e3 é .3.— .- ‘ 1"“: LIBRARY Michigan State Univcrsity 'VHESIS ABSTRACT THE ROLE OF THE BAR IN THE ACQUISITION 0F BAR PRESS AVOIDANCE by Edward Rea ChristOphersen A considerable amount of research has been devoted to the bar press avoidance situation in rats. Specifically, numerous investi- gations have been made in an effort to discover any variable which would account for the relatively poor avoidance performance in the bar press situation as compared to avoidance performance in a wheel- turn or shuttle bar‘situation. The present eXperiment, based upon intensive pilot work, examined the effect, upon performance, of pairing the presentation of a retractable bar with the buzzer (CS). Three groups of six subjects were run in an automated bar press avoidance apparatus. Control group i resembled the typical avoidance schedule in which shocks are presented at fixed intervals unless a bar press is made, in which case the next shock is postponed. The bar, for control group i was always in the extended position, The conditions for the experimental group were identical to control group I except.that the bar was extended with the onset of the CS and retracted with the termination of the CS, i.e., the bar was present only when the CS was being presented. Control group 2 was similar to the experimental group except that, the bar, in addition to being presented with the CS, was presented irregularly throughout the eXperimentai session. Subjects were said to have reached criterion performance when Edward Rea ChristOphersen 2 they avoided 99% of the shocks two sessions in a row. The results indicated that the pairing of the presentation of the bar with the CS yielded significantly better performance than having the bar present throughout the session. The slowest subject in the experiment group reached criterion one session before the fastest subject in control group i. It was suggested that these results could be adributed to (l) eliminating bar-holding, (2) eliminating non-reinforced bar presses, or (3) the discriminative stimulus preperties acquired by the bar thru repeated pairing with the buzzer and the shock. A group with an lntermittendy presented bar was included in an effort to differentiate between the first and third possibilities. The data indicated that the introduction of the bar was a strong cue. Approved: £2€é? } ((2;T“"“\11 M. Ray Denny, ”filer Professof’i ( Date: fl/fire 2 3,1 (J6 THE ROLE OF THE BAR iN THE ACQUISITION 0F BAR PRESS AVOIDANCE BY Edward Rea ChristOphersen A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psycholoqy 1965 AF".-'. ~.4 .‘n'n' ‘ -: vs.» I ACKNOWLEDGMENT The author wishes to express his appreciation to Dr. H. Ray Denny, chairman of his committee, for guidance and assistance in conducting this research. Also he wishes to convey thanks to Drs. S.C. Ratner and J. Hunter for their helpful criticism and advice. The author also wishes to express his sincere gratitude to his wife, Nedean, who willingly gave time and energy to the preparation of this manuscript. TABLE OF CONTENTS Page INTRODUCTION 0 O O O O O O O O O O 0 METHOD 0 O O O 0 O O O O O C O O O 0 RESULTS 0 O O O O O O O O O O O O O O 0 DISCUSSION 0 O O O O O O O O O O O O O O O O O 0 '3 REFERENCES . . . . . . 0 O O O O O O O 0 O O O O ‘6 APPENDIX 0 O O O O O O O O O O O 0 iii LIST OF FIGURES FIGURE Page i A comparison of percent avoidance, mean i one standard deviation, for the experimental group and control group I. o e o o o o o o o e o e e o o o 8 2 A comparison of the number of sessions prior to reaching criterion for the experimental group and control group I. e a e o a o a o o o e o e o a o 9 3 Rate of acquisition of bar press avoidance for each individual subject in control group 2. . . . . . . lO 4 Rate of acquisition of bar press avoidance for each individual subject in the experimental group. . . . . ll 5 Rate of acquisition of bar press avoidance for each individual subject in control group I. . . . . . . l2 Introduction Discrimination avoidance procedures are used extensively by the behavior scientist. In such procedures, an animal is trained to perform a response in the presence of a stimulus which signals the occurence of a noxious stimulus, such as shock, noise, wind, or a strong light. The response, which may consist of pressing a lever, or of running from one side of an alley to the other, terminates the signal and prevents the onset of the noxious stimulus, However, the reported difficulty of establishing a discriminative avoidance habit with rats in a lever-pressing apparatus severly limits the use of this favored experimental technique. Numerous investigations have been undertaken in an effort to identify the variables which account for the poor performance in bar-press avoidance; for example: the role of the tone-shock pairing in warm up, the role of the response in warm up and the role of shock intensity in warm up (Hoffman, Fleshler, and Chorny, l96l); the length of iTl and the length of the CS-UCS interval (Meyer, Cho, and Weseman, I960); UCS intensity (Stone l960; Kimble, l955; Boren, Sidman, and Herrnstein, l959); UCS duration (Chapamn and Bolles, l96h); CS termination verses UCS termination (Kamln, Campbell, Ryan, and Walker, l959); the effect of sensory feedback, post-response time, and reduction of the shock density (Bolles and Papp, l96h); systematic variation of the interval between the response and the onset of the pre-shock warning signal (Ulrich, Holz, and Azrin, l96h); avoidance with inescapable shock (Hurwitz, l96h). All of these studies, however, have yielded essentially negative results. There has to be some kind of procedural difference to account for good avoidance learning in the shuttle box or jump-out box while avoidance learning in a ban press situation remains so poor. Pilot work was conducted in an effort to find an explanation for this difference by attempting to get good bar press avoidance. One of the findings of the pilot work was that the rats tended to hold the bar down during the reSponse-shock interval(oniy the initial depression of the lever counted as a bar press response, holding counted as only one response). A search of the literature yielded a number of studies which reported this same finding. Dinsmoor, Matsuoka, and Hinograd (l958) examined bar-holding Specificially. Their data supported the hypothesis that holding behavior is maintained in part by its preparatory function, i.e.. sequences of behavior in the escape situation indicated that animals turned the shock off more quickly on trials when they were holding the bar down at the onset of shock than on trials when they were off the bar. Anger (personal communication, l965) has found that rats on a Sidman avoidance schedule hold the bar as much as 60% of the time. in the avoidance situation, any time the bar is pressed prior to the onset of the noxious stimulus, the noxious stimulus is avoided. But what of the situation where the subject holds the bar down for two trials in a row? The first trial is counted as an avoidance reSponse and postpones the shock, but on the second trial the subject is shocked for essentially the same response (continuing to hold the bar down). The present study was designed to see what would happen if the subject were not allowed to reSpohJ without reinforcement during the safe period 2 Thus, all responses which were made were reinforced. This was accomplished by using a retractable lever. The lever was in the test chamber only when a bar press could be reinforced; its intro- duction acted, then, as a conditioned stimulus. It was thought that the prevention of bar-holding, coupled with the discriminative prOperties (if any) of the presense of the bar and the prevention of any "unrewarded'l bar presses, would yield significantly better bar press avoidance learning than had previously been demonstrated. Two groups of subjects were run under identical conditions except that, in the experimental group the bar was retractable and in the control group l the bar was present all the time. A third group was run in an effort to differentiate between bar-holding, as such, and the discriminative pr0perties accompanying the introduction of the bar. Method Subjects Eighteen male albino rats, approximately 90 days old, were housed in individual cages and allowed free access to food and water. Apparatus The experimental Space measured (9“ x l0“ x l2%” high). Three sides were constructed of wood; the tOp and the door were of clear plastic, allowing unrestricted observation of the fig, The tap was mounted on switches so that when the subject hit the tap, shock was terminated, i.e., a “tap hit'I served as an escape reSponse. The floor of the compartment consisted of stainless steel rods, 3/32 in. in diameter and Spaced .5 in. apart. A G.E. 3l3 bulb (l.7 amp) on the back wall on the enclosure illuminated the chamber during the sessions. The reSponse lever was a Lehigh Valley Electronics Model lhOSH Retractable Lever; it was located 2 in. above the grid floor. A clearly audible non-aversive buzzer served as the warning signal (CS). The CS-UCS interval was 5 sec. The experimental compartment was contained within a sound attenuating chest. A blower served as a masking noise and as a ventilator. Direct current shocks (iOO volts) l.5 ma in intensity’were delivered through the grid floor until the bar was pressed or until 5 seconds elapsed, whichever occurred first. An Applegate constant current stimulator generated the shocks through a polarity scrambling circuit of the type devised by Hoffman (l962). Bar presses were “rewarded" by termination of the CS and UCS (if on) and also by turning on a safe light (2%in. above the bar) which denoted the "safe period”, i.e., absense from shock. 4 Procedure Experimental Group. During the first four sessions, the subjects (gs) were trained to respond by placing them in the apparatus with the bar extended. After 25 sec. the buzzer was presented. If no bar press occurred within 5 sec. the shock was introduced. If no bar press or “tap hit" occurred, the CS and UCS were tenminated at 5 sec., 2 sec. were allowed to elapse, and the CS was presented again. When a bar press occurred, the CS or CS and UCS were terminated, the bar was retracted, and the "safe light“ was turned on. The safe period duration was 25 sec., after which the bar was reintroduced, 2 sec. elapsed, and the CS was represented. 0n session 5, and all later sessions, the bar was introduced when the buzzer came on, stayed in as long as the buzzer was on, and was retracted when the buzzer was terminated. The re5ponse-shock interval of 32 sec., the shock-shock interval of 7 sec., and the 25 sec. safe light remained constant. The sessions lasted either 2 hours or until lOO shocks had been presented, which- ever occurred first. All bar presses turned on the safe light but only those bar presses occurring prior to the onset of the shock were counted as avoidance reSponses, i.e., if the bar were pressed while it was being retracted, there was immediate cessation of shock concurrent with the onset of the safe light. All §§_were run until they had reached the pre-set criterion of 99% avoidance or above for two sessions in a row. Control Group I. The treatment was identical to the experimental group except that the bar was always extended, i.e., it was never retracted. All §§_were run until they had reached criterion; those §§ not reaching criterion by the llth day were discontinued. Control Group 2. The treatment was identical to the experimental group except that the bar was extended and retracted according to the schedule listed in Table 4 of the Appendix. Also, on every trial, the bar was extended with the buzzer onset and retracted with the buzzer termination. The safe period was divided into A sec. segments during which the bar was either extended or retracted according to the schedule. Both the experimental and the control group subjects also had available, as an escape re5ponse, a ”tap hit”. That is, if the subject, while being shocked, hit the tap, that shock was terminated.I An additional experimental group was run, at a later date, without the ”tap hit“ escape reSponse. The rate of acquisition, however, was similar to the present experimental group. The difference between shocks prior to reaching criterion was not significant (t=.223, df- 7, 01+50> 2 > J'PO) e Results Figure l shows the mean percent avoidance : one atandard deviation for the eXperimental group and control group I (bar always extended). The entire experimental group was avoiding at least 95% of the shocks on the sixth and seventh day. Two subjects of control grOUp l reached criterion on the eighth day. All other subjects from control grOUp 1 reached criterion after the eighth session. Figure 2, a frequency polygraph of sessions to criterion, shows no overlap between the two groups. It should be noted that three of the control group i subjects had not reached criterion by the llth session and were discontinued. A t-test was run on those subjects which reached criterion to determine the significance of the difference in the number of shocks to criterion (t=2.87, df= lo, .010 > g > .005). (See Appendix, Tables l and 2). The experimental group subjects had significantly fewer shocks prior to reaching criterion. Figure 3 represents the rate of acquisition of bar press avoidance for individual subjects in control group 2. The subjects with a high level of avoidance were discriminating between the buzzer CS and the introduction of the bar. (See Appendix, Table 3). Excess bar presses were gradually eliminated, over sessions, in these subjects. However, two subjects showed little or no avoidance learning and no discrimination. Because of the high inter-subject variability, control group 2 was not compared directly with the experimental group or control group i. Figures # and 5 represent, reSpectively the rates of acquisition for individual subjects in the experimental group and in control group i. ._ mecca _ocucou vco ozoLm _oucoe_coaxu ecu ecu .co_um_>uo pcmvcoum oco u come .uucmo_o>m acoucoa mo c0m_cmaeou < ._ .m_u ~L§KK$LN\ mszzwufim. \\ 9 m Q R a m v m N x + Ii I. ifl JI ii :‘I 2 D N N «329216 I II naxrmw \Jxecnatulxw.nllllll __ i.. )5.__._.__._.__.. QJIIJ l QRSSSR ’57] Jzaflnuyquwwy.fuaua%{iggqy~y Q U\ Ahfisx .co_mmum uosu c. concoumcu coconut umzu eaocm o>_uuoamuc Lon muuomnam mo Loose: ecu mucomocaoc ue_oa zoom ._ dooLm .0cucou new mecca _mucoE_coaxo on» to» co_cuu_cu mc_:umoc o» co_ca mcommmom mo Logan: ecu mo :0m_cmqeou < .N .m_u zo_mmh.mu o» mzo_mmum +__ __ o. m m m m m a m _ n/ . \ .//. a \ llr /. L .x a / _ 9.5.5 .955“. III‘ 95.5 .mucoefoaxm. IO Aauanbaaa .N asoco .otucou c. auuaasm _a=u_>.ec. some tea oucoo_o>n mmoce can ac eo_u_m_:ouo mo uuom .m.m.u Acao;-~v mzo.mmum __ o. m m A o m a m N . SA. 4/ - \II*I .. .I . .laqil Av . o~-u > \ / \\ // \ / _\.o / \ ol \ --u _\\\O 3.0 o\\.o/ / . “WNW I I'll. C III-UlllD-I'IDIII|D\\ 0— ON on o: om om cm om om 1N3383d BDNVOIOAV l0 .a30cm .oucoe_coaxo on» c. uuomnam .oso_>_vc. coma Lo» oocov_o>e mmoca can mo co_u~m_:ouu mo «aux .: .m_u “Lao;-~v mzo_mmwm o. m m N m m. a m N _ J H ii 1 + q q I q 4 Iii 1 o— 1 ON m_uw m_um on i; A ~_nm o_um J o: __nu tom . Ow . on . cm 4 cm oo— BJNVOIOAV 1N3383d ._ aaocm .0cuc00 e. uuumnsm .moo_>_oc_ some Lea oucov_o>o mmoca can no co_u_m_:ouo mo ouox .m .m_m Atso;-~v mzo_mmmm —_ 0— m m N w m J m N — I w 4 H *w d. w a 1 d d .L L O— —Nlm IIID n ON ONIw IIId me IIIO w_-m .Illlllla- L on Elm IIl o_|w III. 3 Om om oh ow om OO— 1N3383d EONVOIOAV l1 ‘00 Discussion The data reported here have demonstrated that it is possible to achieve a high percentage of avoidance in the bar press situation. The control condition in which the bar was always extended was similar to previous work with bar press avoidance. The method and the apparatus were meant to reflect previous findings that resulted in better bar press avoidance. For example: Myers (l962) demonstrated superior learning with a buzzer warning signal over a light or a tone; Ulrich, Holz, and Azrin (l964) found that the use of a warning signal reduced the number of shocks received; Chapman and Bolles' (l96h) data argued for shock durations longer than brief shocks of .2 to .3 sec., hence the present study incorporated a 5 sec. maximum duration; Dinsmoor, Hughes, and Matsuoka (l958) found that learning was better if the subjects were required to released the bar, i.e., holding did not count as a reSponse on the next trial, this was incorporated in the present work; Boren, Sidman, and Herrnstein (1959), found that acquisition of the avoidance reSponse was facilitated by using shocks levels above l.2 ma hence the present use of l.5 ma. An additional variable, the“safe light“, was suggested by the findings of Ulrich, Holz, and Azrin, I964. All of the above mentioned procedural variables were included in control group i. The experimental group was identical to control group i except that the bar was extended with the warning signal and retracted when the warning signal terminated. This manipulation clearly facilitated the acquisition of the bar press reSponse in an avoidance situation. This facilitation can be attributed to several possibilities: (i) elimin- ating bar-holding as a competing reSponse, (2) eliminating non- 13 reinforced bar pressed, i.e., those bar presses that occur outside of the CS-UCS interval, and (3) the discriminative stimulus prOperties acquired by the bar thru exclusive pairing with the buzzer and the shock. The group with the intermittently presented bar was included in an effort to differentiate between the first and third poss- ibilities. The data didn't shed much light on possibilty (l), but did indicate that the introduction of the bar was a strong cue. Further consideration of the discriminative prOperties of the bar led to the examination of the training procedure for the group that had the bar extention paired with the buzzer. The warning signal came on and the bar was extended. When 5 sec. elapsed without a response, the shock was delivered. The bar remained extended until it was pressed; when pressed, the bar was retracted and the safe light was turned on for the safe period. The bar, in this situation, is used in a way similar to the way Ulrich, Holz, and Azrin (l96h) used the warning stimulus, i. e., the warning stimulus stayed on (the bar was in) until the response had been made so that, in agreement with Azrin, Holz, Hake, and Ayllon (i963), the safe period was made contingent upon a Specific reSponse and the safe period was selectively associated with the absence of shocks. The data from the present investigation are handled nicely by elicitation theory. In tenms of elicitation theory (Denny and Adelman, l956) conditional fear or anxiety results from the consistent fashion with which the autOno-ic reSponses accompany the attempt to escape. Cessation of the aversive stimulation, in this case, the bar press, elicits a relaxational - approach reSponse which is incipientiy lh conditioned to the immediately preceding stimuli of the bar press and the warning stimuli. With continued trials the relaxational- approach reSponse is consistently elicited and thereby strongly conditioned to the stimuli accompanying the bar press. The animal is not allowed to relax for sufficient time to allow extinction to occur because the relaxation cues come to be associated with the safe light. On those trials where the subject continues to relax after the termination of the safe light, relaxation is punished by the onset of the warning Stimulus and the introduction of the bar. The present procedure allows the subject to relax, during the safe light but terminates the relaxation with the beginning of the next trial. 15 REFERENCES Alrin, N.H., Holz, W.C., Hake, D.F., and Ayllon,T. Fixed ratio ecape reinforcement. J, exp, Anal, Behav., l963, §,hh9-#59. Bolles, R.C. and POpp, R.J.Jr. Parameters affecting the acquisi- tion of Sidman avoidance. Wax... i96ll, 1, 315-32]. Boren, J.J.. Sidman, M., and Herrnstein, R.J. Avoidance, escape, and extinction as functions of shock intensity. J, co 2. physiol, Esxchol.,l959, 5;, h20-h25. Chapman, JJA. and Bolles, R.C. Effect of UCS duration on classical avoidance learning of the bar press reSponse. Esxchol, Rep,, '96“: iii, 559-563. . Denny, M.R. and Adelman, H.M.. Elicitation Theory II: The formal theory and its application to instrumental escape and avoid- ance conditioning. Unpublished theoretical paper, Michigan State University, l956. Dinsmoor, J.A. Punishment: I. The avoidance hypothesis. Esxshgl. 531., __L. 3446. Dinsmoor, J.A.. Hugher. L.H., and Matsuoka, Y. Escape-from-shock training in a free-response situation. Amer. J. Psychol., 1953. 11; 325-337. Dinsmoor, J.A.. Matsuoka, Y., and Winograd, E. Bar-holding as a preparatory response in escape-from-shock training. JJ_§9mn. phxgigl, fisxchol., l958, 51g 637'539. Hoffman, H.S. and Fleshler, M. A relay sequencing device for scrambling grid shock. J, exp, Anal, Behav., i962,§, 329-330. Hoffman, H.S. Fleshler, M., and Chorny, H. Discriminated bar-press avoidance. J, ex2, Anal, Behav., l96l, 5, 309-316. Hurwitz, H. M. 8. Method for discriminative avoidance training. §cience, l964, 1&5” l070-l07l. Kamin, L., Campbell, 0., Judd, R., Ryan, T., and Walker, J. Two determinates of the emergence of anticipatory avoidance. J, comp, physiol, figychol., I959,.§2, 202-205. Kimble, G.A. Shock intensity and avoidance learning. J, co 2. h l S chol., l955, fig, 28l-28h. Meyer, D.R., Cho, C., and Wesemann, Ann. 0n problems of condit- ioning discriminated lever-press avoidance reSponses. Psychol. 3334., 1960, _6_]_, nil-228. l6 Myers, A.K. Effects of CS intensity and quality in avoidance condi- tioning. JLgcomp. physiolLLPsychol., l962, 55, 57-6l. Sidman, M. Avoidance conditioning with brief shock and no extero- ceptive warning signal. Science, I953 ll8 l57-l58. Sidman, M. 0n the persistence of avoidance behavior. J, abngrm. §oc, Esychoi., I955, 59, 2l7-220. Stone, G.C. Effects of some centrally acting drugs upon learning of escape and avoidance habits. J, comp, physiol, fisychol., l960, Sin 33-37. Ulrich, R.E., Holz, W.C., and Azrin, N.H. Stimulus control of avoidance behavior., J, exp, Anal, Behav,, i964,.z, l29-l33. l7 APPENDIX l8 SESSIONS . .co.mmom com .uoomaem coo oocoo_o>o wave can on» can .mxuozm mo Lease: on» .m_o_cu mo Lease: ugh .esoco _nucoe_coexu ._ mum< acooLom on» use .mxuocm mo consaz ecu .m_o_ch mo Loose: och ._o30co _ecu:oU .N m4mo «coo toe on» one .nxuozm mo cease: use .m_a_cu mo Loose: och .N eaoco .ocucoo .n mom