Effects of Biofeedback on the Discrimination of Electrodermal Activity 1

Transcription

1 Biofeedback and Self-Regulation, VoL 2, No. 4, 1977 Effects of Biofeedback on the Discrimination of Electrodermal Activity 1 J. Michael Lacroix ~ Glendon College, York University Twenty-four subjects were tested on their abifity to discriminate between the presence and absence of negative skin potential responses before and after training to control skin potential. Training consisted of 52 discrete 30-second trials during which subjects were asked either to increase or to inhibit palmar sweating. Subjects in groups N and P were provided with analogue feedback on their skin potential activity. Group N was correctly informed that increases in sweating were indicated by increases in the negativity of skin potential; group P was misinformed that these were indicated by increases in the positivity of skin potential Subjects in the control (C) group received no feedback. Reliable evidence of discrimination was obtained only in groups N and P, following training. However, reliable evidence of control was obtained only in group N. Thus, training to control skin potential led to an ability to identify afferentation associated with the more common (i.e., negative) skin potential responses, even though biofeedback training appeared unsuccessful in the case of group P. These findings are discussed in the context of "'discrimination" or "'awareness" accounts of the process of acquiring control of internal responses. While considerable evidence indicates that a number of autonomic responses can be brought under "self-control" or "voluntary" control as a consequence of biofeedback training, the process by which subjects learn to produce appropriate changes in these responses remains unclear. The nature of this acquisition process is now beginning to generate speculation, however, and has been the focus of much recent discussion by Brener (1974a, b, in press; see also Black, Cott, & Pavloski, in press; Roberts, in press). 'This work was supported by grants from York University and from Glendon College. 2Address all correspondence to Dr. J. Michael Lacroix, Department of Psychology, Glendon College, York University, 2275 Bayview Avenue, Toronto, Ontario, Canada M4N 3M This journal is copyrighted by Plenum. Each article is available for $7.50 from Plenum Publishing Corporation, 227 West 17th Street, New York, N.Y

2 394 Lacroix Briefly, Brener views the process of acquiring "voluntary" control of a response essentially as a process of learned discrimination: subjects must learn to discriminate and to identify changes in interoceptive afferentation related to changes in the target autonomic response in order to be able to produce changes in that response. With respect to most autonomic responses, this learning to discriminate changes in interoceptive afferentation requires that exteroceptive feedback be provided contingent upon changes in the target response. Exteroceptive feedback, which is readily discriminable, is correlated with feedback from interoceptive sources and serves to identify this interoceptive feedback. The development of "voluntary" control is viewed as a consequence of this discrimination process. This "discrimination" or "awareness" account of the acquisition process predicts that subjects should be unable to discriminate changes in a response that they cannot control, but that they should be able to discriminate changes in the same response following successful feedback training to control that response. These predictions were examined in the present experiment, with the electrodermal system. This system seemed particularly suited to such an investigation in view of evidence that provision of exteroceptive feedback appears essential to the development of electrodermal control (Lacroix & Roberts, 1976). The question of whether or not subjects can discriminate electrodermal activity prior to and following biofeedback training has been addressed in previous experiments by Baron (1966), Stern (1972), and Diekhoff (1976). Baron's and Stern's results suggest that subjects cannot discriminate electrodermal responses in the absence of training with exteroceptive feedback. On the other hand, Diekhoff observed that subjects could discriminate successfully between the presence and the absence of electrodermal responses in the absence of feedback training, but that they tended to label the two response states incorrectly. However, successful discrimination performance in that experiment may have resulted from a feedback contingency that was implicit in the discrimination task, as Roberts (in press) has noted. Diekhoff's subjects apparently received up to several hundred discrimination trials, over 3 days, and thus had considerable opportunity to learn to discriminate between afferentation associated with response and nonresponse states on the basis of the "feedback" provided by the discrimination probe. In fact, the discrimination performance of Diekhoff's subjects improved over days, in keeping with this interpretation. Thus, whether or not subjects can discriminate electrodermal responses in the absence of biofeedback training remains uncertain. Whether or not subjects can discriminate electrodermal responses following feedback training to control electrodermal activity also remains unclear. In Diekhoff's experiment, some subjects were provided with training to discriminate, not to

3 Discrimination of E ectrodermal Activity 395 control, electrodermal responses, and this training did not enhance their performance on electrodermal discriminations. In Baron's study, in which subjects received training to control electrodermal activity, only 4 of the 10 subjects discriminated electrodermal responses significantly above chance level. Finally, in Stern's experiment the discrimination performance of subjects provided with feedback for electrodermal changes differed only slightly from that of subjects not provided with feedback, and the difference between the two groups was not examined statistically. The poor discrimination performance of subjects provided with feedback training in these studies may be due to the fact that these subjects received only a minimal amount of feedback training to control electrodermal behavior. More extensive training with exteroceptive feedback was provided in the present investigation. METHOD Subjects Nineteen female and five male students made up three groups of eight subjects. They were paid $5 for their participation. Eiectrophysiological Recordings Electrodermal activity was measured as skin potential, which was recorded on a Sanborn oscillographic recorder (model 7701A), operating at a speed of 2.5 mm/second. Skin potential was recorded through Beckman biopotential Ag/AgC1 skin electrodes, 7 mm in diameter, which were attached to the skin by means of Beckman electrode collars, 10 mm in diameter. The face of each electrode was filled with Unibase (.07 molar with respect to NaCl) when in use. When not in use, electrodes were kept shorted together in a weak salt solution. In recording skin potential, the active electrode was placed on the distal phalange of the middle finger of the left hand, while the reference electrode was placed over the left forearm, approximately over the fascia of the palmaris longus muscle. The reference site was abraded lightly with a pumice stone at the beginning of each session. Procedure All subjects participated in two separate sessions, which were almost always on consecutive days. Subjects were provided with a standard set of

4 396 Lacroix instructions at the beginning of each session. These contained information about skin potential (particularly that it evidenced "sweating" and "reabsorption" waves) and its recording. These instructions also informed the subjects that the purposes of the study were (a) to demonstrate that provision of exteroceptive feedback was (or was not, in the case of control subjects) essential to the development of control of skin potential, and (b) to study the relationship between the discrimination and control of changes in skin potential. The provision of the instructions was followed by three series of 12 discrimination trials (series A, B, and C), which examined the subjects' ability to discriminate skin potential responses. These were followed by 26 training trials to control skin potential activity. On half the training trials (Raise trials), the subject was instructed to produce increases in palmar sweating; on the other half (Lower trials), his task was to inhibit palmar sweating. Increases in sweating were identified (accurately) as increases in the negativity of skin potential for subjects in Group N, and (inaccurately) as increases in the positivity of skin potential for subjects in group P. The second session began with 26 more training trials, which were followed by four series of 12 discrimination trials (A,B,C, and A). At the end of the second session, subjects were also required to complete a questionnaire. Series A discrimination trials tested the subject's ability to discriminate between the presence and absence of negative skin potential responses. On half the trials, a buzzer was presented on the rising phase of a negative skin potential response. On the other half of the trials, the buzzer was presented when no skin potential response of any kind had been evident in the electrodermal record for at least 5 seconds. The subject's task was to indicate verbally whether he believed that he was or was not producing a negative skin potential response on each buzzer presentation. Series B and C discrimination trials were designed to be carried out similarly but to test, respectively, the subject's ability to discriminate between the presence and absence of positive skin potential responses, and between negative and positive skin potential responses. Series A trials were carried out as intended in all subjects. However, the frequency of positive responses was too low for series B and C trials to be carried out as intended in the majority of subjects, including most of those in group P. Consequently, only the discrimination results from series A trials will be reported here. Knowledge of results was not provided on any of the discrimination trials. Training trials were 30 seconds in duration and were presented according to a predetermined random sequence, which was the same for all subjects. Intertrial intervals ranged from 30 to 50 seconds, with a mean of 35 seconds, Subjects in the two experimental groups (N and P) received feedback on their performance on the training trials. This was provided by

5 Discrimination of Electrodermal Activity 397 allowing subjects to watch their skin potential record "on-line." Feedback was presented during 20 of the 26 training trials in each session. It was not presented during 6 "Transfer" trials, designed to examine the extent to which electrodermal control was evident when feedback was withdrawn. Subjects in the control (C) group received no feedback whatsoever on their electrodermal performance. The questionnaire given at the end of the second session was addressed at three issues. These were (a) the extent to which the subjects complied with the discrimination and control instructions, (b) the extent to which they felt confident in their ability to discriminate and control skin potential changes, and (c) the extent to which they actively attempted to produce changes in skin potential during the discrimination trials. Data Analysis. Performance on the discrimination tasks was measured according to the method of signal detection theory, in terms of d'. Performance on training trials was evaluated as follows. First, skin potential was sampled at the midpoint of consecutive 5-second intervals, beginning 30 seconds before trial onset. Evaluations of control of skin potential were based on change scores based on the difference between the last three trial measurements and the last three baseline measurements. This measure takes baseline behavior into account and is unlikely to be contaminated by any orienting responses to trial onset. It was intended to provide a general index of the extent to which subjects were able to modify their skin potential, which was sensitive to both tonic and phasic changes, although it did not distinguish between the two. Further analyses then examined differences between trial and baseline periods in skin potential response frequency and amplitude. Electrodermal control was examined on the six trials in each session (trials #3, 4, 12, 15, 23, and 24) on which exteroceptive feedback was provided in groups N and P, which most closely matched the six Transfer trials (trials #1, 2, 13, 14, 25, and 26). These are referred to as Feedback trials. The extent to which electrodermal control was maintained when feedback was withdrawn was ascertained by examining performance on the six Transfer trials in each session. RESULTS Dixcrim&ation Figure 1 portrays the mean discrimination performance (d') of groups N, P, and C, on series A discrimination trials, before and after training to control skin potential. Performance on the two series A trials following

6 398 Lacroix 1"4 = STANDARD ERROA 1"2 I'0 "8 'ol "S "4 "2 O / -'2 -'4 -.$ -'G BEFOR~ TRAINING AFTER TRAINING Fig. 1. Mean d' of groups N, P, and C, on series A discrimination trials, before and after training to control skin potential. training did not differ significantly in any of the groups, and has been combined. Examination of performance on the discrimination trials prior to training revealed that none of the groups discriminated reliably between the presence and absence of negative skin potential responses on these trials; nor was there reliable evidence of discrimination prior to feedback training when the three groups were considered together. However, discriminations between the presence and absence of negative skin potential responses were performed successfully and reliably by both the experimental groups following biofeedback training [for group N, t (7) = 5.61, p<.001; for group P, t (7) = 3.08, p<.02]. Group differences in discrimination performance were examined by means of an analysis of variance (ANOVA) that took Groups and Training (before and after training) as variates. This analysis yielded statistically significant main effects of Groups [F(2,21) = 8.06, p<.01] and Training [F(1,21) = 12.26, p<.01], as well as a significant Groups X Training interaction [F(2,21) = 3.68, p<.05]. Post hoc tests indicated that the discrimination performance of groups N and P following training did not differ significantly, but that the posttraining performance of both these groups differed significantly from their pretraining performance, and from both the pre- and posttraining performance of group C (Newman-Keuls, all p<.05).

7 Discrimination of Eiectrodermal Activity 399 These differences between groups with respect to discrimination performance were not accompanied by significant group differences in the extent to which subjects attempted to comply with the discrimination instructions, as measured by their responses to the questionnaire. Further, Group differences, in the degree to which subjects felt confident that they could discriminate their skin potential responses, were not statistically significant. Product-moment correlations computed between posttraining d' scores and the degree to which subjects felt confident of their ability to discriminate skin potential responses were low and unreliable, both when the groups were considered separately and when they were considered together. These data suggest that training with exteroceptive feedback led to an ability to discriminate electrodermal responses, but that the extent of that ability remained unrecognized by the subjects. Further evidence suggested that the discrimination performance of subjects in groups N and P following training did not stem from overt manipulations of skin potential on the discrimination trials. Although subjects estimated that they actively attempted to control their skin potential activity on slightly more than half the posttraining discrimination trials, the three groups did not differ at a statistically significant level with respect to those estimates. Moreover, product-moment correlations computed between these estimates and d' scores did not approach statistical significance, in any of the groups. Finally, group differences in skin potential activity on the final four series of discrimination trials were examined by means of three separate ANOVAs employing Groups and Series as variates, which were carried out on the total frequency of skin potential responses, on the frequency of negative skin potential responses, and on the amplitude of negative responses. None of the main effects or interactions proved statistically significant, indicating that the three groups did not differ systematically with respect to skin potential activity on the different series of discrimination trials. Training Figure 2 presents mean skin potential activity in the three groups on the last two Raise (upper half) and the last two Lower (lower half) Feedback trials, where control was expected to be maximal. Inspection of the upper half of the figure shows that the three groups generated different patterns of skin potential activity on Raise trials. This was confirmed by a comparison of the groups' performance by means of a one-way ANOVA applied to the change scores on Raise trials. This analysis yielded a statistically significant effect of Groups [F(2,21) = 3.51, p<.05]. On the other hand, inspection of the lower half of Figure 2 indicates that the three groups generated sim-

8 400 Lacroix BASELINE RAISE TRIAL -2 o..-, *2 en OP Z~C,.J BASELINE LOWER TRIAL o ~- -2 w u o +1 FIVE- SECOND INTERVALS Fig. 2. Mean skin potential activity in groups N, P, and C, on the last two Raise (upper graph) and the last two Lower (lower half) Feedback trials of the second session. Skin potential was measured at the midpoint of consecutive five-second intervals, beginning 30 seconds before trial onset. ilar patterns of skin potential activity on Lower trials, and in all cases performance on Lower trials appeared to reflect a continuation of a decreasing trend in baseline skin potential. A one-way ANOVA applied to the change scores obtained on Lower trials did not yield a reliable effect of Groups. Group differences in the degree of electrodermal control on Raise trials are unlikely to stem from differences in motivation, since the groups did not differ significantly in the extent to which they attempted to comply with the instructions to control skin potential, as measured by their responses to the questionnaire. Further analyses examined the performance of each group in greater detail. Group N. Examination of Figure 2 reveals that the largest difference in skin potential activity between Raise and Lower trials was obtained in group N. Statistical analyses, based on change scores on the last two Raise and Lower Feedback trial pairs, indicated that the difference between performance on Raise and Lower trials was significant [t (7) = 3.28, p <.02], reflecting both an increase in skin potential on Raise trials [t(7) = 2.46, p<.05] and a decrease on Lower trials [t (7) = 3.85, p<.01]. A similar

9 Discrimination of Electrodermal Activity 401 pattern of results was obtained on Transfer trials, where exteroceptive feedback was withdrawn. However, the difference between performance on Raise and Lower trials was not significant on Transfer trials. Direct comparison of the difference in performance between Feedback and Transfer trials in group N was carried out by means of an ANOVA employing Types of Trial (Raise, Lower) and Feedback (present, absent) as variates. The resulting Types of Trial X Feedback interaction approached statistical significance IF(l,7) = 3.87, p<.10]. Thus, while subjects in group N were able to control skin potential on trials where exteroceptive feedback was provided, electrodermal control did not fully generalize to trials where the feedback stimulus was removed. Further analyses examined whether the electrodermal control that was evident on Feedback trials reflected primarily a change in the frequency or the amplitude of skin potential responses. Both the frequency and the amplitude of negative skin potential responses were examined on the last two Raise-Lower Feedback trial pairs. The results of this analysis are displayed in Table I, which presents the mean frequency and amplitude of negative skin potential responses during the 30-second baseline, and during Raise and Lower trial periods. These data suggest that control of skin potential was accomplished by increasing both the frequency and the amplitude of negative skin potential responses on Raise trials, and by decreasing the amplitude of those responses on Lowe r trials. However, this was only partially confirmed by statistical analyses. ANOVAs employing Types of Trial and Trials (baseline, trial period) as variates were applied to both the frequency and the amplitude data. Of the resulting Types of Trial X Trials interactions, only that with respect to response amplitude proved statistically significant [F(1,7) = 9.99, p<.05], reflecting significantly larger response amplitudes on Raise trials than on Lower trials (Newman-Keuls, p <.05). Thus, control of skin potential in group N appeared to involve primarily manipulations of response amplitude. Finally, acquisition effects were examined. These were evaluated in two ways. First, mean performance on the first session was compared with mean performance on the second session. And second, performance on the first Raise-Lower trial pair of the first session was compared with performance on the last Raise-Lower trial pair of the second session. Change scores, as well as the frequency and the amplitude of negative skin potential responses, were examined. These data were evaluated by means of ANOVAs employing Types of Trial and Time as variates, carried out separately on the data from Feedback and Transfer trials. None of the analyses yielded a statistically significant Types of Trial X Time interaction. Thus, the acquisition of skin potential control in Group N did not appear to follow a uniform course.

10 402 Lacroix Table I. Mean Frequency and Amplitude of Negative Responses in Group N on the Last Two Raise-Lower Feedback Trial Pairs Raise Baseline Trial Lower Baseline Trial Frequency /30 sec..~ SE Ampfitude (my) X SE Group P. Examination of the performance of group P in Figure 2 indicates that this group exhibited no substantial change in skin potential on Raise trials, and a slight decrease in skin potential on Lower trials. A comparison of the change scores on the last two Raise-Lower Feedback trial pairs did not approach statistical significance. Similar results were obtained when other possible indices of control were examined, such as the frequency of positive and bidirectional responses, the frequency of negative responses, and the amplitude of negative responses. These data were evaluated by means of ANOVAs employing Types of Trial and Trials (baseline, trial period) as variates. None of the resulting Types of Trial X Trials interactions approached statistical significance. Evaluation of skin potential control in group P was also carried out by examining whether suppression of negative skin potential responses might have occurred on Raise trials as a function of training. Both frequency and amplitude of negative responses were compared on the first and last Raise-Lower trial pairs, by means of ANOVAs employing Types of Trial and Time as variates. Neither Types of Trial X Time interaction approached statistical significance. In sum, the above analyses, which examined performance on Feedback trials, yielded no evidence of electrodermal control in group P. Parallel analyses, carried out on performance on Transfer trials, yielded very similar results and did not provide any evidence of electrodermal contrel in this group. Group C. Examination of the performance of group C in Figure 2 yields no evidence of skin potential control, as this group exhibited decreases in skin potential of similar proportions on both types of trials. Performance was examined on trials corresponding to the last two Raise and Lower Feedback trials in groups N and P. Evaluations of change scores on these trials revealed that the difference in performance on Raise and Lower trials was unreliable. Likewise, examination of other possible indices of control yielded no evidence of electrodermal control. The frequency and amplitude of negative skin potential responses, as well as the frequency of positive and bidirectional responses were examined as for group P, by

11 Discrimination of Electrodermal Activity 403 means of ANOVAs employing Types of Trial and Trials as variates. None of the resulting Types of Trial X Trials interactions approached statistical significance. DISCUSSION The present investigation was primarily concerned with whether or not subjects could discriminate changes in skin potential activity in the absence of feedback training to control skin potential, and when provided with such training. Subjects were unable to discriminate skin potential responses in the absence of training with exteroceptive feedback, under the conditions of this experiment. These findings are in keeping with previous reports by Baron (1966) and Stern (1972). However, the present results also indicate that training to control skin potential by providing exteroceptive feedback for electrodermal activity leads to successful discrimination of skin potential responses. Subjects in group N, who were instructed to produce increases and decreases in palmar sweating on Raise and Lower trials, and who received accurate feedback on their performance, learned to control their skin potential activity appropriately on Feedback trials, and this control appeared to be accomplished primarily by manipulating the amplitude of negative-going responses. These subjects also learned to discriminate skin potential responses successfully as a consequence of this training. Further, training with exteroceptive feedback appears necessary to develop the ability to discriminate electrodermal responses. This is suggested by the performance of subjects in group C. These subjects were given the same rmmber of discrimination trials as those in group N, received identical instructions on training trials, and the same number of training trials, but were not provided with exteroceptive feedback on these trials, In keeping with the previous report by Lacroix and Roberts (1976), subjects in group C did not learn to control their electrodermal activity appropriately. Moreover, these subjects were unable to discriminate electrodermal responses following training. These results, as well as related findings with respect to the discrimination of changes in cardiac activity (Brener & Jones, 1974), support the view that the process of acquiring control of a given autonomic response involves learning to discriminate Changes in interoceptive afferentation associated with the target autonomic response. However, this conclusion must be viewed with caution. In the case of discriminations of interoceptive events, presentation of the stimuli to be discriminated is under the control of the subject. This differs from more typical discrimination situations, where the subject attempts to discriminate the occurrence of exteroceptive

12 404 Lacroix stimuli, which are under the control of the experimenter. Thus, in the former case, it is possible that the subject's performance on the discrimination tests is based not on his ability to discriminate interoceptive events but rather on his ability to produce these events. This possibility was evaluated here by comparing electrodermal behavior on the different discrimination series. While the results of these comparisons did not support the view that the subjects' performance on the discrimination tests was based on their production of the skin potential changes to be discriminated, it is possible that these evaluations were not sensitive enough to detect possible "voluntary" manipulations of electrodermal activity. Nonetheless, the present results do suggest that learning to control an internal response involves learning to discriminate afferentation related to changes in that response. However, the nature of the afferentation that subjects learn to discriminate as a consequence of biofeedback training is as yet undetermined. Brener (1974a) has suggested, in the case of cardiac discriminations, that these discriminations may be based on afferentation originating from the visceral effector. This is unlikely in the case of electrodermal discriminations, as there is no evidence of any afferentation arising directly from sudomotor activation, and the innervation of palmar sweat glands appears to be purely efferent (Kuno, 1956). Thus, successful discriminations of electrodermal changes are likely based on afferentation from one or more other response systems (e.g., respiratory) which feedback training has identified to be correlated with (and possibly "mediating") changes in electrodermal responding. The specific nature of the afferentation that subjects learn to discriminate as a result of biofeedback training, and the relationship between that afferentation and the target autonomic response, are questions that will have to be answered satisfactorily before a genuine understanding of the process of acquiring control of an internal response can emerge. Another issue that will have to be addressed concerns the relationship between control and discrimination processes. While learning to control electrodermal behavior was accompanied by a learning to discriminate electrodermal responses in the present experiment, there are indications that the relationship between control and discrimination processes may not always be so straightforward. For example, a recent experiment by Black et al. (in press) showed that subjects who were able to control occipital alpha EEG nonetheless performed unsuccessfully on tests of their ability to discriminate alpha from non-alpha. This may mean that the ability to discriminate a response is not always a prerequisite for control of that response. However, it is also possible that the discrimination tests employed in the experiment of Black et al. did not provide a valid measure of the subjects' ability to discriminate their EEG.

13 Discrimination of Eleetrodermal Activity 405 Conversely, it appears that under some circumstances subjects may be able to discriminate a response but nonetheless perform unsuccessfully on tests of their ability to control that response. Group P in the present experiment provides an example of such a case. Like group N, group P underwent training to control skin potential and was provided with exteroceptive feedback. However, subjects in group P were misinformed about the relationship between their responses and changes in the feedback display, as negative-going responses (which reflect increases in palmar sweating) were identified to the subjects as decreases in palmar sweating, and vice versa. Subjects in group P did not learn to produce the desired changes in skin potential on training trials. However, they clearly learned something about skin potential activity as a result of biofeedback training, since they were able to discriminate negative skin potential responses following this training. It may be that these subjects would have been able to produce negative changes in skin potential had they been asked to generate such changes. These subjects' inability to control skin potential on training trials may not have reflected a real inability to control skin potential but rather may have indicated that they were not given a proper opportunity to demonstrate their ability to control this activity. Thus, even if control and discrimination processes are related in a direct and simple way, the ability to perform one type of task successfully may not always lead to successful performance on the other, because performance on discrimination or control tests may be somehow impeded, a point which has also been made by Roberts (in press). Clearly the relationship between the discrimination and control of internal events is not well understood at present. However, in view of evidence, such as that presented here, that the two processes appear related, and in view of the heuristic benefits accruing to discrimination accounts of the process by which subjects learn to control internal responses (Brener, 1974a; Roberts, in press), studies of the relationship between discrimination and control processes would seem well worth pursuing, REFERENCES Baron, J. An EEG correlate of autonomic discrimination. Psychonomic Science, 1966, 4, Black, A. H., Cott, A., & Pavloski, R. The operant learning theory approach to biofeedback training. In G. E. Schwartz & J. Beatty (Eds.), Biofeedback: Theory and research. New York: Academic Press, in press. Brener, J. A general model of voluntary control applied to the phenomena of learned cardiovascular change. In P. A. Obrist, A. H. Black, J. Brener, & L. V. DiCara (Eds.), Cardiovascular psychophysiology: Current issues in response mechanisms, biofeedback, and methodology. Chicago: Aldine, (a)

14 406 Lacroix Brener, J. Factors influencing the specificity of learned cardiovascular control. In L. V. DiCara (Ed.), Limbic and autonomic nervous systems research. New York: Plenum, (b) Brener, J. Sensory and perceptual determinants of voluntary visceral control. In G. E. Schwartz & J. Beatty (Eds.), Biofeedback: Theory and research. New York: Academic Press, in press. Brener, J. & Jones, J. M. Interoceptive discrimination in intact humans: Detection of cardiac activity. Physiology and Behavior, 1974, 13, Diekhoff, G. M. Effects of feedback in a forced-choice GSR detection task. Psychophysiology, 1976, 13, Kuno, Y. Human perspiration. Springfield, Illinois: Thomas, Lacroix, J. M., & Roberts, L. E. Determinants of learned electrodermal and cardiac control: A comparative study. Psychophysiology, 1976, 13, 175. (Abstract) Roberts, L. E. The role of exteroceptive feedback in learned electrodermal and cardiac control: Some attractions of and problems with discrimination theory. In J. Beatty & H. Legewie (Eds.), Biofeedback and behavior. New York: Plenum, in press. Stern, R. M. Detection of one's own spontaneous GSR's. Psychonomic Science, 1972, 29, (Revision received May 2, 1977)