THESES SIS/LIBRARY TELEPHONE: +61 2 6125 4631 R.G. MENZIES LIBRARY BUILDING NO:2 FACSIMILE: +61 2 6125 4063 THE AUSTRALIAN NATIONAL UNIVERSITY EMAIL: library.theses@anu.edu.au CANBERRA ACT 0200 AUSTRALIA USE OF THESES This copy is supplied for purposes of private study and research only. Passages from the thesis may not be copied or closely paraphrased without the written consent of the author.
THE EFFECTS OF LATENT INHIBITORY PROCESSES ON SUBSEQUENT CONDITIONING: TEMPORAL FACTORS AND ORDER PHENOMENA Bernard J. Thorley This thesis was submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy in The Australian National University September 1968
ii This statement is to certify that the experiments described in thi.s thesis were my o~n original work. Bernard J. Thorley
iii TABLE OF CONTENTS Page ACKNOWLELGEMENTS LIST OF ABBREVIATIONS ABSTRACT viii ix X CHAPTER I CHAPTER II THEORETICAL BACKGROUND RELEVANT BACKGROUND RESEARCH Pre-exposure of the CS Table 2,1 Table 2,2 Related experiments Concluding comments 1 17 17 24 32 34 40 CHAPTER III GENERALIZATIONS, HYPOTHESES A.ND IMPLICATIONS The nature of the process underlying the effects of unreinforced pre-acquisition exposure to the CS Implications: Negative transfer and partial reinforcement Implications: Temporal variables Concluding comments 42 42 46 47 50 CHAPTER IV THE DEVELOPMENT AND RETENTION OF THE CER AS A FUNCTION OF PRE EXPOSURE TO THE CS A.ND PARTIAL REINFORCEMENT Classical, conditioning technique Rationale for the experiments Experiment I Experimental design Table 4.1 Subjects Apparatus Figure 4.1 Procedure Results Drinking rates in the absence of the CS 53 53 54 56 57 57 58 58 60 61 62 62
iv CHAPTER IV (Continued) Table 4,2 CS suppression of dr.inking behaviour Table 4.3 Retention Table 4.4 Discussion Experiment II Subjects Apparatus Procedure Results Drinking rates in the absence of the CS CS suppression of drinking behaviour Table 4.5 Table 4.6 Table 4.7 Retention Incidental observations Discussion Conclusions Pre-exposure to the offset of a stimulus Partial reinforcement Retention Page 63 64 65 67 68 69 71 72 73 7.3 74 74 74 75 76 78 79 80 81 83 83 83 84 CHAPTER V ORDER EFFECTS AND ITI SCHEDULES IN HUMAN EYELID CONDITIONING Fatigue factors associated with 85 trial spacing Control of fatigue factors 88 Initial unreinforced presentations 89 of the CS using human subjects Experiment I 90 Rationale for the experiment 90 Experimental design 91 Predictions 92 Subjects 92 Apparatus 92 Procedure 93 Figure 5.1 94 Results 96 Table 5.1 97 85
v CHAPTER V (Continued) Figure 5.2 Discussion Experiment II Rationale for the experiment Summary of the design of the experiment Subjects Apparatus Procedure Results Figure 5. 3 Table 5.2 Table 5.3 Discussion Conclusions Page 98 99 101 101 107 107 108 108 109 110 111 112 113 114 CHAPTER VI EXTENSION OF THE THEORETICAL SCHEMA TO A SIMPLE INSTRUMENTAL CONDIT.IONING PROCEDURE Fatigue factors in runway acquisition training The effect of partial reinforcement on runway acquisition Latent learning The rationale for experiment I Experiment I Table 6,1 Discussion of the independent variables Summary of the purposes of the experiment Subjects Apparatus Figure 6.1 Procedure Results Pre-training Figure 6.2 Acquisition Figure 6. 3 Figure 6.4 Table 6.2 Transfer as a function of reinforcement schedule Table 6,3 117 118 119 121 126 127 128 129 131 132 1.32 133 134 138 139 140 141 142 143 145 146 146
vi CHAPTER VI (Continued) Response decrement as a function of initial training and trial spacing Table 6,4 Table 6.5 Table 6,6 Table 6.7 Table 6, 8 Pre-training and the interval between subsequent rewarded trials Table 6.9 Extinction following continuous reinforcement Figure 6,5 Extinction following partial reinforcement Table 6.10 Figure 6,6 Table 6,11 Discussion Experiment II Table 6.12 Dependent variable Experimental design Subjects Apparatus Procedure Results Table 6.13 Figure 6,7 Figure 6,8 Table 6.14 Discussion Conclusions Table 6.15 Page 147 148 149 151 154 155 156 157 159 160 161 162 163 165 166 171 17.3 17.3 174 175 175 175 176 177 178 181 182 183 183 184 CHAPTER VII SUMMARY AND GENERAL CONCLUSIONS The latent inhibition paradigm as a basic negative transfer paradigm The general problem of order effects General conclusions from the classical conditioning research 188 189 194 197
vii CHAPTER VII (Continued) The extension to instrumental conditioning General implications Page 197 206 APPENDIX A APPENDIX B APPENDIX C REFERENCES CER EXPERIMENTS Table A.l Table A,2 HUMAN EYELID CONDITIONING EXPERIMENTS Table B.l Table B.2 Table B,_3 Table B.4 INSTRUMENTAL CONDITIONING (RUNWAY) EXPERIMENTS Table C.l Table C.2 Table C,J Table c.4 Table C.5 Table c.6 209 210 212 214 215 216 217 218 219 220 224 232 240 248 249 250
viii ACKNOWLEDGEMENTS The writer wishes to acknowledge the support and encouragement of the following people: The members of my supervisory committee, Dr J. R. Trotter, Dr W, H. Gladstone and }fr H. L. Cook, for advice and assistance during the course of the project, Professor C, A, Gibb :for assistance in obtaining equipment, supplies and experimental :facilities. Professor K. A. Provins for valuable criticisms o:f the manuscript. Hr Chia-chong Chen for invaluable assistance with the equipment and general advice on the management o:f the rats, Dr D. G, Beswick :for assistance with computer programmes which materially eased the computational burdens involved, Mrs S, Page and Mr P. G. Ward for statistical advice. Mr K. A. Rae for drawi.ng the sketches of the equipment. Hy wi:fe, Christobel, for constant help and support. The sta:f:f o:f the Australian National University Thesis Typing Scheme who typed and printed the thesis. The writer is indebted to the Australian National University for the award o:f a post-graduate research scholarship, and to the New South Wales Government :for granting study leave :for the period of the scholarship.
ix LIST OF ABBREVIATIONS The thesis: ~ollowing abbreviations have been used in this CER cps or Hz CR cs ITI Conditioned emotional response Cycles per second or hertz Conditioned response Conditioned stimulus Intertrial interval N Number o~ subjects PI RI s Ss Proactive inhibition Retroactive inhibition Standard deviation Subjects TFNR 'I'ri.al ( s) ~ollowing non-re in~orcemen t TFR us Trial(s) following reinforcement Unconditioned st:imulus
X ABSTRACT Some interference theories of extinction assert that the conditioned response decrement results from the attachment of incompatible associations to a stimulus which formerly had a strong tendency to evoke the conditioned response. In Chapter I, it is argued that this type of theory implies that the incompatible associations acqui.red via the extinction procedure will be subject to pro-active inhibition by virtue of the fact that extinction forms the second stage of an A-B, A-D negative transfer paradigm. On these grounds, it is possible to interpret the apparent retardation of extinction by spaced trial intervals, and the spontaneous recovery of the conditioned response with rest, as order effects resulting :from pro-active inhibition. The possible confounding of extinction phenomena by order effects limits understanding of the extinction process. This thesis stresses the importance of reversing the two stage conditioning-extinction sequence. An initial series of unrein:forced presentations of a stimulus :followed, at a later stage, by a series of reinforced presentations of that stimulus, constitutes a procedural reversal. This reversal is referred to herein as 1 the latent inhibition parad:igm', and is considered as a possible counterbalanced arrangement :for assessing order effects in. extinction. Studies, prior to the inception of the programme of research outl:ined in this thesis, were not viewed in this perspective. Nevertheless, these early studies, which are reviewed in Chapter II, had shown that initial unrein:forced
xi exposure of the CS inhibits subsequent conditioning, Moreover, it becomes apparent in the review of the literature, that this inhibitory effect is not necessarily contingent upon the pre-exposure and the conditioning phases being given in a single sessiont Hence, it seemed appropriate to infer that the inhibitory process, underlying the effects of initial unreinforced presentations of the CS, has elements of permanence (that is, learning characteristics), The research programme in this thesis represents a preliminary attempt to determine the relationship between the inhibitory effects of unreinforced preexposure to the CS, and those produced by the introduction of non-reinforced CS presentations during or following a classical conditioning series. The basic theoretical proposition is that a common.learning process, which interferes associatively with classical conditioning, underlies the effects of the pre-exposure, the partial reinforcement and the extinction procedures. This hypothesis implies that these three variables will be non-independent in their effects - an implication that can be tested by examining the conjoint action and interaction of the variables, However, the primary emphasis was placed on the order in which the mutually interfering associations (acquired by way of nonreinforced presentations of the CS on the one hand, and reinforced presentations of the CS on the other) are established, From this viewpoint, the effect of the duration of the time interval following a reinforced, or a non-reinforced trial, will depend upon whether previous training has provided a source of pro-active
xii inhibition. For example, as a consequenee of pre-exposure to the CS, the shorter the interval following a reinforced trial, the higher should be the probability of a conditioned response. The relative retention of the conditioned response, formed against a background of non-reinforced trials, should be poorer than usual. This form of analysis also suggested that the marked interfering effect of intermittent reinforcement on classical conditioning could be reduced or increased by appropriate manipulation of the intervals following reinforced or non-reinforced trials, Two experi.ments were conducted using the CER classical conditioning technique, A significant interaction was obtained between the effects of pre-exposure to the CS and the effects of intermittent reinforcement. Similar effects were prevalent using the interruption of an ongoing stimulus as the CS in one experiment, and the onset of a stimulus as the CS in the other. The retention of the CER was tested after a twenty-four hour rest period, There was some evidence for poorer retention as a consequence of pre-exposure to the CS, but the results were inconclusive because of the poorer conditioning level of the pre-exposed subjects. In a human eyelid conditioning experiment, an equal number of short and long duration intertri.al intervals were irregularly alternated. When conditioning was preceded by an initial series of non-reinforced trials, the subjects had a higher frequency of conditi.oned responses on trials followi.ng short, rather than long, intertrial intervals. This difference in frequency was significant when evaluated over the first half of the
xiii conditioning series, Control sub,jects responded with equal frequency on trials following short and long intertrial intervals. In comparison with the control group, the experimental group exhibited significantly poorer conditioning during the early stages - when the comparison was made in terms of performance on trials following long intervals. The difference was small and non-reliable when a similar comparison was made in terms of performance on trials following short intervals, In a second human eyelid conditioning experiment, two groups of subjects were given a 50 per cent random reinforcement conditioning series, Group SL had short intervals following reinforced trials and long intervals following non-reinforced trials. Group LS had the reverse contingencies - long i.ntervals following reinforcement and short intervals following nonreinforcement, Although the mean intertrial interval was the same in both groups, Group SL had a significantly higher frequency of conditioned responses than Group LS. This result was predicted on the basis of associative interference and pro-active inhibition. Collectively, the classical conditioning data were interpreted as favouring an interference theory of the effects of unreinforced CS presentations introduced prior to, or during, conditioning. In general, the effects of temporal variables were in accordance with the principles of pro-active inhibition. In the final phase of the research programme, an attempt was made to extend the concepts and variables, which had been applied to the latent inhibition paradigm, to an analogous instrumental conditioning paradigm.
xlv Thi.s entailed an investigation of the effects of an initial series of non-rewarded trials in a runway on a subsequent, instrumentally conditioned running response in that situation. The influence of initial non-rewarded trials was tested in conjunction with the effects of percentage reinforcement during acquisition, trial distribution during acquisition, and trial distribution during extinction, A significant latent learning (positive transfer) effect was obtained as a consequence of initial nonrewarded trials, Neither of the two trial distribution variables interacted significantly with the initial training variable, Intermittent reinforcement did not significantly reduce the degree of positive transfer. Nonetheless, the initial non-rewarded trials and the intermittent non-rewarded trials were not independent in their effects, A combination of the two treatments created greater resistance to extinction than did intermittent reinforcement in the absence of initial non-rewarded trials. Furthermore, the non-rewarded trials, interspersed in the early acquisition training, produced significantly sharper decrements in the conditioned running performance of rats which had previously undergone initial non-rewarded training. These large decrements indicated that the reintroduction of non-rewarded trials was reinstating incompatible associations acquired during initial training. The depressed performance which accompanied the reintroduction of non-rewarded trials was, however, extremely transient; with the restoration of reward there was a precipitous rise in the performance curve. Thus, as a result of initial non-rewarded training, two strong
XV contrasting response tendencies, which could be executed in rapid succession with little mutual interference, were associated wi,th the rewarded and the non-rewarded tr:ials of an intermittent reinforcement acquisition series. Possible reasons for the discrepanc:ies between the results obtained using the latent inhibition paradigm, and those obtained from the analogous instrumental conditioning procedure were discussed, These included (a) the complexity of the stimulus situation in instrumental conditloning as opposed to the discrete CS in classical conditioning, (b) opportunlties to practise segments of the instrumental response during initial non-rewarded exposure to the runway situation, and (c) the greater extent to which mediatlng reactions can influence instrumental condltioned responses. In view of the complexities of the instrumental runway situation, it was concluded that the problems of associative interference and order effects (as outlined earlier) might best be pursued in the realm of classical conditioning at this stage. Further research is necessary to establish the generality of the classical condition.ing phenomena obtained in the present research programme,