where it wasattached to the relay to where it bent at 90 0 to form thesmaller arm. This smaller arm, which was 3.5 cm long and rubbertipped, was positioned perpendicular to, and halfway up, thewall of the petri dish. When the relay was activated, the tipof the tapper, which was touching the dish at rest, wouldoscillate at a peak amplitude of approximately 2.5 mmperpendicular to the tangent of the point where it contacted thedish. This contact created vibrations which were transmittedthrough the dish and the agar.The stimulus used in these experiments was a brief(600-ms) train of six taps that delivered a peak force of 1.1 N(refer to Appendix I) to the side of the petri dish. To producethis stimulus the relay was electrically connected to a GrassFig. 1. The apparatus used to test and observe worms. Depictedin the top of the figure are the stereomicroscope, videoequipment, and stimulator equipment used. In the lower part ofthe figure are the mechanical tapper and holder used forstimulating the worm. (From Mah, 1991).15STIMULUSGENERATOR JVIDEOCAMERA CAENORMABDITISELEGANS00000 o oo oVIDEOCASSETTERECORDERTIME-DATEGENERATORIVIDEOMONITORMECHANICALTAPPER/////PETRIPLATEPETRI PLATEHOLDER\^\ \.^1\\ \it MICRO-^\^\ MANIPULATOR/^& STAND^\ \MICROSCOPE1617S88 stimulator which was set to deliver a signal of six 25-mspulses at 60 V and a rate of 8.5 pulses per second.Procedure In all experiments individual worms were transferred to afresh agar-filled petri plate from a colony plate about 2 minprior to testing. In each of the following experiments apre-determined number of trains was given according to aspecified interstimulus interval schedule. To monitorspontaneous recovery from habituation, single stimuli were givenat 30 s, 5 min, 10 min, and 20 min following the lasthabituation stimulus. Specific details for each experiment aredescribed in the corresponding procedure section.Response Analysis The length of reversals (distance travelled while swimmingbackwards) given in response to trains of taps was the responsemeasure that was used. Responses were scored by reviewingvideotapes using stop-frame video analysis and tracing the path(the distance travelled) of each reversal onto an acetate sheet.The tracings were then digitized using a digitizing tablet(Summagraphics Bit Pad Plus) interfaced with a Macintosh SEmicrocomputer and Macmeasure software.Reversals were considered to be caused by the stimulus onlyif they occurred within 1 s after the last tap in the train wasdelivered. If the worm appeared to be unaffected by a stimulusthat response was given a score of zero. If the worm was in the18process of reversing when the stimulus occurred, or if itaccelerated in response to it, the response was assigned ablank. About one in every five responses was scored as a blank,either for the reasons just mentioned, or because technicaldifficulties prevented the scoring of a response.Statistical Analysis Many of the statistical methods that were used on thesedata have already been established through other researchexamining changes in habituation in C. elegans (e.g., Beck &Rankin, in press; Mah & Rankin, in press; Rankin & Broster,1992). In general, data involving response magnitude wasanalyzed using t-tests or ANOVAs with Fisher's protected leastsignificant difference planned comparisons (PLSD) whenstatistical significance was achieved. All between-groupcomparisons were made using data that were standardized bydividing all responses of a given animal by its initialresponse. All within-group analyses were conducted with thedata untransformed. Any time multiple t-tests were employed thetype I error rate was adjusted for each test to keep thefamily-wise error rate below .05.The specific characteristics of habituation and spontaneousrecovery that were examined included the following: presence ofhabituation, rate of habituation, level of habituation, andextent of recovery. The presence of habituation was tested by19comparing the first response to the average of the last fourresponses of habituation within each group. Differences betweengroups in rate of habituation were examined by comparing meanslopes of the first few (12 or 25, depending on the ISI used andthe protocol) responses in the habituation series. Differencesbetween groups in response level prior to recovery (i.e., atasymptote) were assessed by comparing the average of the lastfour habituation responses.With regard to recovery, in almost all situations therewere no significant differences between the 5-min, 10-min and20-min responses for a particular group, therefore, these pointswere usually pooled. The resulting mean was used in anycomparisons involving recovery. Within each group, an overallrepeated measures ANOVA involved the initial response, the meanhabituated response, and the mean recovery response. The extentof recovery was assessed by comparing overall recovery to boththe asymptote level (to determine whether significant recoverytook place) and the initial response level (to determine whetherresponsiveness returned to baseline levels).In Experiment 2 the mixed ISI habituation curves consistedof two components, one for each ISI. To assess the effects ofISI transition, separate mean habituated responses were computedfrom the last four responses at each ISI, and these averageswere included in the overall ANOVA. Also included were theinitial responses for each ISI.Experiment 1: Variable vs. fixed interstimulus intervalsThe aim of this experiment was to investigate the effectof continuous ISI variation on habituation and recovery in C.elegans. This was done by comparing two groups habituated withthe same average ISI, one given a wide range of intervals in anirregular order, and another given the same interval on aregular basis. Sokolov (1963), in his stimulus-model comparatortheory of habituation, suggested that regularity of stimulustiming is an important part of learning not to respond to thatstimulus. It was expected that varying the ISI might,therefore, slow the rate of habituation. Because Davis (1970)found that animals treated with variable ISIs had less retentionof habituation, it was also expected that such treatment wouldinfluence the extent of recovery observed. Because the outcomemight depend on whether the average ISI was long or short, thisexperiment was carried out at both a 10- and 60-s ISI.MethodA total of 80 worms were used. During habituation onegroup of 20 worms received variable stimulation at, on average,a 10-s ISI (scheduled intervals ranging from 2 s to 40 s; seeAppendix II) and another 20 animals received stimuli atprecisely a 10-s ISI. An additional group of 20 worms were20given stimulation at regular intervals of 60 s, and acorresponding group of 20 animals were stimulated at scheduledvariable intervals having an average of 60 s (minimum of 5 s toa maximum of 4 min; see Appendix II). All worms received atotal of 60 stimuli during habituation, and were given fouradditional stimuli to test for recovery. These stimuli wereadministered at 30 s, 5 min, 10 min, and 20 min after the lasthabituation stimulus.Results and Discussion 10-s ISI fixed vs. variable interval. For thefixed-interval group, an overall repeated-measures ANOVA withFisher's PLSD comparisons indicated that there was a significantdecrease in response level from the beginning to the end ofhabituation training, F(2, 38) = 88.039, p = .0001 (refer toFigure 2). This analysis also showed that there was significantrecovery from the habituated level (the mean of the last fourhabituation responses), however, there wasn't enough recovery toreturn response levels to baseline. For this test, recoverylevel was represented by taking the mean of the 5-, 10-, and20-min post-habituation responses for each animal. An ANOVAfound no significant differences between these points even wheneach of the blank cells had been replaced by the group mean.The variable-interval group showed similar habituation buta slightly different recovery pattern. There were significant2122Fig. 2. A comparison between the 10-s ISI fixed and variabletreatment groups during habituation and recovery. Responsemagnitude (+/- SEM) is expressed in terms of a percentage of eachworm's initial response (n = 20 for each group). A) Habituation:60 stimuli at, on average, a 10-s ISI. B) Recovery: Recoverystimuli were given at 30 s, 5 min, 10 min, and 20 min after thelast habituation stimulus.INIT is the response to the firsthabituation stimulus. HAB is the average of the last fourhabituation stimuli.10 20 30 40 50 60-a- 10S V.I.-o- 10S F.I.STIMULUS23B • 10S V.I.• 10S F.I.INIT^HAB^30S^5M^10M^20MSTIMULUSdifferences between the initial response, the mean habituatedresponse, and the mean recovery response, E(2, 38) = 72.07, g =.0001. A difficulty with this analysis was that the 5-min,10-min, and 20-min recovery points were found to besignificantly different, F(2, 38) = 6.59, g = .0035, and,therefore, it was not appropriate to pool them into a recoverymean. When the analysis was recalculated with the recoveryseparated into a 20-min point and a pooled 5- and 10-min point,the pattern of significant differences was very similar; the20-min point was still significantly below baseline, E(3, 39) =35.55, g = .0001.When the variable- and fixed-interval groups were directlycompared the data used were standardized to initial response(see Statistical Analysis). Unpaired t-test results showedthere to be no significant differences between these two groupsat habituation asymptote. There was, however, a significantdifference between the fixed-interval mean overall recoverylevel and the variable-interval 5- & 10-min mean recovery level,t(37) = 3.20, g = .0028. After 20 min of recovery time hadpassed this difference disappeared. Thus, although neithergroup recovered back to baseline, the fixed-interval groupshowed more complete recovery earlier (at 5 and 10 minpost-habituation) than those animals habituated withvariable-intervals.2425Slope analysis revealed that these groups also differed intheir rate of habituation prior to reaching asymptote. A slopefor each animal was derived from the regression line that bestfit the first 12 responses. It was decided that, for the 10-sISI group, the first 12 stimuli were best for assessing the rateof habituation because by the 12th stimulus both groups hadspent exactly the same amount of time being tested, and bothgroups had reached asymptotic response levels. A t-testrevealed that the fixed-interval group had a steeper slope,indicating a faster rate of habituation than thevariable-interval group, t(38) = 2.54, p = .0153.Two other characteristics of these groups were explored.First, it was suspected that the asymptotic portion of thehabituation curve of the fixed-interval group might contain somesort of periodic variation of response level that could not beexplained by any variation in the time between each stimulus.This periodicity was a property investigated purely for the sakeof interest, and therefore is described in Appendix III.A second issue explored was the extent to which each meanresponse during habituation of the variable-interval group couldbe correlated with the amount of time elapsed since the previousstimulus. It was hoped that this might permit a betterunderstanding of how time interval can affect the responseoutcome of any given stimulus. A correlation taken over theentire habituation process resulted in r(59) = .401, p < .005.26Interestingly, there was a tendency for this interval-responsecorrelation to vary, depending on whether the data involved wastaken from the period prior to or during asymptote. Whenseparate correlations were done for before (approximately thefirst 15 intervals) and during (the remaining 44 intervals)asymptote, there was a noticeably, but not significantly, largercorrelation during asymptote, r(15) = .376, 2 < .10, and z(44) =.582, 2 < .0005. To determine how responsiveness was related totime since the last stimulus, the average response of eachanimal for each interval type was calculated and plotted (seeFigure 3). This idea was prompted by the work of Davis (1970),who found response amplitude to be proportional to intervallength. The intervals used for this calculation were taken onlyfrom the asymptotic portion of the curve (the last 44intervals). Intervals during pre-asymptote were excludedbecause there was unequal representation of them during thisphase, which, considering the large magnitude of many of thoseresponses, might have biased the results. These data should,therefore, be thought of as a reflection of theinterval-response relationship that exists only once a fairlystable habituation level, using a variable ISI protocol, hasbeen established. For a comparison, these data are plotted nextto similar asymptotic response data from the 60-s ISI variablegroup.Fig. 3. A plot of the average response magnitude for differentinterval lengths during habituation for both the 10-s and 60-svariable ISI groups (n = 20 for each group). Responses used didnot include those occurring prior to asymptote (i.e. the first 15stimuli for the 10-s ISI group and the first 23 stimuli for the60-s ISI group).27100 -90 -80- ■ 10S ISI70 -^W 60S ISI- a^III 1 1 I60 -50-40 -30 -10^ 1120 -0282S 5S 10S 15S 20S 25S 30S 40S 60S 120S 180S 5MINSTIMULUS2960s-ISI fixed vs. variable interval. Analysis of thefixed-interval 60-s ISI group indicated that there was asignificant decrease in response level from the initial responseto the mean habituated response (the average of the last fourhabituation responses), F(2, 38) = 26.605, R = .0001 (refer toFigure 4). The mean of the 5-, 10-, and 20-min recoveryresponses, which a separate ANOVA showed were not significantlydifferent, and, therefore, could be pooled, was found to besignificantly greater than the habituated level. Recovery didnot, however, return to baseline.Analysis of the variable-interval group showed a verysimilar pattern of habituation and recovery, F(2, 38) = 47.277,= .0001. There was significant habituation and significantrecovery from habituation, but not enough recovery to reach theinitial response level. As with most of the other groups inthis experiment, most of the recovery took place in the first 5min post-habituation, with no significant gain in recoveryoccurring at 10- or 20-min points.T-tests, with the alpha-level adjusted downward to keep theoverall error rate below .05, were used to directly comparethese two groups. There were no significant differences betweenmean habituated response levels or between mean recovery levels.To analyze relative rates of habituation, slopes were calculatedover the first 25 responses for each animal. The first 25stimuli were used because by that point both groups had beenFig. 4. A comparison between the 60-s ISI fixed and variabletreatment groups during habituation and recovery. Responsemagnitude (+/- SEM) is expressed in terms of a percentage of eachworm's initial response (n = 20 for each group). A) Habituation:60 stimuli at, on average, a 60-s ISI. B) Recovery: Recoverystimuli were given at 30 s, 5 min, 10 min, and 20 min after thelast habituation stimulus. INIT is the response to the firsthabituation stimulus. HAB is the average of the last fourhabituation stimuli.300 1 0 20 40 6050111 60S V.I.IN 60S F.I.INIT HAB 30S^5M 10M 20MA-o-- 60S V.I.-o- 60S F.I.130STIMULUSBSTIMULUS3132tested for the same amount of time, and both groups were withinasymptotic response levels. The fixed-interval group was foundto have a significantly steeper slope, reflecting a more rapidrate of habituation, than the variable-interval group, 1(38) =2.26, R = .0296.As with the 10-s ISI variable group, a correlationco-efficient was calculated on the 60-s ISI variable intervaldata to determine the extent to which the response magnitude foreach stimulus was dependent on the length of the interval whichpreceded it. This correlation was found to be significantlydifferent from zero, r(59) = .672, p < .0001.Figure 3 shows how response amplitude varies as a functionof the time since the last stimulus. For this graph, responseaverages for each interval were calculated without inclusion ofthe first 22 intervals, because the unequal representation ofthese intervals prior to asymptote biased the results. Figure 3also shows a comparison of the 60-s variable ISI data with the10-s variable ISI data. Included were the two 5 min recoveryresponses, to assess how the response level for each of thesegroups compared over very long intervals. A repeated-measurestwo-factor ANOVA conducted on the common intervals (5 s, 10 s,30 s, and 5 min) revealed that there were significantdifferences between the two variable ISI groups at the twoshortest intervals (5 and 10 s), F(2, 38) = 9.82, p = .0033.Although there is a strong correlation between ISI and response33magnitude, these data indicate that the ISI alone does notcontrol response amplitude. Here the same intervals areproducing different response amplitudes. The response magnitudeto short intervals appears to be influenced by the cumulativeeffect of previous stimulation. Something about the 10-svariable ISI treatment has reduced the mean response level, or,alternatively, something about the 60-s variable ISI treatmenthas facilitated the mean response level, for these intervalsduring asymptote. The longer intervals appear more similar toeach other.In general, for both a 10- and 60-s ISI, treatment with afixed interval during habituation appeared to result in asharper decline in response levels prior to asymptote than thatseen when a variable interval was used. There were, however,differences between the two ISIs in the effect that regularityof stimulus delivery had on extent of recovery. While nodifference in recovery levels was observed for the 60-s fixedand variable groups, the 10-s fixed interval group recovered toa greater extent during the early part of the recovery phase (at5 and 10 min post-habituation).An additional point to mention is that the results of thisexperiment are not incompatible with an associativeinterpretation of habituation (e.g., see Whitlow & Wagner,1984). In Instrumental Learning stimuli delivered at variableintervals produce slower acquistion rates and longer retentionthan stimuli delivered at fixed intervals. This is similar tothe effect of habituating stimuli delivered at variable andfixed ISIs, especially the 10 s ISI group.Experiment 2: The effect of mixed ISIs during habituationIn this experiment the ways that ISI exerts its powerfuleffects on habituation and recovery were further investigated byusing two different ISIs, each for an equal number of stimuli,during habituation. Of interest was the effect that suchtreatment would have on the pattern of habituation and theextent of subsequent recovery. The two ISIs used, 10 and 60 s,were chosen because previous research (Rankin & Broster, 1992)has shown them to be distinctly different, both in terms oftheir habituation pattern and their recovery rate. It was feltthat these differences would make it easier to determine whichISI was having a bigger impact on habituation and recoveryresponse levels. Based on the findings of Rankin and Broster(1992) it was hypothesized that the order of presentation (i.e.,which ISI was given during the first half of habituation andwhich was given in the second half) might be an importantfactor.Method A total of 80 worms were used for this experiment. One3435group of 20 worms was given 15 stimuli at a 10-s ISI followed by15 stimuli at a 60-s ISI. Another group of 20 worms was given15 stimuli at a 60-s ISI followed by 15 stimuli at a 10-s ISI.Both groups then received four recovery stimuli, each one givenat 30 s, 5 min, 10 min, and 20 min, respectively, after the 30thstimulus of habituation.For controls, one group of 20 worms was habituated with 30stimuli at a 10-s ISI, and another group of 20 worms was giventhe same number of stimuli at a 60-s ISI. Both groups thenreceived recovery test stimuli at 30 s, 5 min, 10 min, and 20min after habituation.Results and Discussion Single ISI control groups. In order to have a baselinecondition to compare with the experimental groups, the resultsof the two control groups were examined first. Within each ofthese groups there were no significant differences between the5-, 10-, and 20-min recovery points, thus, any further recoveryanalysis employed a mean of these three responses.A repeated-measures ANOVA of the 10-s ISI control groupshowed that animals were significantly less responsive by theend of habituation (as represented by the average of the lastfour responses), compared to the start, F(2, 38) = 32.11, p =.0001 (refer to Figure 5). Recovery was not only significantlyabove the mean habituated level, it was extensive enough to benot significantly different from the initial response level.Fig. 5. A comparison of the two 30 stimuli control groups (10-sand 60-s ISI; n = 20 per group). Response magnitude is expressedin terms of a percentage of each worm's initial response, andincludes +/- SEM. A) Habituation: 30 stimuli at either a 10-s ora 60-s ISI. B) Recovery: Recovery stimuli were given at 30 s, 5min, 10 min, and 20 min after the last habituation stimulus.INIT is the response to the first habituation stimulus. HAB isthe average of the last four habituation stimuli.36—0--- 10S ISI—*-- 60S ISI37A^120 -W(1)^100 -ZOa.ca^80 -WCC-JQ 60 -HZ_1.-^40 -ZWVQ 20 -Wa.00^ 10^20^30STIMULUS■ 10S ISIMI 60S ISIINIT^HAB^30S^5M^10M^20MSTIMULUS38The 60-s control group also showed a significant decreasein response level during the course of habituation, F(2, 38) =28.66, 2 = .0001, but, unlike with the 10-s group, there was nosignificant recovery. The mean recovery level was significantlybelow the initial response level and was also not significantlydifferent from the mean habituated level.These two 30-stimuli control groups were compared in termsof their rate of initial habituation prior to asymptote byexamining differences in the mean slopes of the regression linescalculated for the first 12 stimuli of each group. Using thismethod, the 60-s ISI group was found to habituate at asignificantly slower rate than the 10-s ISI group. Oncehabituated to asymptote, the 60-s group also showed less overalldecrement in responsiveness compared to the 10-s group, t(38) =4.19, 2 = .0002. Thus, these two groups were different in manyaspects of their response curves; the 60-s ISI group wascharacterized as having habituation that was slower and lessextensive and recovery that also was slower and less extensivecompared to the 10-s ISI group.Mixed ISI experimental groups. For both the 10-to-60-s ISIand the 60-to-10-s ISI groups initial response levels, the meanhabituatep response levels for each ISI given (the average ofthe last four responses given in each series), the responselevel after the ISI was shifted, and the mean recovery responselevel (which, in both cases was pooled after no significant39difference was found between the 5-, 10-, and 20-min points)were analyzed using an overall repeated-measures ANOVA andFisher PLSD comparisons.The 10-to-60-s group (refer to Figure 6) was characterizedby a sharp and significant decrease in response level that istypical of habituation with a 10-s ISI. When the ISI wasswitched to 60 s there was an immediate and significant increasein response level over the 10-s ISI habituated level, but notenough to return to baseline, F(4, 76) = 47.02, 2 = .0001. Overthe course of the 15 stimuli given at a 60-s ISI there was asignificant decrease in responsiveness, similar to that seenwith other 60-s ISI groups. The mean response level at the endof the 60-s ISI part of the habituation curve was significantlyhigher than the mean response level at the end of the 10-s ISItreatment. Recovery was similar to that seen with the 60-s ISIcontrol group in that there was no significant improvement inresponsiveness from the 60-s habituated level.With the 60-to-10-s ISI group (refer to Figure 7) there wasa slow, but significant, decrease in responsiveness over thecourse of the 15 stimuli given at a 60-s ISI, F(4, 76) = 44.47,2 = .0001. Additional multiple comparisons showed that as soonas the ISI was shifted to 10 s there was an immediate andsignificant drop in the response level that appeared to take thegroup down to the level of a typical 10-s ISI asymptote, as wasindicated by the finding that the mean of the first two 10-sFig. 6. The effect on reversal response magnitude of shiftingfrom a 10-s to a 60-s ISI during habituation (n = 20). Responsemagnitude (+/- SEM) is expressed in terms of a percentage of eachworm's initial response. A) Habituation: The first 15 stimuliwere given at a 10-s ISI; the latter 15 stimuli were given at a60-s ISI. B) Recovery: Four recovery stimuli were given at 30 s,5 min, 10 min, and 20 min following the last habituationstimulus.40A^120•^I^.^I^.10 15 20^25^30STIMULUS0 10I^i IB100 -80 -60 -40 -20 -4142Fig. 7. The effect on reversal response magnitude of shiftingfrom a 60-s to a 10-s ISI during habituation (n = 20). Responsemagnitude (+/- SEM) is expressed in terms of a percentage of eachworm's initial response. A) Habituation: The first 15 stimuliwere given at a 60-s ISI; the latter 15 stimuli were given at a10-s ISI. B) Recovery: Four recovery stimuli were given at 30 s,5 min, 10 min, and 20 min following the last habituationstimulus.43A^120 -100 -80 -60 -40-20 -00I^.^I^•^I^r^I^I5 10 15^20^25 30STIMULUSB1201008060402030 S^5 MIN^10 MIN^20 MINRECOVERY STIMULUS44responses was no different from the mean of the last four. Therecovery of this group more closely resembled that of the 10-sISI control group in that it was significantly above both the60-s and 10-s habituated levels Recovery did not return tobaseline, perhaps due to a lingering effect of the 15 60-s ISIstimuli. The overall recovery of this group was higher thanthat of the 10-to-60-s ISI group, t(38) = 2.13, 2 = .0397.For each of the ISI transition groups, recovery rate seemedto be primarily determined by the ISI most recent to the onsetof recovery, as opposed to the ISI to which the animals werefirst habituated. It is unclear whether the early ISI had anyeffect at all on recovery rate, though there is the suggestionthat initial habituation with a 60-s ISI may have slightlyreduced the amount of recovery that took place after subsequenthabituation with a 10-s ISI. With respect to any effecthabituation with the first ISI may have had on habituation withthe second ISI, initial treatment with a 60-s ISI seemed to haveinfluenced the habituation pattern observed with a 10-s ISI.There is no evidence that initial habituation with a 10-s ISIhad any effect on the habituation with a 60-s ISI.The ways that habituation with one ISI were affected byprior habituation with another ISI were further explored bytesting whether the rate and level of habituation were dependenton whether an ISI came first or second in the habituationseries. Within each ISI, t-tests were used to test for45differences between the mean response level at the end of 15stimuli for the group that received the ISI first and the groupthat received it second. For both the 10- and 60-s ISIS therewas no difference in the respective habituated (asymptotic)response levels. Thus, the asymptotic response level achievedat a 60-s ISI was not significantly affected by the 15 10-s ISIstimuli that preceded it, and, likewise, the asymptotic responselevel achieved at a 10-s ISI was not significantly affected byprior stimulation at a 60-s ISI.For each block of 15 stimuli, t-tests were used todetermine whether there was significant decrement in responsemagnitude from the 2nd to the 15th stimulus by comparing themean of the first two responses following the first interval atthat ISI (i.e., stimuli 2 & 3, and stimuli 16 & 17) with themean of the last four habituation stimuli at that ISI. Therewas an effect of prior stimulation on how quickly the asymptotewas reached, but it was only seen during the 10-s ISIhabituation when it followed habituation at a 60-s ISI. In the10-s portion of the 10-s to 60-s ISI habituation, asymptote wasnot reached by the second or third stimulus, t(16) = 5.479, 2 =.0001, and the same was true for the second (60-s ISI) half ofthe habituation, t(19) = 2.83, 2 = .0108. For the 60-s to 10-shabituation, the results were the same, asymptote was notreached by the second or third stimulus during the 60-s ISI halfof habituation, t(19) = 2.738, 2 = .0131; however, with the 10-s46ISI habituation that followed, asymptote was reached rapidly, asthere were no significant differences between the first two andthe last four 10-s ISI habituation responses.These findings support the idea that there was sometransfer of habituation from the 60-s ISI to the 10-s ISI, butnot vice-versa. It is also interesting to note that asymptoticlevel, once established, was unaffected by previous stimulation(refer to Figure 8), but, the rate of response decrement wasaffected when 10-s ISI habituation that was preceded by 15stimuli at a 60-s ISI. This suggests that the responsedecrement phase of habituation, prior to asymptote, is moresensitive to prior habituation than the asymptotic phase itself.ThiS' supports the hypothesis that one or more processes in thepre-asymptote phase are different from the processes of theasymptotic phase of habituation.Fig. 8. A comparison of the mean habituated response levels foreach of the mixed ISI groups. HAB 10S is the mean of the lastfour response at a 10-s ISI. HAB 60S is the mean of the lastfour response at a 60-s ISI. The first pair of bars are from the10-s to 60-s ISI group (n = 20) and the second pair of bars arefrom the 60-s to 10-s ISI group (n = 20). Response magnitude(+/- SEM) is expressed as a percentage of each worm's initialresponse.471 00 -9 0 -80-70-60-50  -40-30-20 -10-0 ^48HAB1 OS^HAB6OS^HAB6OS^HAB1 OS10S TO 60S GROUP^60S TO 10S GROUP MEAN HABITUATION LEVEL49General DiscussionThere is a wealth of research that has emphasized theimportance of the effect that interstimulus interval has onhabituation. It is highly conceivable that one of the reasonsthe neural underpinnings of this form of learning have thusfarbeen so elusive is that the underlying mechanisms may bedifferent for habituation to different ISIs. One way ofapproaching this problem at a behavioural level is to firstunderstand the overall effect a particular ISI protocol has, andthen look at how each individual stimulus event might havecontributed to this effect. Understanding the relationshipbetween these two aspects of habituation in terms ofbehavioural dynamics may, in turn, make it easier to elucidatethe cellular mechanisms underlying them.These two experiments have made it easier to apply thisapproach by employing protocols in which the ISI changes duringthe habituation procedure. By examining the effect that thesechanges had on recovery from habituation it was possible toassess them in terms of their overall effect. By examining theimmediate effect during habituation of these ISI variations itwas possible to come to a better understanding of how ISI mightbe interacting with events on a stimulus-by-stimulus basis.In the first experiment, the importance of the regularityof ISI timing was explored. It was found that, with a 10-s ISI,fixed time intervals during habituation produced faster recovery50than if the intervals were varied. No difference in recoverywas noted when the average interval was 60 s. That the outcomeis dependent on which ISI the experiment is conducted with mightindicate that animals are more affected by ISI variability whenshorter, rather than longer, intervals are used. The differenceseen in recovery with the 10-s ISI version of this experimentdiffers from the results of Davis (1970), who found that therewas more pronounced retention of habituation after treatmentwith fixed intervals than with variable intervals. Thiscontradiction is probably due to differences in thepost-habituation test protocol used. Davis used a large batteryof post-habituation tests, rather than just a few probe stimuli,and this may have made his protocol more like rehabituation.The results of Davis (1970) might, therefore, be replicated ifExperiment 1 were re-run using rehabituation instead ofrecovery.Another overall effect examined was the impact thatirregular ISIs had on rate of habituation. With both the 10-and 60-s comparisons, slope analysis revealed a faster rate ofhabituation when the stimulation was given at regular intervals.Laming and McKinney (1990) also used slope analysis to assessrate of habituation to light in the goldfish, and they too foundthat it was slower when there was variation in the ISI. Onehypothesis is that regularity of stimulation might make iteasier for the animal to learn more quickly that a stimulus is51not relevant. In general, though, the literature on thissubject is mixed, with results often changing with subtledifferences in protocol and method of analysis.An important issue concerning the interpretation of thesedata is that it is difficult to determine whether the effectsobserved are due to the presence of regularity (or the lack ofit), or an imbalance between the effects of the longer intervalsand the shorter intervals that are included in the treatment.For instance, the slower recovery seen with the 10-s ISIvariable group may be there because any long-term inhibition ofrecovery contributed by one type of interval (for example, thelonger ones) may outweigh any facilitation of recovery thatmight be provided by another type of interval (for example, theshorter ones). It is unlikely that the effects of overallregularity can be disentangled from the specific contributionmade by intervals of different length until the latter is betterunderstood.Rankin and Broster (1992) have speculated that very shortISIs might enhance recovery, or at least spare it from beingrestricted, by producing some sort of effect, such as sensoryadaptation, that reduces the amount of transmitter lost from thenerve terminal. From this, one might have predicted that theinclusion of very short intervals in the 10-s ISI variableschedule would have contributed to faster recovery for thevariable group. The results of Experiment 1 do not support this52prediction, but they also can not rule out the idea because anyeffect that short ISIs might have been outweighed by an oppositeeffect brought about by the inclusion of much longer ISIs in thevariable habituation schedule.In Experiment 2, the use of just two different ISIs duringhabituation made for easier interpretation. The ISI used mostrecent to the onset of recovery was the one that had the largestinfluence on rate of recovery. This is an interesting finding,especially when analyzed in light of the findings of Rankin andBroster (1992), who found that as few as eight stimuli at an ISIwere enough to produce a recovery curve that was typical forthat ISI, and that further stimulation did little to influencethe shape of that curve. Their results seemed to indicate thatrecovery rate had become, in a sense, fixed once habituation toasymptote had occurred. The present experiments havedemonstrated that recovery rate can be changed substantiallyeven once asymptote has been established. For instance, animalsfirst habituated to asymptote at a 10-s ISI and then treatedwith a 60-s ISI exhibited the limited recovery that is usuallyseen after full habituation with a 60-s ISI. This suggests thathabituation stimuli during the asymptote are capable of bothmaintaining recovery rate once it has been established (if thesame ISI is used throughout habituation) and changing it if ithas been previously set by some other ISI. Thus, it is likelythat each stimulus during habituation plays some role in53determining recovery rate. Depending on what has preceded thestimulus, this contribution can either help set, change, ormaintain the recovery rate.Another important aspect of Experiment 2 is the effect thatthe ISI change had on habituation itself. This should first beexamined in terms of the general effect that the first ISI hadon the second, i.e., whether habituation with the first ISIcarried over to the second. When the ISI was switched from 10 sto 60 s there was an elevation in response level that broughtthe group almost back to baseline, and the subsequenthabituation curve was much like a typical 60-s ISI habituationcurve. So, in this case, although the 10-s ISI habituationsubstantially reduced the response level prior to the change, ithad minimal impact on habituation at a 60-s ISI.In contrast, the effect of initial habituation with 15stimuli at a 60-s ISI seemed to carry over to the 10-s ISIhabituation. When the ISI was switched from 60 to 10 s, theresponse level went from a 60-s ISI asymptotic habituation levelto being not significantly different from a 10-s ISI habituationlevel. That a substantial transfer effect was observed with the60-to-10-s ISI protocol and not the reverse protocol suggeststhat the underlying mechanisms responsible for habituation atthe two ISIs may be linked in such a way that the eventsoccurring during habituation at a 60-s ISI can potentially mimicsome of the events that occur during habituation at a 10-s ISI.54This contrast between the 10-s-to-60-s change and the60-s-to-10-s change could be explained in terms of two types ofhabituation processes, one type that is transient and moreassociated with shorter ISIs and another type that is longerlasting and more associated with longer ISIs. According to thismodel (Staddon, personal communication) habituation at a shortISI may involve processes that result in a rapid and pronouncedresponse decrement during habituation as well as faster and morecomplete recovery following habituation. In addition,habituation at a long ISI may involve processes that cause aslow and shallow habituation that takes a long time to recoverfrom. During the change from a 10-s to a 60-s ISI the first 60s interval may have provided enough time for most of the 10-sISI habituation effects to recover. During the change from a60-s to a 10-s ISI there would not have been enough time torecover from the initial habituation at a 60-s ISI, and,therefore, the effects of this treatment would be carried overto the 10-s ISI habituation. If one considers the slow recoveryafter habituation stimuli at a 60-s ISI to be indicative of theamount of time that the effects of habituation at this ISI last,then it is interesting to note that there was little trace ofthese long-lasting effects in the recovery of the group thatreceived 2.5 min of 10-s ISI stimulation between 60-s ISIhabituation and recovery. This suggests that 10-s ISI stimulimay be actively reversing the effects of 60-s ISI stimulation.55Assuming that the long-lasting effects of 60-s ISIhabituation were somehow reduced by 10-s ISI habituation, theremay be one or more processes that can improve, or facilitate,(up to a point) rate of recovery. This idea has been expressedby others, such as Gingrich and Byrne (1985), who suggested thatan abnormally high increase in calcium levels at key areasduring high frequency (short ISI) stimulation may temporarilyImprove the ability of a neuron to replenish depletedtransmitter, hence, generating faster recovery.At this point it may be useful to approach any furtherspeculation about the results of these experiments by generatinga model that gives some account of what, in general, isoccurring between each individual interval during habituationstimulation. This model makes the assumption that, immediatelyafter a stimulus occurs, the ability of the circuit to respondto another stimulus is essentially zero. As time goes on, theresponse potential of the circuit, i.e., its ability to respondto another stimulus, rises. (The interval-response plots shownin Figure 3 would be examples of how response potential may varywith time since the last stimulus). The rate at which theresponse potential rises (or, recovers) may depend on what sortof stimulation treatment was received prior to the previousstimulus. The response amplitude of the next stimulus could bepredicted by observing how high the response-potential curve isfor the ISI that has just been experienced.56Habituation data suggests that the shape of the responsepotential curve changes with repeated stimulation. If it wasalways the same after each stimulus then, with regularstimulation at a single ISI, one would observe a drop inresponse level when the second stimulus was given (assuming thiswas done with a reasonably short ISI, that wasn't long enough toprevent interstimulus recovery back to baseline), but, therewould be no further drop in response level, as long assubsequent stimuli were given at the same ISI. Furtherstimulation would always interrupt the response-potential curveat the same point and, therefore, if the shape of the curvedidn't change, roughly the same level of response would resulteach time stimuli were given at that ISI. What actually happenswith the first few stimuli is very different from this, as mosthabituation graphs initially show a steady decrease in responselevel with repeated stimulation, at least until the asymptote isreached.This observed pattern may reflect a gradual slowing of theresponse-potential curve, with each stimulus during thepre-asymptote phase of habituation further decreasing the speedof inter-stimulus recovery of response-potential. Onceasymptote is reached, the observed response level flattens out,suggesting that the response-potential curve is no longer beingmodified within the time window of the ISI given; each newstimulus at that ISI produces roughly the same level of response57as the previous stimulus. An illustration of this is shown inFigure 9.This hypothetical response-potential model may be usefulfor investigating habituation in a number of ways. One of themost fundamental issues that can, and should, be examined usingit is whether or not the properties of the response-potentialcurve are affected in different ways by different ISIs, and, ifso, how. A comparison of the 10-s and 60-s ISI recovery curvesobserved in these experiments (e.g., the 30 stimuli controlgroups) indicates that, by 5 min after the last habituationstimulus, the 10-s ISI group has recovered very rapidly, both interms of absolute recovery and net recovery (recovery subtractasymptote), compared to the 60-s ISI group. This implies thatthere is a difference in the long-range portion of theresponse-potential curve for the two ISI groups after 30stimuli. By 5 min after the last stimulation, theresponse-potential curve for the 10-s ISI group would appear tobe at a higher point than that for the 60-s ISI group. Thus, itwould appear, in the context of this model, that habituation ata 60-s ISI probably produces some sort of long-lastinginhibition of response-potential, and/or, habituation at a 10-sISI probably causes some limited facilitation of recovery.An overview of the research that has been done onhabituation and recovery in C. elegans suggests that it may bebest to distinguish between those changes in theFig. 9. An illustration of hypothetical trends in theresponse-potential model. The response-potential curve may changewith repeated stimulation at a constant ISI. Each pointrepresents potential response amplitude for a stimulus deliveredat the post-stimulus interval shown on the x-axis. The curve ofrecovery of response potential for the first stimulus isrelatively fast, but it becomes slower with each successivestimulus.58Stim 1Stim 2Stim 3Stim 410080604020•50 100INTERVAL LENGTH (S)5960response-potential curve that are seen in the earlypost-stimulus period and those that are seen in the latepost-stimulus period. As discussed above, assessment ofrecovery responses indicates that there are different ISIeffects on the late post-stimulus part of the response-potentialcurve; habituation with a 60-s ISI seems to depress it, andhabituation with a 10-s ISI may facilitate it to some degree.Furthermore, the research of Rankin and Broster (1992) suggeststhat the effects of each.ISI on this portion of theresponse-potential curve are fully realized by the timeasymptotic response levels have been reached, as groupshabituated just to asymptote exhibit typical recovery curves forthe ISI they were treated with. This evidence reinforces theidea that the response-potential curve is stabilized byasymptote.With regard to any ISI differences in the earlypost-stimulus portion of the response potential curve, it ismore difficult to speculate about what may be happening. A hintof ISI differences here is given in the 10-s variable ISI vs.60-s variable ISI plot depicted in Figure 3. During the 60-svariable ISI treatment, the shorter ISIs (5 and 10 s) resultedin significantly higher responses than were seen at the sameintervals during 10-s variable ISI treatments. One possibleexplanation of this result is that the use of generally shorterISIs inhibits the early part of the response-potential curve.There are, of course, other explanations of these data,including the possibility that the 60-s variable ISI treatmentproduces some sort of facilitation of response potential duringthis period, or, that something about the irregularity of thesetwo groups is responsible for the differences between them.If different ISIs during habituation affect theresponse-potential curve in different ways, the next logicalquestion to ask is to what extent does each stimulus contributeto these differences. The results reported here suggest that itmay take relatively few stimuli at a particular ISI before thespecific cumulative effects on the response-potential curve aremaximized. As mentioned before, the results of Experiment 2suggest that stimuli can either set, change, or maintain theasymptotic habituation level, as well as the rate of recovery.Also, it appears that it takes relatively few stimuli (15 orless) for the second ISI to alter both the level of habituationand the pattern of recovery that had been, presumably, alreadyset by the first ISI. The dynamic nature of these dataunderscores the importance of the role played by individualstimuli. The cellular events that an individual stimulusprecipitates are likely the same every time, but their effect onresponse level is probably more a function of the interactionthat these events have with the cumulative effect of otherevents and factors, as reflected by the state of theresponse-potential curve at the time. For example, the6162inhibition of the early portion of the post-stimulusresponse-potential curve that might be occurring in short ISIhabituation may reflect the presence of a transient inhibitoryprocess that lasts for a few seconds after every stimulus(regardless of ISI), but which remains unseen unless severalshort ISIS in a row allow for an accumulation of the effects ofsuch a process.It would be interesting to follow up this research withexperiments that test how few stimuli are needed to change arecovery pattern that had already been set by one ISI. Forinstance, if five stimuli at one ISI were to follow 15 stimuliat another ISI, would that be enough to convert the recoverypattern observed? Perhaps the most important follow-upexperiments to be conducted, though, would be ones that moresystematically explored the relationship between length of theISI and the response magnitude observed. Such experiments mightstart by looking at the effect of the interjection of a singleinterval that is different from the ISI used during the rest ofhabituation. The effect of this interval may change dependingon when during habituation it is given, what the absolute lengthof the interval is, and how its length compares to the other ISIused. Experiments exploring the effect of one differentinterval could then be followed by ones that looked at theeffects of having two interjection stimuli in a row, and thenthree in a row, and so on. Experiments involving single63interjection stimuli offer two useful pieces of information. Onone hand, they may permit a full plotting of theresponse-potential curve for a given habituation protocol, aslong as a range of different single ISIs are used. Secondly,they may clarify the amount of impact that a single stimulusinterval can have, and how this impact varies with the treatmentthat follows or precedes the stimulus. Experiments involving afew interjection stimuli in a row may reveal the cumulativeeffects of this impact.The present experiments have produced a betterunderstanding of the role that ISI plays in habituation, and,more importantly, have introduced ways of viewing ISIs that mayprove very useful in the future. By understanding what may behappening between each stimulus, and how that relates to thecumulative effects of ISI it may be possible to more readilyconnect these observations at a behavioral level with themolecular events that might be responsible for them.ReferencesBeck, C. D. 0., & Rankin, C. H. (in press). Effects of aging onhabituation in the nematode Caenorhabditis elegans. TheNeurobiology of Aging.Brenner, S. (1974). The genetics of Caenorhabditis elegans.Genetics, 71, 71-94.Castellucci, V. F., & Kandel, E. R. (1974). A quantal analysisof the synaptic depression underlying habituation of thegill-withdrawal reflex in Aplysia. Proceedings of the National Academy of Sciences of the United States of America, 71, 5004-5008.Carew, T. J., & Sahley, C. (1986). Invertebrate learning andmemory: From behavior to molecules. Annual Review of Neuroscience, 9, 435-487.Chalfie, M. (1984). Neuronal development in Caenorhabditiselegans. Trends in Neuroscience, 7, 197-202.Chalfie, M., Sulston, J. E., White, J. G., Southgate, E.,Thomson, J. N., & Brenner, S. (1985). The neural circuitfor touch sensitivity in Caenorhabditis elegans. Journal of Neuroscience, 5, 956-964.Coulson, A., Sulston, J., Brenner, S., & Karn, J. (1986).Towards a physical map of the genome of the nematodeCaenorhabditis elegans. Proceedings of the National Academyof Sciences, 83, 7821-7825.64Davis, M. (1970). Effects of interstimulus interval length andvariability on startle response habituation in the rat.Journal of Comparative and Physiological Psycholoay, 72,177-192.Eisenstein, E. M., & Peretz, B. (1973). Aspects of habituationin invertebrates. In H. V. S. Peeke & M. J. Herz (Eds.),Habituation (Vol. 2) (pp. 1-34). New York: Academic Press.Farel, P. B., Glanzman, D. L., & Thompson, R. F. (1973).Habituation of a monosynaptic response in the vertebratecentral nervous system. Journal of Neurophysioloay, 36,1117-1130.File, S. E. (1973). Inter-stimulus interval and the rate ofbehavioural habituation. Quarterly Journal of Experimental Psycholoay, 25, 360-367.Geer, J. H. (1966). Effect of interstimulus intervals andrest-period length upon habituation of the orientingresponse. Journal of Experimental Psychology,Gingrich, K. J., & Byrne, J. H. (1985). Simulation of synapticdepression, posttetanic potentiation, and presynapticfacilitation of synaptic potentials from sensory neuronsmediating gill-withdrawal reflex in Aplysia. Journal of Neurophysioloay, 53, 652-669.Groves, P. M., & Thompson, R. F. (1970). Habituation: A dualprocess theory. Psychological Review, 77, 419-450.65Hawkins, R. D. (1988). A simple circuit model for higher-orderfeatures of classical conditioning. In J. H. Byrne and W.0. Berry (Eds.), Neural models of plasticity (pp.73-93).San Diego: Academic Press.Hodgkin, J. , Edgely, M., Riddle, D. L., & Albertson, D. G.(1988). Genetics. In W. B. Wood (Ed.), The nematode Caenorhabditis elegans (pp. 491-586). Cold Spring Harbor:Cold Spring Harbor Laboratory.Hodgkin, J., Horvitz, H. R., & Brenner, S. (1979).Nondisjunction mutants of the nematode Caenorhabditiselegans. Genetics, 91, 67-94.Kandel, E. R. (1976). Cellular basis of behavior: An introduction to behavioral neurobiology. San Francisco:Freeman.Klein, M., Shapiro, E., & Kandel, E. R. (1980). Synapticplasticity and the modulation of the calcium current.Journal of Experimental Biology, 89, 117-157.Laming, P. R., & McKinney, S. J. (1990). Habituation in goldfish(Carassius auratus) is impaired by increased interstimulusinterval, interval variability, and telencephalic ablation.Behavioral Neuroscience, 104(6), 869-875.Mah, B. K. (1991). An analysis of the tap withdrawl response inmale Caenorhabditis elegans. Unpublished master's thesis,University of British Columbia, Vancouver.Mah, B. K., & Rankin, C. H. (in press). An analysis ofbehavioral plasticity in male Caenorhabditis elegans.Behavioral and Neural Biology.66Mackworth, J. F. (1968). Vigilance, arousal and habituation.Psychological Review, 75, 308-322.Peeke, H. V. S., & Peeke, S. C. (1973). Habituation in fish withspecial reference to intraspecific agressive behavior. InH. V. S. Peeke & M. J. Herz (Eds.), Habituation (Vol. 1)(pp. 59-83). New York: Academic Press.Rankin, C. H., & Beck, C. D. 0. (1992). Caenorhabditis elegans:a simple systems approach to the genetics of behavior. InD. Goldwitz, D. Wahisten, & R. E. Wimer (Eds.), Techniques for the Genetic Analysis of Brain and Behavior (pp.445-463). New York: Elsevier Science Publishers BV.Rankin, C. H., Beck, C. D. 0., & Chiba, C. M. (1990).Caenorhabditis elegans: A new model system for learning andmemory. Behavioral Brain Research, 37, 89-92.Rankin, C. H., & Broster, B. S. (1990). Factors affectinghabituation and recovery from habituation in C. elegans.Society for Neuroscience Abstracts, 16, 626.Rankin, C. H., & Broster, B. S. (1992). Factors affectinghabituation and recovery from habituation in the nematodeCaenorhabditis elegans. Behavioral Neuroscience, 106(2),239-242.Rankin, C. H., & Chalfie, M. (1989). Analysis of non-associativelearning in C. elegans: I Neural circuit mutations. Societyfor Neuroscience Abstracts, 15, 1118.67Rankin, C. H., & Chiba, C. M. (1988). Short- and long-termlearning in the nematode C. elegans. Society for Neuroscience Abstracts, 14, 607.Rankin, C. H., & Wicks, S. R. (1991). Circuit analysis ofinteractions between two antagonistic reflexes in C.elegans. Society for Neuroscience Abstracts, 1/, 1055.Ratner, S. C. (1972). Habituation and retention of habituationin the leech (Macrobdella Decora). Journal of Comparative and Physiological Psychology, 81(1), 115-121.Ruchkin, D. S. (1965). An analysis of average responsecomputations based upon aperiodic stimuli. IEEETransactions of Biomedical Engineering, 2, 87-94.Sokolov, E. N. (1963). Higher nervous functions: The orientingreflex. Annual Review of Physiology, 25, 545-580.Sulston, J. E., & Brenner, S. (1974). The DNA of Caenorhabditiselegans. Genetics, 77, 95-104.Sulston, J. E., Schierenberg, E., White, J. G., & Thomson, J. N.(1983). The embryonic cell lineage of the nematodeCaenorhabditis elegans. Developmental Biology, 100, 64-119.Thompson, R. F., & Spencer, W. A. (1966). Habituation: A modelphenomenon for the study of neuronal subtrates of behavior.Psychological Review, 173, 16-43.Velleman, P. F., & Hoaglin, D. C. (1981). Applications, basics, and computing of exploratory data analysis. Boston: DuxburyPress.6869White, J. G., Southgate, E., Thomson, J. N., & Brenner, S.(1986). The structure of the nervous system ofCaenorhabditis elegans. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 111, 1-340.Whitlow, J. W., & Wagner, A. R. (1984). Memory and habituation.In H. V. S. Peeke & L. Petrinovich (Eds.), Habituation. Sensitization. and Behaviour (pp. 103-153). New York:Academic Press.Wood, D. C. (1970). Parametric studies of the response decrementproduced by mechanical stimuli in the protozoan, Stentorcoeruleus. Journal of Neurobiology, 1(3), 345-360.Wood, W. B. (Ed.) (1989). The nematode Caenorhabditis elegans.Cold Spring Harbor: Cold Spring Harbor Laboratory.Yerkes, A. W. (1906). Modification of behavior in Hydroideadianthus. Journal of Comparative and Neurological Psychology, 16, 441-450.Appendix IThe force imparted by the mechanical tapper was measuredusing a Showa strain-guage that was electrically connected to a6 V DC power supply, a wheatstone bridge and an oscilloscope.The strain guage was mounted on a metal cantilever such that anydeformation of the cantilever created by a force applied to oneend of it resulted in a deviation from zero of the oscilloscopereading. The peak amount of deviation created by the tapper wasthen matched by calibrated masses suspended from the cantilever(at the same point where the tapper force was applied) by acarriage of neglibile weight. The mass that matched the tapperdeflection was multiplied by 9.81 m/s 2 to give a force value inNewtons. The average value obtained over several readings was1.1 N.70Appendix IIThe order of the variable intervals (as listed below) foreach of the variable-ISI groups was chosen with the followingconsiderations in mind. The first and last 5-7 stimuli in eachset were chosen to be as varied in length as possible to avoid apossible effect where predominance of one type of interval couldhave a large impact on either the initial rate of habituation orthe rate of recovery. Also, any sort of regular pattern (e.g.,a larger interval every 4 or 5 stimuli) was avoided. Finally,having more than two short or long intervals in a row was alsoavoided.lOs ISI Group Variable Schedule: 10s, 2s, 15s, 5s, 10s, 5s,10s, 2s, 40s, 10s, 5s, 2s, 10s, 5s, 30s, 5s, 5s, 10s, 2s, 25s,5s, 10s, 2s, 25s, 5s, 10s, 2s, 15s, 2s, 5s, 25s, 2s, 10s, 20s,5s, 2s, 10s, 20s, 5s, 2s, 25s, 5s, 15s, 5s, 20s, 2s, 15s, 5s,10s, 2s, 30s, 5s, 15s, 5s, 20s, 2s, 15s, 5s, 10s.60s ISI Group Variable Schedule: 10s, 3m, 30s, 2m, 5s, 40s,10s, 60s, 4m, 5s, 2m, 30s, lOs, 20s, 40s, 30s, 10s, 2m, 30s, 5s,4m, 10s, 30s, 60s, 5s, 40s, 10s, 2m, 20s, 40s, 5s, 60s, 3m, 10s,40s, 2m, 30s, 5s, 60s, 30s, 10s, 40s, 2m, 5s, 60s, 30s, 3m, 10s,4m, 40s, 60s, 2m, 40s, 60s, 5s, 10s, 3m, 60s.71Appendix IIIThere appeared to be some degree of cyclicity in thefluctuation in response level during habituation of thefixed-interval 10-s ISI group. That is, the response levelseemed to slightly rise and fall in periodic fashion above andbelow the response level that would have been predicted by astraight regression line running through the data points. Thisphenomenon was further explored by performing an autocorrelationon the mean response levels for each stimulus during asymptote(stimuli 13 through 60). Only asymptote data was used, sincethis made the analysis simpler. An autocorrelation done on theraw data indicated that there was some measure of cyclicity inthe data. To make the period clearer, the data was smoothed,and the autocorrelation was recalculated (refer to Velleman &Hoaglin, 1981). The results are shown in Figure 10A.As can be seen in this graph, any one observation is highlycorrelated with the one that immediately follows it (lag of 1),and, with each subsequent stimulus the correlation goes downuntil about seven or eight stimuli removed, where thecorrelation is near zero. Through the eight stimuli that followthis point (lag of 9 through 16) the correlation becomes morenegative, before once again heading back toward zero. Thus, thefull period describing these data appears to be about 16 stimulilong, with peak autocorrelation values ranging from 0.867 to72-0.418. The periodicity was also evident in a plot of theseresponses once they were smoothed (refer to Figure 1013).73Fig. 10. A) Autocorrelation values (between each response andthose that follow it) for smoothed data from the last 48responses during habituation of the 10-s ISI fixed-interval group(n = 20). The lag is the number of stimuli removed for which thecorrelation has been calculated. B) A plot of the smoothed datafrom the asymptotic habituation stimuli of the 10-s ISIfixed-interval group in Experiment 1. These data are based onthe last 48 stimuli of habituation. Reversal response magnitudeis expressed in terms of raw mm.74A16151413121110^09876543211^ 0^ 1CORRELATION75B^20 -15 -10 -5010 20^30^40^50^60STIMULUS


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items