W H \I‘W'HL ‘l s \ h L WI HI \ $ 1 WI \ ‘ ”IN _\ 4:. .p. Hm Ice moo THE CONE??§GYEL LEC FUNSTEGN C}? ”' ~35 €53»ch EF-i'i'EM’fixi. hesis fear {'31s {399708 03’ F"... A. l:(‘. Y; P ‘ 9C 1 r"irc-y\; ‘s .tr‘iv 0 ' ‘0 m' w o’t'licnf‘ 1.!"0 w Y ‘ 2.: 33.3‘q‘gi KW»: RHPQHSE 1N RATS» .115 A 3% mzévs LIBRARY Michigan State University THE CONDITIONED LICKING RESPONSE IN RATS AS A FUNCTICN OF THE CS-UCS INTERVAL By ROBERT BOICE ABSTRACT OF A MASTER'S THESIS COMPLETED SUMMER TERM, 1964 DEPARTMENT OF PSYCHOLOGY Using Weismanis new method of classically condi- tioning the licking response in rats, the acquisition of test trial responses and anticipatory CR's was compared over five CS-UCS intervals (0.5", 1", 2", 4", 6-). It was expected that the Optimal interval would be longer than one half second because of the nature of the UCR and the type of organism.employed. In addition, a comparison between two strains of rats, albino and grey, was made. The results were as follows: (a) The Optimal inter- val was two seconds with four seconds being nearly as good. (b) The one-half, one, and six second groups were not sig- nificantly different from one another. (c) The two Optimal intervals were significantly better than each of the other three intervals. (d) Data for CR's and test trials indica- ted the same trends although the differences were greater Robert Boice for the CR's. Reasons for that difference were discussed. (e) There was no genetic difference or interaction over Approved: 27?, flfi (92% Committee (gairman / Date: 629?: A}; /f F ‘/ temporal groups. Thesis Committee: M. Ray Denny, Chairman John King Thomas Nelson THE CONDITIONED LICKING RESPONSE IN RATS AS A FUNCTION OF THE CSAUCS INTERVAL By ROBERT 130103 A THESIS Submitted to the College of Social Science of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1964 ACKNOWLEDGMENTS The author wishes to eXpress his appreciation to Dr. M; Ray Denny, chainman of his committee, for guidance and assis- tance in the planning of this research. Also he wishes to convey thanks to Mr. Ronald Weisman for his patient advice and cri ticism. ii Introduction 149131106. 0 0 0 Results . . Discussion . Summary. . . Bibliography TABLE OF CONTENTS iii 11 18 21 22 LIST OF TABLES Table Page 1. Experimental Studies of CS-UCS Interval EffeCts .‘OOOOOOCCOOOOOOOOOOOOO 2 2. Summaries of Analyses of Variance ...... l5 3. Summary of Duncan's Multiple Range TeSt oooooocooooooooocooocococoo 16-17 4. Base Rates of Random CR.Responding .... 19 iv 00.009.00.00 00... LIST OF FIGURES Figure Page 1. The Plastic Lickinngcx . . . . . . . . . 6 2. Response Tapes. . . . . . . . . . . . . . 9 3. Cumulative Learning Curves fer Test Trial Responses . . . . . . . . . . . . . 12 4. Parametric Functions by Days. . . . . . . 15 5. Total Parametric Functions. . . . . . . . 14 INTRODUCTION The notion that there is no single Optimal CSAUCS interval in classical conditioning dates back to Pavlov (1927) who fOund a variety of tenporal intervals up to five seconds to be equally efficacious. American investigations in the interim, however, have largely neglected using the consummatory response as a UCR. A preponderance of data from.eyelid and finger withdrawal conditioning studies has consistently indicated an Optimal one-half second.interval. Exceptions are evident in procedures utilizing the responses of pupillary dilation and GSR (Jones, 1962). In work with other organisms, an Optimal interval of a different length is often found. Table 1 lists a cross section of’fihe var- ious procedures and organisms employed in classical condi- tioning. (l) TABLE 1 EXPERIMENTAL STUDIES OF CS-UCS INTERVAL EFFECTS OPTIMAL INVEST IGATOR UCR ORGANISM C S-UCS INTERVAL Kimble (1947) Eyeblink Human .4” Kimble, Mann, and Dufort (1955) Eyeblink Human .5" Schneiderman and Nictitafing Gormezano (1964) Membrane Rabbit .25" Cohen (1950) Flexion Show 0 Kappauf and Schlosberg (1937) Respiration Rat 1.5” Woodward (1958) Dilation Human 1.5" Bierbaum (1959) GSR Human 3' Jones (1962)) GSR Human 1" Spooner and Finger 3911088 (1947) Withdrawal Human .5” Wolfe (1930) Finger Withdrawal Human . 5 " Pavlov (1927) Salivation Dog 0" - 5" Goldstein _e_j:_ a1. Nictitating (1964) Membrane Fro g 2" (2) .- C-u: (3) In recent work at this university, R. G. Weisman has develOped a.new method of classical conditioning in the rat. The UCR is the consummatory response of licking for water. Weisman's doctoral thesis (1964) is based on a con- current technique which allows comparisons of classical conditioning and instrumental conditioning using the same licking response. His data on successive acquisitions and extinctions, on partial reinforcement effects, and the con- trols for pseudo-conditioning and superstitious behavior combine to demonstrate evidence fer a genuine dassical con- ditioning procedure. Indications that the licking response in the rat is highly reflexive (a good UCR) can be found in studies by Davis and Keehn (1959.) and Keehn and Arnold (1960). They found the mean licking rate in adult rats to be a consis- tent six or seven licks per second -- over various levels of thirst, for water, sucrose, saccharin, and saline solu- tions. The question of innateness was answered by Schaffer and Premack (1961) who observed weanling rats to lick in the same range as adults upon first contact with water. With.a new procedure of classical conditioning available and the obvious critics awaiting data (Kimble, 1964), the need for a CSAUCS interval study was present. A standardized CS-UCS study would permit comparisons with other techniques and also establish.the most efficient ISI for this method. The apparatus designed by Weisman was Kl» .19.— (4) easily modified to allow variance of stimulus intervals and confinement of the _S_ to insure minimal competing responses during the delay of reinforcement period (after Carlton -- in Spence, 1956). In view of the research done with the other consummatory response, salivation, the expectancy of results was only that the Optimal interval need not be 0.5 seconds to reflect a true classical conditioning method. 4‘ METHOD Subjects — The §fs were forty experimentally naive rats about four months of age, which were divided equally into two groups of Long-Evans greys and Spartan Colony albinos. Strict conditions Of deprivation.were imposed on all §fs as follows: First, five days of ten minutes of water per day; second, two days of four minutes of water per day; third, on training days, only two minutes of water outside the apparatus. All §fs were maintained in individual cages fer the twelve day period with.§d lib. fOod. Each animal was run.at approximately the same time each day, and all train- ing was administered in the evening. Apparatus - The apparatus used was basically that built by Weisman to test concurrent classical and instrumental condi- tioning. Two matched systems were incorporated to allow the simultaneous conditioning of two animals. During the experimentation, each licking apparatus was enclosed in a converted refrigerator to insure prOper sound insulation. Each refrigerator contained a 60 ft./min. exhaust fan and a white noise speaker for masking apparatus clicks. An ideal- ized representation of one of the "licking boxes" is shown in Figure l. The wooden insert in the licking box was de- signed to permit more confinement of the animal that in Weisman's case. The resultant interior dimensions were five inches in length, one and three-quarters inches in (5) 1‘ .- v '. I l . U ‘ c . 3“ p f '1 a L K I A t . . ' . (6) xom 02:83 9515 mi Hesse .7 :m a. MQOO wmo I a news fie: ifizuj . a , 2.? c m. . . l m” \n ex. 4 :\ :KH Hmmmz 2w0003 (7) width and four inches in height. The drinking well side and front face were constructed of plexiglass. The floor consisted of a one-eighth inch.stainless steel grid. The drinking well was a fully enclosed one and one- half inch extension on the front face of the box. Water was presented in small quantities, controlled by a sole- noid valve, one drOp per trial through a #11 needle (ground flat and smooth) which.projected through a small Opening at the bottom of the well. A small copper ring encircled this Opening to prevent gnawing and to measure licking. The CS ‘was a ten watt bulb mounted on the outside of the box next to the well. The general level of illumination in the boxes was 5 ft. Co Licking was measured by Grason—Stadler drinkometer, Model B 4690A, and recorded with a Gerbrands event recorder. Hunter timers and a Gerbrands punch tape timer provided the temporal intervals. A stepping switch was installed to allow randomized test trials. Design - The two experimental groups of albinos and greys were randomly divided into five groups of fbur each, making a total of five interval groups with eightlg's each. The five CSéUCS intervals chosen on the basis of pilot work were one-half’second, one second, two seconds, four seconds, and six seconds. All CS-UCS presentations were of the de- layed type and had a two second overlap. (8) Procedure - The basic procedure consisted of fOur days of training for all S's. The animals were habituated on the first day in order to familiarize the animal with.the apparatus and to insure that the UCS (water) would be a consistent elicitor of licking, once acquisition trials were begun. The criterion of habituation for all §fs was .fifty UCR's on a one-hundred and twenty second v.1. sched- ule with no CS present. The three days of conditioning followed immediately, utilizing various ISI intervals as previously indicated. Two pairs of S's were run each evening and the three hour session for each had trials presented on an ITI of one-hundred and twenty seconds, v.1. Test trials were inserted on the average of one in ten trials during the three ninety-trial sessions. Measures - The two basic measures of conditioned responses were recorded on paper tape by the event recorder. Two samples of that tape appear in Figure 2. FIGURE 2 RESPONSE TAPES SHOWING A CR AND TEST TRIAL RESPONSE ch T‘P‘ : Dwec‘hon ‘ -«______ (1 Mm. I See.) ”(3 .{ a s i 4 L U QB- _.__._- ___,___ __ _ ____-.. L414,” :1 _ .._._._ .- --,__..._-._---_-l__-__._._ _____l __,.. L (L5 (.7... _, . .._--.,__,4 n ..., “.51.... - ___--_l_ __ 1'.» Thai C No 9515) (9) (10) Anticipatory responses were scored as CR's on the non-test trials if they occurred during the CS-UCS interval and were discriminatory in nature (not part of intertrial respond- ing). To compensate for the obvious disadvantage of measuring anticipatory CR's with the two shortest intervals, test trials were instituted as the primary measure of per- formance. The criteria for a test trial response were that it also be discriminatory in nature and occur during the presence of the CS. .‘Both measures, CR's and test trials, were used for all groups. RESULTS The results are graphically presented in terms of cumulative acquisition curves (Figure 3), as parametric fUnctions by days (Figure 4), and as total parametric func- tions (Figure 5). In Figure 5 test trial responses were totaled for all three days of acquisition, whereas CR re- sponses represent only the third day of acquisition. A visual examination of the three graphs indicates a.marked superiority of the two and four second groups in acquisition rates and final levels of conditioning. In fact, two seconds appears to be generally Optimal - although probably not significantly so. The one-half, one, and six second intervals show no important differences at any level. The OR levels for those three intervals are much lower than the corresponding test trial levels in Figure 5. Test trial functions for the three days in Figure 4 seem to maintain the same general relationship among the intervals. In the two and four second groups, the level of anticipatory CR re- sponding on the third day was seventy-four percent. The CR function in Figure 5 does not include the first one hundred and eighty trials because the three less Optimal groups showed little conditioning (CR's) during that period. The design of the experiment was a 2 X S factorial with an organismic variable. The same appropriate analysis (11) -. . . .uoo ' . . . V . . ‘ K . . T ’ r— c u I Q ' . i Q o . . l I . ‘1. t e . '. I I _ , . I ' ' . ‘ (12) 7 mmmzommmm H40 H“. Ilv mumzodmum w>§<43230 low low 100. a. I 0.? I om vow WOON (l3) IOMD1' I§I'CMNY so- \/_\ _.| I/lén in 2|! 4|! 6' g IOOT O 2ND DAY *- LL. 0 so- '— Z 3 . . . . . m I/zu 'u 2» 4:: 6:: LIJ CL. IOC)j 38.12 DAY 50‘ I/‘zlt 1Ill 5" 4"“ All (25- UCS INTERVAL (SECONDS) FIGURE 4. PARAMETRIC FUNCTIONS BY DAYS PERCENT OF TOTAL (14) IOO " 75" 50- / / ANTICIPATORY) 25 - / CR‘s ,J (a no DAY) J2“ p: £91 4p: 6" CS "UCS INTERVAL (s ECONDS) FIGURE 5. TOTAL PARAMETRIC FUNCTIONS (15) of variance was applied to both sets of data. The summary tables for both analyses appear in Table 2. TABLE 2 SUMMARY OF ANALYSIS OF VARIANCE FOR TOTAL NUMBER OF TEST TRIAL RESPONSES SQURCE__OF VARIATION SS (1. f. M.S. F 1) Between Genetic 18.8 1 18.8 Temporal 856.11; 4 2L4.0 8.17 .005 G X T 65.1 4 15.8 Within :73; 30 24-4 TOTAL 39 SUMMARY OF ANALYSIS OF VARIANCE FOR THIRD DAY ANTICIPATORY CR'S SOURCE OF VARIATION SS d. f. M.S. F p Between Genetic l4 :1: 14 Temporal 19L742 4 4.936 50.85 .005 G X T 139 4 35 Within 4. 809 30 460 TOTAL 59 (16) The organismic variable produced no significant difference between albino and grey rate over temporal groups and no interaction with tenporal groups. For both responses the intervals produced significant differences beyond the .005 level. To determine the differences between the group means, Duncanis Multiple Range Test was administered. The results of the comparisons appear in Table 3. Group Means A 11.8 B 12.4 C 22.5 D 20.0 E 11.8 TABLE 3 SUMMARY OF MULTIPLE COMPARISONS BETWEEN INTERVAL GROUP MEANS FOR TOTAL TEST TRIAL RESPONSES APPLYING DUNCAN'S MULTIPLE RANGE TEST %"Group l"Group 2"Group 4"Group 6”Group Shortest C E Significant 11.8 12.4 22.5 20.0 11.8 (.gfnfgjel) .6 10.7 8.2 O 6.8 10.1 7.6 .6 7.1 2.5 10.7 7.3 8.2 7.4 SUMMARY OF MULTIPLE COMPARISONS BETWEEN INTERVAL GROUP MEANS FOR THIRD DAY CR'S APPLYING DUNCAN'S MULTIPLE RANGE TEST %"Group l"Group 2”Gr0up 4"Group 6"Group Shortest Significant $2.3: sf. 1:. .22 58?. 159 (are... a 13.1 .8 49.2 45.4 5.8 17.23 B 12.5 50.0 46.2 6.6 17.88 c 62.3 5.8 45.4 18.42 D 58.5 39.6 18.83 E 18.9 As indicated in all previous comparisons, the means for both measures were significantly higher in the two and four second groups. The two Optimal intervals were not signif- icantly different according to Duncan's Test. (17) DISCUSSION Effects of CSAUCS interval on test trial response acquisition - Hughes and Schlosberg (1958) found asymptotic levels of conditioning in rats to occur at about seventy- five trials. This investigation used a different modality and technique and the peak of performance was not reached in the two second group until beyond two hundred trials. The three non-Optimal intervals reached only about a fifty percent level of conditioned responding during test trials. An interesting characteristic of these three groups was the inconsistency of responding over trials which resembled a sort of inattention on the part of the‘g. The two and four second groups, on the other hand, displayed almost perfect consistency in responding over all three hour sessions. Gormezano (1964) has mentioned some similar lapses in uncon- ditioned salivary responding. The effect of introducing some partial reinforcement during acquisition by omitting one tenth of the UCS's was felt to be negligible in view of extensive pilot work where no differences were observed. Effects of CS-UCS interval on anticipatory CR acquisition - The fact that the shorter interval CR's reached a much lower level than their test trial counterparts was expected because longer temporal intervals allow more ran- dom licks and because of the latency of the licking response (18) (19) when the animal's head is slightly removed from the licker. Base rates for responses randomly falling in the CS-UCS interval during habituation (no CS) appear in Table 4. TABLE 4 BASE RATES OF RANDOM CR RESPONDING INTERVAL RATE__ 0.5” -- 2 a l n -- 4 % 2n -- as 4" u 8% 6 n - 7 a The marked inferiority of CR‘s as compared to test trials in the six second group was a bit more surprising. The cumulative function for the CR's (not shown) in the longest interval showed a levelingeoff to the rate of the shorter intervals. Examination of the latency of response on the paper tapes indicated that a process similar to Pavlov's inhibition of delay might be the cause for the dec- rement in later trials. Pavlov (1927) mentioned that with intervals over five seconds the onset of the UCR is delayed in a manner prOportional to the interval's length. In other words, when the animal learns to wait for the U08 before licking, there will be a decrease in anticipatory CR's. (20) Effects of genetic strain on conditioning measures - The analyses of variance and a visual inepection of the data indicate a.remarkable similarity in performance between.the grey and albino rats. Theoretical implications - As with much previous work which either classically conditioned a consummatory re- sponse or used subhuman organisms as'fi's, an Optimal inter- val between the onset of the CS and UCS of more than one second was found. Pavlov (1927) considered intervals up to five seconds to be almost equally efficacious. This study singled out two and four second temporal intervals as being superior. Kappauf and Schlosberg (1937) found an Optimum of over one second in rats. Some superficial similarity of these results can be noted in instrumental studies where the ISI has been inves- tigated. Berah (1951), for example, found half-second, one second, and two second intervals to similarly influence the establishment of a secondary reinforcer. Meet important in making a comparison to instrumental conditioning is the con- sideration that almost all delay of reinforcement (ISI) studies in Operant techniques show a direct inverse relation of performance to interval in early learning if any compet- ing responses are available (Spence, 1956). SUMMARY Using Weisman's new method of classically condition- ing the licking response in rats, the acquisition of test trial responses and anticipatory CR's was compared over five CS-UCS intervals (0.5”, l", 2", 4',6”). It was eXpected that the Optimal interval would be longer than one half sec- ond because of the nature of the UCR and the type of organism.employed. In addition, a comparison between two strains of rats, albino and grey, was made. The results were as follows: (a) The Optimal interval was two seconds with four seconds being nearly as good. (b) The one-half, one, and six second groups were not sig- nificantly different from.0ne another. (c) The two optimal intervals were significantly better than each of the other three intervals. (d) Data for CRfls and test trials indicat- ed the same trends although.the differences were greater for the CR's. Reasons for that difference were discussed. (e) There was no genetic difference or interaction over tem- poral groups. (21) BIBLIOGRAPHY Bersh, P. J. The influence of two variables upon the estab- lishment of a secondary reinforcer for Operant responses. g. 935p. Psychol., 41, 62-73. Bierbaum, 14.3. The temporal gradient in GSR conditioning. .J... gen. PSYChOIQ, 1958, 22, 97-1030 Cohen, J. Observations on strictly simultaneous conditioned reflexgs. g. comp. ppysio . Psychol., 1950, 42, 211-21 . Davis, J.D., and Keehn, J.D. Magnitude of reinforcement and consummatory behavior. Science, 1959, 1:2, 269. Gerall, A.A., and Woodward, J.K., Conditioning of the human pupillary dilation response as a function of the 05"ch interval. :10 8X2. P82Ch010, 1958' 22, 501‘ 507. Goldstein, A.C., Spies, G., and Sepinwall, J. Conditioning of the nictitating membrane in the frog. ,1. comp. 22128.10 0 PBEChOIO, 19649 51, 456“4580 Gormezano, I. Classical conditioning. In J..'B. Sidowski (Ed.) Experimental methods and instrumentation _ip psychology. New YorkT McGraw-Hill, 1964. Jones, J.E. The CS-UCS interval in conditioning short and long-latency responses. g. exp. Psychol., 1961, §_2_, 612—617. Kappauf, W.E., and Schlosberg, H. Conditioned responses in the white rat. III. Conditioning as a function of the length of the period of delay. ,1. genet. P82011010, 1937’ 2, 27-450 Keehn, J.D., and Arnold, E.M.M. Licking rates in albino rats. SCience, 1960’ 122, 739-7410 Kimble, G.A. Conditioning as a function of the time between conditioned and unconditioned stimuli. g. __§p. PSYChOlc, 1947,‘21’ 1-150 Hilgard and Mar uis’ conditioning. New York: Century Croft, 1961. Comment. P§ychon. Sci., 1964, l, 40. (22) '8 (O r. a. (O t. a. (23) Kimble,ig.A., Mann, L.I., and Dufort, R.H. Classical and strumental eyelid conditioning. g. m. ngchol. 1955. 59, 407-417. ’ McAllister, W. Eyelid conditioning as a function of the CS-US interval. ,1. app. ngcholn 1953, 42, 417-422. Moeller, G. The CS-UCS interval in GSR conditioning. :10 6X2. PSRChOlo, 1954, fig, 162-1660 Pavlov, I.P. Conditioned reflexes. (Translated by G.V. Anrep) London: Oxford Univ. Press, 1927. Schaeffer, R.W., and Premack, D. Licking rates in infant albino rats. Science, 1961, 154, 1980-1981. Schneiderman, N., and Bormezano, I. Conditioning of the nictitating membrane of the rabbit as a function of CS-US interval. ,1. comp. ppysio . Psychol., 1964, 51, 188-195. Spence, K.W. Behavior theory and conditioning. New Haven: Yale UnIversIty Press, 1956. Spooner, A., and Kellogg,W.N. The backward-conditioning curve. Amer. g. Paychol., 1947, _6_Q, 321-534. Weisman, R.G. A new method of classical conditioning in the rat: Comparisons with an instrumental conditioning technique using the same response. (Unpublished doctoral dissertation. Michigan State Univ., 1964.) Wolfe, H.M., Time factors in conditioning finger withdrawal. ,1. gen. ngchol” 1950, 4, 572-578. g I n. AX. . O 9 . a o f O O 0 a n J I o u D n O o o 0- . Ix . I O . I u C c I a J b O 9 . .r. a Q 0 h .. . .. O O I 0 e 0 0 fl. 0‘ o o 3 v o I I n i n a. n O A I R u 0 f I I u 0 P\ . p O o .9. d . . u n, ' A Rfi‘fi‘t USE Gs‘éfiLY -~. ’ “T ‘. :~ I " ‘uiiiv‘ ”'TITifilfifl’lfilfl'lfii fifigflifiififlififltgiflfflflfi'“