A comparison of instrumental reward and avoidance training with classical reinforcement technique in conditioning the eyelid response

A comparison of instrumental reward and avoidance training with classical reinforcement technique in conditioning the eyelid response) Wesley d. Hansehe. TULANE UNIVERSITY David A. Grant. UNIVERSITY OF WISCONSIN Ablltraet Successful conditioning of the human eyelid response is reported, using intrumental reward training. Instrumental reward training of the eyelid appears to lead to more resistance to extinction than that found following classical or avoidance training. When reward training is added to avoidance training resistance to extinction is increased, but no corresponding effect is found when reward training is added to classical reinforcement. The eyelid response is readily conditioned by classical methods where the delivery of the DeS, usually an airpuff to the cornea, is not contingent upon the S's response. Successful instrumental avoidance training has also been reported for the eyelid response by Logan (1951) and Kimble, Mann & Dufort (1955). Although the journals contain numerous instances of instrumental reward conditioning of human responses, earlier unpublished exploratory work using verbal instructions and reinforcement was unsuccessful in conditioning the eyelid response in the Wisconsin laboratory. The present study reports a successful application of instrumental reward training with symbolic reinforcement of the eyelid response, a comparison of instrumental reward and avoidance training with classical eyelid conditioning, and an investigation of the addition of instrumental reward reinforcement to avoidance and classical training of the eyelid. M4'tbod The basic apparatus and procedure was the same as that described by Hartman & Grant (1962). The S sat looking into a rectangular wooden enclosure, painted flat white, and illuminated to approximately 10 ml, so as to avoid dark adaptation. The S wore a modified basketball spectacle protector carrying the air jet, and microtorque potentiometer which was coupled to his eyelid by means of a wire, hypodermic tube, and Scotch tape. The eSwas always a 1 sec. illumination of a 10 cm diameter circular ground glass window to 42 m!. The classical DeS, when given, consisted of a puff of nitrogen delivered to the cornea 660 msec. after onset of the es. The puff duration was 100 msec. and the pressure at the source was 150 mm of Hg. The eyelid response and its first time-derivative were recorded by an inkwriting system. A voltage level pickoff unit operated a relay .05 sec. after the derivative voltage reached a low preset level. Timing circuits were arranged so that if a new or Psychon. Sci., 1%5, Vol. 2. additional eyelid closure occurred between 200 msec. and 660 msec. after onset of the es, the relay operation could be used to: (a) eliminate the presentation of the DeS, (b) provide a symbolic reward reinforcement for the response by means of illuminating a 50 watt bulb and operating a Veeder-Root electric counter in the direct view of the S, or (c) provide both DeS elimination plus symbolic reward reinforcement. In the classical conditioning procedure, the pickoff unit was not activated. All Ss received 60 acquisition trials followed by 25 extinction trials. The intertrial interval was varied randomly from 16 to 48 sec. and averaged 32 sec. The experimental design was an orthogonal variation of symbolic reward reinforcement (present or absent) with Des contingency (classical or noncontingent, avoidance or response-contingent, and no DeS). There were thus six experimental procedures, and 15 men and women students from Introductory Psychology classes were randomly assigned to each procedure. If symbolic reward training was to be used, the S was instructed, "I cannot tell you what you are supposed to do but I can say that when you have done the right thing the light bulb will light. When you have made a correct response the light bulb will light, and the counter will count the number of times the light bulb goes on." If symbolic reinforcement was not to be given, the S was given standard neutral instructions not to control his natural responses to the stimuli-to let his responses take care of themselves. R4'suits A eR was defined as any response of greater than one mm recorded magnitude which occurred in the interval from 200 msec. after the onset of the es and before the onset of the DeS. Acquisition and extinction curves for all groups are plotted in Fig. la, 1b, and 1c. Percent eRs are plotted in 10-trial blocks during acquiSition and 5-trial blocks during extinction. The initial point on all acquisition curves was computed from mean performance on the first two trials. Figure 1a shows that acquisition with symbolic reward training was intermediate between the superior acquisition with classical DeS training and the inferior acquisition with avoidance training. The latter result has been found by Logan (1951) and Kimbleet al (1955). Resistance to extinction seems greatest following symbolic training. In Fig. 1b it is apparent that addition 305 100 EXTINCTION ACQUISITION 90 Ia eo ~ :~j~ 50 ~~ ~ AVOIDANCE \. ~ 40 CONTROL BLOCKS OF TRIALS ~~__~__~L-___~L-__~4____~____~ 100 ACQUISITION CLASSICAL. PLUS 90 S1M . REtNf . I 2 ~ 4 • EXTINCTION Ib ----------~~--- BLOCKS OF TRIALS ~ 4 5 ~ 100 ::: /;~""' ACQUISITION :: ~ I /'~OANCE ::b 40 10 / / / CONTROL ~ BLOCKS OF TRIALS 1 2345 EXTINCTION Ie ~ Fig. 1. Acquisition and extinction curves for e yelid responses under six conditions of training. of symbolic reward training to classical training increased rate of acquisition slightly but seems to have no effect upon extinction, which resembles extinction following classical training alone more than it resembles extinction following reward training alone. In contrast, Fig. lc shows that adding symbolic reward training to avoidance training leads to greater resistance to extinction than that produced by avoidance training alone. In the analyses of variance of total acquisition performance and total extinction performance all main effects and interactions except one in extinction were statistically significant, but the extinction differences disappeared when analysis of covariance was used to control for final level of responding in acquisition. 306 DI8e0881on The primary novel result of the experiment was the successful symbolic reward training of the eyelid response by means of a simple procedure. The eyelid responses, thus produced, closely resembled the eyelid CRs occurring with classical airpuff reinforcement, although there was more voluntary form responses (Spence & Ross, 1959) than in the classical group. 2 This finding is extremely useful when for theoretical reasons a non-aversive reinforcement procedure need be used with the eyelid response. This success in reward training is probably due to the immediacy of the reinforcement. A second noteworthy feature of the results is that the eyelid CR appears to be relatively resistant to extinction following symbolic reward train (...truncated)