Operant GSR Conditioning with Cool Air Reinforcement H. D. KI~MEL AND F. W. GtnatrCHAra~I*
University of South Florida
Abstract-Two groups of 20 Ss each were run individually in a soundproof chamber heated to 115 F and humidified to 100~. Ss in the experimental group received brief presentation of cool air contingent upon emitted GSRs during 20 rain of acquisition. Control Ss were matched one at a time in operant rate and were yoked one at a time to receive the cool air on a response-independent schedule. Both groups increased significantly in rate of unelicited GSRs during acquisition, with the experimental Ss' curve rising more rapidly and both groups attaining equivalent terminal levels. During extinction, the groups maintained high levels of responding and did not differ. Examination of basal skin conductance data ruled out an activation hypothesis. In the control Ss, terminal response frequency was significantly con'elated with percentage of fortuitous response-contingent reinforcement. It was concluded that both groups were conditioned, with the controls receiving intermittent reinforcement sufficiently often to elevate their responding. The cool air was judged to be an effective reinforcer of unelicited GSRs, perhaps more effective than those used in previous studies.
A VAmETY OF POSlTXarE REINFORCERS have been employed in research on operant conditioning of the GSR in humans (Kimmel, 1974), including pleasant odors (Kimmel & Hill, 1960), lights presented to Ss in a dark room ( F o w l e r & Kimmel, 1962; Kimmel & Kimmel, 1963; Rice, 1966), tones signalling m o n e y earned (Shapiro, Crider & Tursky, 1964), candy and "good" (Kimmel, Pendergrass & Kimmel. 1967), class "points" ( G r e e n e & Nielsen, 1966), and pictures of nude women presented to male Ss (Schwartz & Johnson, 1968). Although the great majority of these studies have found some evidence of an operant conditioning effect, none of the rein* Based upon MA thesis done at the University of South Florida. Reprint requests: H. D. Kimmel, Department of Psychology, University of South Florida, Tampa, Florida 33620. 239
240
~rMM~_~ AND G U B U C ~ I
Pay. 3. BioL Sci. Oet.-Dee. 1975
forcers employed could be said to be highly effective in the sense that large increases in response frequency were generated. As has been noted in this connection by Katkin and Murray (1968), "Clearly, future research is needed to determine just which reinforcers are most effective in producing instrumental ANS conditioning. ~ In the present study Ss were seated in an uncomfortably hot and humid chamber and cool air was used as reinforcement of unelicited* GSRs. It was assumed that this type of primary reinforcement might be more effective than those previously used. Cool air has proved to be highly effective in conditioning skeletal operants in humans (Paul, Eriksen & Humphreys, 1962), and it has also been shown to be an effective US in classical GSR conditioning (Furedy, 1967). In addition to a group of Ss who received the cool air reinforeement contingent upon emitted GSRs, on a CRF schedule, yoked controls were run who separately received the cool air presentations on response-independent schedules temporarily identical to those actually administered to the CRF Ss to whom each was yoked. For comparative purposes, the present experimental procedures were modelled after previous ones (e.g., Coffman & Kimmel, 1971), with the exception of the use of a heated and humidified chamber and cool air as reinforcement. Method Forty male undergraduate students at the University of South Florida volunteered to serve as Ss. They received either $1.00 or two class points (Introductory Psychology), at their option. The Ss were assigned to experimental and yoked control groups as follows: During a 10 min baseline period, the average frequency of response per minute during the 5th through 8th minutes was obtained. On the basis of this measure of operant rate, Ss were placed into four different categories, 1-2, 3-5, 6-8, or 9-11 responses per minute. The first S to fall in a particular category became an experimental S; the next S in that category was his yoked control. This was repeated until there were 20 Ss in each group. The Ss were told that this was an experiment on physiological reactions to repeated stimulations, that periodic presentations of cool air would occur, and that they should remain as alert and motionless as possible and breathe evenly. These instructions were delivered via intercoau before the S was assigned to his experimental condition. * The term "unelicited" is used in this paper to mean that no known stimulus was deliberately correlated with such response,
Volume 10 Number 4
GSR C O N D I T I O N I N G
~-~l
Data were collected in a double-walled IAC chamber equipped with a standard office chair with padded armrests. A thermostatically controlled baseboard electric heater and humidifier kept chamber temperature at 115 F mad humidity at 100g. The cool air was delivered from an air conditioning unit outside the chamber via a 10 em duct which was fastened to an adjustable mount inside the chamber, allowing the end of the duct to be positioned in front of and slightly above the S's face, 12.7 cm away. An electrically operated valve permitted the cool air to enter the duct and reach the S's face. The GSR was picked up as a de resistance change from the palm and back of the S's right hand via 1.9 cm zinc-zinc sulphate electrodes in lucite cups filled with saline electrode paste. It was amplified by a Biophysical Instruments Company amplifier coupled to a Hewlett-Packard logarithmic converter, and recorded on a Texas Instruments Company RectiRiter with a paper speed of 7.6 cm/min. Reinforcement duration of 3 sec was electronically controlled and signalled on the recorder. An automatic 3 sec timeout followed each reinforcement. The occurrence of the reinforcement for the experimental Ss was also recorded as an audio signal on a tape recorder for subsequent playback and control of responseindependent reinforcements for the yoked controls. Upon arriving at the laboratory the S removed his shirt and the GSR electrodes were attached. He was then seated in the IAC chamber, the cool air duct was positioned, the chamber doors were closed, and the instructions read via the intercom. A baseline period of 10 min was then administered, during which the S was assigned to the experimental or control condition. Next, 20 rain of acquisition occurred. For the experimental Ss this involved visual monitoring by E of the GSR record and manual operation of a switch to deliver reinforcements as quickly as possible following criterion responses. The definition of a response and the reinforcement criterion were both 0.01 log microhms. For yoked controls, the record of reinforcements given to the experimental S was played back and E administered yoked reinforcements in response to each tone on the tape. The Ss were thus matched in operant level and yoked in acquisition temporal schedule and number of reinforcements. After acquisition, 10 rain of extinction was run and the experiment was concluded.
Results and Discussion The primary dependent measure was the number of criterion GSRs emitted per minute of baseline, acquisition, and extinction,
242
Pav. J. Biol. Sei. Oct.-Dec. 1975
K 1 M M E L AND GUflUCHARRI
2.1
2.0
Experimental
I
\
I 1.9
I
\
I
/
/
\\\ i I
I
E
/
I 1.8
-
q
//',....S Yoked Controls
\
University of South Florida
Abstract-Two groups of 20 Ss each were run individually in a soundproof chamber heated to 115 F and humidified to 100~. Ss in the experimental group received brief presentation of cool air contingent upon emitted GSRs during 20 rain of acquisition. Control Ss were matched one at a time in operant rate and were yoked one at a time to receive the cool air on a response-independent schedule. Both groups increased significantly in rate of unelicited GSRs during acquisition, with the experimental Ss' curve rising more rapidly and both groups attaining equivalent terminal levels. During extinction, the groups maintained high levels of responding and did not differ. Examination of basal skin conductance data ruled out an activation hypothesis. In the control Ss, terminal response frequency was significantly con'elated with percentage of fortuitous response-contingent reinforcement. It was concluded that both groups were conditioned, with the controls receiving intermittent reinforcement sufficiently often to elevate their responding. The cool air was judged to be an effective reinforcer of unelicited GSRs, perhaps more effective than those used in previous studies.
A VAmETY OF POSlTXarE REINFORCERS have been employed in research on operant conditioning of the GSR in humans (Kimmel, 1974), including pleasant odors (Kimmel & Hill, 1960), lights presented to Ss in a dark room ( F o w l e r & Kimmel, 1962; Kimmel & Kimmel, 1963; Rice, 1966), tones signalling m o n e y earned (Shapiro, Crider & Tursky, 1964), candy and "good" (Kimmel, Pendergrass & Kimmel. 1967), class "points" ( G r e e n e & Nielsen, 1966), and pictures of nude women presented to male Ss (Schwartz & Johnson, 1968). Although the great majority of these studies have found some evidence of an operant conditioning effect, none of the rein* Based upon MA thesis done at the University of South Florida. Reprint requests: H. D. Kimmel, Department of Psychology, University of South Florida, Tampa, Florida 33620. 239
240
~rMM~_~ AND G U B U C ~ I
Pay. 3. BioL Sci. Oet.-Dee. 1975
forcers employed could be said to be highly effective in the sense that large increases in response frequency were generated. As has been noted in this connection by Katkin and Murray (1968), "Clearly, future research is needed to determine just which reinforcers are most effective in producing instrumental ANS conditioning. ~ In the present study Ss were seated in an uncomfortably hot and humid chamber and cool air was used as reinforcement of unelicited* GSRs. It was assumed that this type of primary reinforcement might be more effective than those previously used. Cool air has proved to be highly effective in conditioning skeletal operants in humans (Paul, Eriksen & Humphreys, 1962), and it has also been shown to be an effective US in classical GSR conditioning (Furedy, 1967). In addition to a group of Ss who received the cool air reinforeement contingent upon emitted GSRs, on a CRF schedule, yoked controls were run who separately received the cool air presentations on response-independent schedules temporarily identical to those actually administered to the CRF Ss to whom each was yoked. For comparative purposes, the present experimental procedures were modelled after previous ones (e.g., Coffman & Kimmel, 1971), with the exception of the use of a heated and humidified chamber and cool air as reinforcement. Method Forty male undergraduate students at the University of South Florida volunteered to serve as Ss. They received either $1.00 or two class points (Introductory Psychology), at their option. The Ss were assigned to experimental and yoked control groups as follows: During a 10 min baseline period, the average frequency of response per minute during the 5th through 8th minutes was obtained. On the basis of this measure of operant rate, Ss were placed into four different categories, 1-2, 3-5, 6-8, or 9-11 responses per minute. The first S to fall in a particular category became an experimental S; the next S in that category was his yoked control. This was repeated until there were 20 Ss in each group. The Ss were told that this was an experiment on physiological reactions to repeated stimulations, that periodic presentations of cool air would occur, and that they should remain as alert and motionless as possible and breathe evenly. These instructions were delivered via intercoau before the S was assigned to his experimental condition. * The term "unelicited" is used in this paper to mean that no known stimulus was deliberately correlated with such response,
Volume 10 Number 4
GSR C O N D I T I O N I N G
~-~l
Data were collected in a double-walled IAC chamber equipped with a standard office chair with padded armrests. A thermostatically controlled baseboard electric heater and humidifier kept chamber temperature at 115 F mad humidity at 100g. The cool air was delivered from an air conditioning unit outside the chamber via a 10 em duct which was fastened to an adjustable mount inside the chamber, allowing the end of the duct to be positioned in front of and slightly above the S's face, 12.7 cm away. An electrically operated valve permitted the cool air to enter the duct and reach the S's face. The GSR was picked up as a de resistance change from the palm and back of the S's right hand via 1.9 cm zinc-zinc sulphate electrodes in lucite cups filled with saline electrode paste. It was amplified by a Biophysical Instruments Company amplifier coupled to a Hewlett-Packard logarithmic converter, and recorded on a Texas Instruments Company RectiRiter with a paper speed of 7.6 cm/min. Reinforcement duration of 3 sec was electronically controlled and signalled on the recorder. An automatic 3 sec timeout followed each reinforcement. The occurrence of the reinforcement for the experimental Ss was also recorded as an audio signal on a tape recorder for subsequent playback and control of responseindependent reinforcements for the yoked controls. Upon arriving at the laboratory the S removed his shirt and the GSR electrodes were attached. He was then seated in the IAC chamber, the cool air duct was positioned, the chamber doors were closed, and the instructions read via the intercom. A baseline period of 10 min was then administered, during which the S was assigned to the experimental or control condition. Next, 20 rain of acquisition occurred. For the experimental Ss this involved visual monitoring by E of the GSR record and manual operation of a switch to deliver reinforcements as quickly as possible following criterion responses. The definition of a response and the reinforcement criterion were both 0.01 log microhms. For yoked controls, the record of reinforcements given to the experimental S was played back and E administered yoked reinforcements in response to each tone on the tape. The Ss were thus matched in operant level and yoked in acquisition temporal schedule and number of reinforcements. After acquisition, 10 rain of extinction was run and the experiment was concluded.
Results and Discussion The primary dependent measure was the number of criterion GSRs emitted per minute of baseline, acquisition, and extinction,
242
Pav. J. Biol. Sei. Oct.-Dec. 1975
K 1 M M E L AND GUflUCHARRI
2.1
2.0
Experimental
I
\
I 1.9
I
\
I
/
/
\\\ i I
I
E
/
I 1.8
-
q
//',....S Yoked Controls
\
