Biological Psychology, 3,197.5,271-293.

STIMULUS

OMISSION

ELECTRODERMAL COMPONENTS

DAVID

@ North-Holland Publishing Company

AND RECOVERY

AND DIGITAL

VASOCONSTRICTIVE

OF THE ORIENTING

A. T. SIDDLE*

and

PETER

OF THE

RESPONSE

A. HERON

Department of Psychology, University of Southampton, Southampton,

Accepted

for publication

U.K.

6 June 1975

Three experiments are reported which investigated Sokolov’s (1968) hypothesis that, after a number of stimulus presentations, complete omission of a stimulus leads to increased responsiveness of the orienting response (OR). The skin conductance response @CR) and finger pulse volume (FPV) response components of the OR were studied. In experiment 1 (N = 60), the effect of number of pre-omission training trials on response recovery was investigated, while experiment 2 (N= 120) investigated the effects of stimulus intensity (70 or 90 dB) and interstimulus interval (12 or 21 set) on recovery to stimulus omission following a fixed number of training trials. In experiment 3 (N = 40), an attempt was made to control for possible below-zero habituation effects by training each subject to a habituation criterion before stimulus omission. All experiments employed a 1000 Hz tone of 3 set duration which was presented at a constant interstimulus interval. Although recovery of the SCR did occur under some conditions, the results were largely negative. Manipulation of the number of training trials, training stimulus intensity and interstimulus interval had little effect on response recovery. A consistent finding, however, was that subjects who displayed SCR recovery also displayed significantly more spontaneous fluctuations in skin conductance during the pre-stimulus period and required significantly more training trials to reach the criterion of habituation than did subjects displaying no recovery. Moreover, the SCRs displayed by ‘labile’ subjects on omission trials were significantly larger than those displayed on either the last training trial or during a control interval just prior to stimulus omission.

1. Introduction Considerable research effort has been devoted to an examination of the orienting response (OR) which is defined (Sokolov, 1963) in terms of the complex of sensory, muscular, EEG and vegetative changes which occur in response to a discriminable change in sensory input. The most comprehensive model of OR elicitation and habituation is that proposed by Sokolov (1960, 1963, 1968) who states that the OR possesses two important characteristics : it * Address for correspondence: Southampton, Southampton

D. A. T. Siddle, SO9 5NH, U.K. 277

Department

of Psychology,

University

of

278

D. A. T. Siclrll~ onrl P. A. Hwm

displays an amplitude decrement as a function of stimulus repetition (habituation), but is re-evoked by any detectable change in one or more parameters of the stimulus. With regard to the second characteristic, Sokolov (1968) has reported that the OR is re-evoked ‘when the signal is intensified, weakened, lengthened, shortened; when it is presented before the usual time; when it has been omitted at the usual time’ (p. 674). OR re-evocation is also said to occur if a new stimulus is added or if one of the elements in a compound stimulus is omitted. Moreover, Sokolov (1963) has suggested that the size of the OR elicited by a change in stimulation is a positive function of the amount of change. Thus, OR evocation is said to be a function of stimulus change per SC and independent of the energy properties of the change stimulus. To account for this feature of the OR, Sokolov (1960) has proposed that two stages are involved in its elicitation. In the first, stimulus input is compared with the stored characteristics of previous stimulation, while in the second, the OR is either elicited or inhibited, depending upon the result of the comparison process. Storage is said to occur in a ‘neuronal model’. A match between incoming stimulation and the neuronal model leads to OR suppression, while a mismatch results in OR evocation. The clear implications of this model for information processing (Sokolov, 1966) and classical conditioning methodology (e.g. Badia and Defran, 1970) have stimulated research aimed at providing more empirical evidence to support Sokolov’s proposals. The general experimental paradigm has involved presentation of a standard stimulus, either for a fixed number of presentations (training trials), or until subjects no longer respond, followed by introduction of changes in one or more parameters of the stimulus. The ORs elicited by the various change stimuli have then been compared with each other and with response level immediately prior to the introduction of change. Less frequently, they have been compared with response level under a control condition not involving stimulus change. The change conditions employed include intensity changes in visual and auditory stimuli (e.g. Bernstein 1969; O’Gorman, Mangan and Cowan, 1970) frequency changes in auditory stimuli (e.g. Geer, 1969) changes in the order of stimulus presentation (Furedy, 1968a) and omission of one or more elements from a compound stimulus (Gliner, Harley and Badia, 1971). Some studies have also used a change in stimulus modality (e.g. Houck and Mefferd, 1969) or some other qualitative change (e.g. Yaremko and Keleman, 1972). Although the results of these studies have been reviewed in terms favourable to Sokolov’s model (Graham, I973), O’Gorman (I 973) has concluded that the data on OR recovery do not require a model which includes a matching process. O’Gorman has argued that the only change conditions which reliably re-evoke an OR are increases in stimulus intensity and change in stimulus modality, and that these data can be interpreted more parsimoniously by using an arousal construct. However, both O’Gorman (1973) and Groves and Thompson (I 970) have suggested that a demonstration

Stimulrrs omission and SCR and FP V response recowry

279

of OR recovery’ to omission of a stimulus would constitute strong evidence in support of a matching process. Although the effect of stimulus omission on evoked potentials has been studied (Sutton, Tueting, Zubin and John, 1967), evidence on autonomic response recovery is meagre. Allen, Hill and Wickens (1963) reported that omission of either the second and third or first and third elements from a threeelement compound stimulus resulted in an increase in skin conductance response (SCR) amplitude. Similarly, Badia and Defran (1970) reported that omission of one element from a two-element stimulus produced heightened SCR amplitude. However, Glitter et al. (1971) have argued that these results might have been due to the instructions employed, in that subjects were required to judge the intensity of the second stimulus element. In their own study, Gliner et al. reported that SCR recovery to the omission of one element from a two-element stimulus was inconsistent, despite the fact that all subjects reported being aware of the change. Complete stimulus omission was employed by Fried, Welch, Friedman and Gluck (1967) who concluded that electrodermal activity was not sensitive to omission of a stimulus after 5, 10 or 15 training trials. However, their analysis was based on tonic resistance levels rather than on a more conventionally employed phasic response measure, and Furedy (1968b) has demonstrated the relative insensitivity of tonic measures to stimulus change. Sokolov (1968) reported that complete omission of a stimulus which had been presented regularly resulted in recovery of the EEG blocking response, but presented data from only one subject. This paper presents the results of three studies designed to investigate recovery of the SCR and finger pulse volume (FPV) response components of the OR to complete stimuhrs omission. The basic paradigm involved presenting pure tones at a constant interstimulus interval (ISI), which, as Graham (1973) has noted, is also the paradigm employed in studies of temporal conditioning. However, while temporal conditioning studies have been concerned with anticipatory responses occurring during the interval immediately preceding the time of the stimulus (Lockhart, 1966; Harley, 1973), the present approach involved responses occurring immediately U@P the time at which a stimulus should have been presented. Control groups which did not receive stimulus omission were also included so that OR recovery could be assessed by both between (control versus experimental groups) and within (omission trial versus last training trial) group comparisons. 2. Experiment 1 This experiment was designed to investigate the effect of number of training trials on recovery of the SCR and FPV response to stimulus omission. Since ‘The term ‘recovery’ is used to denote either an increase in response amplitude in comparison with a control condition, or response evocation following zero responding (habituation).

280

D. A. T.

SiMc~ and P. A. Hrron

Sokolov (1960, 1963) has argued that the characteristics of stimulation are stored, at least temporarily, in a neuronal model, some authors (e.g. Geer, 1969; James, Daniels and Hanson, 1974) have suggested that the mismatch between a training and change stimulus will be greater following a relatively large number of training trials than following a relatively few. Since Sokolov (1968) has stated that with a constant ISI, the neuronal model contains information concerning the distribution of stimuli in time, it seems reasonable to suppose that this information will increase in accuracy as a positive function of the number of training trials. It was hypothesized, therefore, that as the number of training trials is increased, larger SCRs and FPV responses will be evoked by stimulus omission. 2. I. Method 2.1.1. Subjects and design. The subjects were 60 undergraduate volunteers (age range 17-25 yr) who were allocated randomly to one of six groups. The records of six subjects who failed to display any evoked SCRs during training were discarded and six more subjects recruited from the same population. There were six males and four females in each group. The experimental design is shown in table I. For group I, responses were sought at the time when stimulus number 7 should have occurred, while for group 2, responses were sought at the time when stimulus number IO should have occurred, and so on. The responses of group 6, the control group, to stimuli 7, IO, 13, 16 and 21 were used as a baseline with which the responses of the experimental groups to omitted stimuli could be compared. 2.1.2. Apparatm andprocedure. Electrodermal and digital vasomotor activity were measured in response to an auditory stimulus of 3 set duration, presented at a constant ISI of 21 set (offset to onset). Tone frequency was 1000 Hz, and the intensity 70 dB (re : 0.0002 dyn/cm2) as rated by Dawe’s sound level meter at the headrest of the chair in which subjects sat. The stimuli were produced by an Advance signal generator (model HI B) and delivered through a speaker 100 cm directly behind the subject. Stimulus duration and IS1 were controlled by a Birkbeck timer in conjunction with timing and counting circuits. Bipolar recording of skin conductance was made using domed Ag-AgC1 Table

1.

Design of experiment 1. Group

Number Omitted

of training trials stimulus number

I

2

3

4

5

6

6 7

9 10

12 13

15 16

20 21

21 -

Stimulus omission and SCR and FP V response recovery

281

electrodes filled with 0.05M NaCl electrolyte and attached to masked areas on the index and second fingers of the subject’s left hand. The electrodes were placed over the whorls of the fingerprints on the distal phalanges. A constant voltage of 0.5 V (Lykken and Venables, 1971) was applied across the electrodes and conductance recorded on a Grass model 7 polygraph with a 7Pl B preamplifier sensitivity of 1 mm of pen deflection equal to 0.02 umho. The total area from which conductance was recorded was 0.35 cm’. Digital vasoconstriction was measured using a Grass photoelectric transducer attached to the distal phalange of the index finger of the subject’s right hand. It was recorded via a 7P6 pre-amplifier with a time constant of 0.03 sec. Respiration was monitored via a Grass nasal thermocouple to detect SCRs and FPV responses arising from coughs, deep breaths or other respiratory irregularities. Subjects were seated in a semi-reclining padded chair in a sound-proofed room and the apparatus monitored from an adjoining room. They were informed that after a short rest period, some tones would be heard from time to time and that no responses were required. They were asked merely to relax but not to go to sleep. After a 5 min rest period, the stimulus programme commenced without prior warning. Subjects were tested under eyes-closed conditions with an ambient room illumination of 2 lx. Temperature was within the range 19-23°C and the relative humidity within 49-55 x,. 2.2. Scoring 2.2.1. Electrodermal activity. SCRs greater than 0.02 umho, occurring from 1 to 5 secafter stimulus onset, were considered changes evoked by the stimulus. Since presumably, a certain period of time is required for subjects to become aware that an expected stimulus has not occurred, the latency criterion for SCRs to omitted stimuli was modified. Here, SCRs greater than 0.02 umho, occurring from 1 to 9 set following the time at which the stimulus should have been presented, were considered changes evoked by stimulus omission. SCR amplitude was measured as a change in log umhos. Tonic skin conductance level (SCL) and spontaneous activity were assessed during the rest period. SCL (log umho) was measured at every minute, while spontaneous activity was measured by counting the number of artifact-free fluctuations greater than 0.02 umho. 2.2.2. Digital vasoconstriction. The records were examined for constriction occurring within the interval beginning 1 set after stimulus onset and ending after 10 beats. For a decrease in pulse amplitude occurring during this interval to be considered a stimulus-evoked response, it had to be maintained for three consecutive beats, with the added constraint that the largest of these three beats was smaller than the smallest of the three pre-stimulus beats. For FPV responses to omitted stimuli, the response interval began 1 set after the time at which the stimulus should have been presented and ended after 15 beats. The

282

D. A. T. Sicldk~ am/P. A. Hwon

ratio of the mean of the three pre-stimulus beats to the mean of the three poststimulus beats displaying constriction was computed, and the log of this ratio established. Thus, FPV response ampIitude was measured as a change in log mm (Morgenson and Martin, 1968). 3. Results and discussion 3.1. Electrodermal actiaity Although there was no difference between the groups in terms of spontaneous activity during the rest period (F< I), there was a significant difference in terms of XL (F= 3.02, df= 5.54; p < 0.05). individual comparisons (Newman-Keuls test) indicated that group 3 displayed higher SCL than did groups 5 and 6 (p < 0.05 in both cases). Although the original intention was to analyse the data in terms of both the frequency and amplitude of SCRs to omitted stimuh, inspection of the data indicated that the frequency of SCRs to omitted stimuti was too low to permit parametric analysis of response amplitude. The number of SC% occurring in response to omitted stimuli, i.e. to stimuli 7, 10, 13, 16 and 21 in the appropriate groups, and the number of SCRs evoked by actual stimuli 7, 10,13, 16 and 21 in the control group are shown in table 2. These data were analysed using a procedure similar to that outlined by Still (1967). The number of SCRs displayed by each group was weighted by fitting the coefficients of the linear polynomial, adjusted for unequal intervals (Grandage, 1958). The weighted scores of the experimental and control groups were then compared using a Friedman non-parametric analysis of variance for correlated samples. This indicated that the apparent interaction shown in tabIe 2 is not significant (xl = 1.8, df = I), despite the separation of exper~menta1 and control groups

Table 2. Number of SCRs and FPV responses which occurred to omitted and actual stimuli and during control intervals. Stimulus number Variable SCRs

FVP responses

7

10

no. of SCRs on last training trial prior to stimulus omission no. of SCRs to stimuius omission no. of non-specific SCRs during control interval no. of SCRs on test trials in control group

54233 3 2 22344 54221

no. of responses on last training trial prior to stimulus omission no. of responses to stimulus omission no. of responses on test trials in control group

5 6 64 56332

13

16

21

4

4

5

5 5

4 6

2 6

Stimulus omission and SCR and FPV response recovery

283

at 20 training trials (p < 0.07, using a Fisher exact probability test). In order to examine further the relationship between SCR recovery and number of training trials, the number of SCRs which occurred to omitted stimuli in groups 1 and 2 combined was compared with the number which occurred in groups 4 and 5. No significant difference was found (x2 = 0.98, df = 1). Table 2 also shows the number of SCRs which occurred in the experimental groups to the last training trial prior to stimulus omission. Within-group analyses could be performed, therefore, by comparing (McNemar tests) the number of SCRs evoked on the last training trial with the number which occurred to stimulus omission. Significant differences were not found for any of the experimental groups. It could be argued that both the between- and within-group analyses outlined above were contaminated by stimulus energy factors, i.e. responsiveness during a period of no stimulus energy was compared with responsiveness during a period of stimulus energy. In order to avoid this contamination, a third type of analysis was performed. Here, the number of SCRs evoked by stimulus omission was compared with the number of spontaneous or nonspecific SCRs which occurred during a control interval between the last training trial and the omission trial. The control interval of 8 set began 16 set before, and ended 8 set before, the time at which the omitted stimulus should have been presented. The number of these intervals containing an SCR greater than 0.02 pmho is shown in table 2. However, no significant differences were obtained. It is clear from table 2 that many subjects (64%) did not display SCR recovery to stimulus omission, and an attempt was made to identify electrodermal differences between these subjects and those who did display response recovery. The data from groups I, 2,3,4 and 5 were combined, and comparisons made between subjects displaying recovery (Group R, N = 18) and those displaying no recovery (group NR, N = 32) to an omitted stimulus. Comparisons were made in terms of tonic SCL and spontaneous activity during the rest period and SCR amplitude to the first stimulus in the training serie?. Although group R (8= 25.44, SD = 13.71) displayed significantly more spontaneous fluctuations thandidgroupNR (8= 9.13, SD = 9.31 ; fJ = ?7,p < 0.01) there were no significant differences in either SCL (t = I .86, df = 48) or first SCR amplitude (t = 0.34, df = 48). The relationship between spontaneous activity and response recovery was examined further by designating the 25 subjects displaying the most spontaneous activity as ‘labiles’ and the 25 displaying the least spontaneous activity as ‘stabiles’ (Lacey and Lacey, 1958). Fifteen labiles, but only three stabiles displayed SCR recovery (x2 = 11.94, df = I ; p < 0.001). As a check on the extent to which labile subjects displayed SCR recovery, response amplitude on the omission trial (8= 0.0132 log umho, SD = 0.0132) 2A comparison in terms ofnumber of stimulus presentations required to produce habituation could not be made since the majority of subjects recived too few stimuli for this to be assessed.

284

D. A. T. Siddle and P. A. Heron

was compared (Wilcoxon matched-pairs test) with that on the last training trial (x = 0.0052 log umho, SD = 0.0048) and that which occurred during the control interval prior to the omission trial (8= 0.0039 log umho, SD = 0.0056). If multiple responding occurred during either the omission or control intervals, the largest response was included in the analysis. SCR amplitude on the omission trial was significantly greater than on either the last training trial (T= 32, p < 0.01) or during the control interval (T= 11.5,~ < 0.01). 3.2. Digital vasoconstriction The FPV response data are presented in table 2. A comparison of the frequency of responding in the control and experimental groups was made in the same manner as for electrodermal activity. There was no significant difference in the weighted scores of the control and experimental groups (xz = 0.2, df = 1). Moreover, within-group analyses indicated no significant differences between the frequency of responding on omission trials and on the last training trial. After 20 training trials, however, there was a tendency for a greater number of FPV responses to occur in the experimental group than in the control group (17 < 0.08, using a Fisher exact probability test). An examination was made of subjects’ consistency across response systems in terms of responding to the omitted stimulus. Only 58 % of subjects responded in the same way in both response systems. These results offer little evidence to support Sokolov’s (1968) claim that stimulus omission results in OR recovery. Although the frequency of omissionevoked SCRs and FPV responses rose in experimental groups as a positive function of number of training trials while the number of evoked responses in the control group fell, the difference in trends was not significant. Moreover, within-group analyses indicated that the frequency of responding on omission trials was not significantly greater than that which occurred on either the last training trial or during a control interval prior to stimulus omission. Nevertheless, the difference between control and experimental groups was reasonably clear after 20 training trials. The majority of those subjects designated as ‘labile’ on the basis of frequency of spontaneous fluctuations in skin conductance during the rest period did display electrodermal recovery to stimulus omission, and for these subjects, SCR amplitude on test trials was significantly greater than that on either the last training trial or during the control interval prior to stimulus omission. Two important variables, ISI and stimulus intensity, were not manipulated in the present experiment. On the basis of his failure to observe SCR recovery to a decrease in stimulus intensity, Bernstein (I 969) has argued that an evaluation by subjects of stimulus significance is an important factor in OR recovery, and some evidence has been reported which supports this view (Bernstein, Taylor, Austen, Nathanson and Scarpelli, 1971). It seems reasonable to suppose that greater significance will be attached to a more intense stimulus,

Stimulus omission and SCR and FPV response recovery

285

and with regard to ISI, it seems likely that a short ISI will be encoded more precisely in a neuronal model than will a long one. Accordingly, a second experiment was designed to investigate the effects of stimulus intensity and ISI on SCR and FPV response recovery to stimulus omission. 4. Experiment

2

4. I. Method Subjects and design. The subjects were 120 undergraduate volunteers (age range 17-27 yr). The records of seven subjects who failed to display any evoked SCRs during training were discarded and seven more subjects recruited from the same population. Half the subjects were allocated randomly to experimental groups and the other half to control groups. Experimental subjects received 20 stimulus presentations, with stimulus number 21 being ‘omitted’, while control subjects received 21 stimulus presentations. Within each of these groups, half the subjects received tones of 90 dB, while the other half received tones of 70 dB (re: 0.0002 dyn/cm’ in both cases). For half of the subjects in each intensity group, stimuli were presented at a constant ISI of 21 set (offset to onset), while for the other half, IS1 was 12 sec. Thus, a 2 x 2 x 2 design was employed with the main factors being stimulus omission (control versus experimental), stimulus intensity (70 dB versus 90 dB) and IS1 (21 set versus 12 set). There were seven males and eight females in each group. The control of stimulus presentation, recording techniques, experimental procedure and scoring were the same as in experiment 1. 4.1.1.

Table 3. Number of SCRs which occurred to stimulus 20, during the control interval and to the omitted stimulus 21 in cxpcrimental groups, and to the actual stimulus 21 in contrcl groups. Intensity

Group

ISI (set)

Experiment

Control

Stimulus

90

70

12

stimulus 21 stimulus 20 control interval

9 6 3

7 4 3

21

stimulus 21 stimulus 20 control interval

6 6 2

4 3 1

12 21

stimulus stimulus

5 8

1 3

21 21

286

D. A. T. Sic/die ad

P. A. HIM

5. Results and discussion 5. I. Elcctrodermal activitll There were no differences between the groups in terms of either spontaneous activity (F = I .OO, df = 7, I 12) or SCL (F < 1) during the rest period. The number of SCRs which occurred in response to stimulus omission in the experimental groups and in response to stimulus number 21 in the control groups are shown in table 3. Comparisons of each experimental group with its appropriate control condition (Fisher tests) indicated that only in the case of the 70 dB, 12 set IS.1 group was the frequency of responding higher than in the control condition (p < 0.05). Within-group analyses were performed for experimental groups (McNemar tests) by comparing the number of SCRs which occurred to stimulus omission with the number which occurred on the last training trial (stimulus 20) and during a control interval which began 10 set before, and ended 2 set before, the time at which stimulus 21 should have been presented. These data are also shown in table 3. The comparisons involving the last training trial yielded no significant differences. The omission trial-control interval comparisons indicated that within the 90 dB, I2 set ISI group, significantly more SCRs occurred to stimulus omission (x2 = 4.16, df = I ; p < 0.05). The effects of stimulus intensity and ISI on frequency of responding in the experimental groups were examined using x2 tests. The data were collapsed across ISI, and a x2 performed on the number of subjects responding and not responding to stimulus omission following training with either a 70 or 90 dB stimulus. No significant effect was obtained (x’ = 0.61, df = I). The data were then collapsed across stimulus intensity and a x2 performed on the number of subjects responding and not responding following training with ISIS of either 12 or 21 sec. Again, no significant effect was obtained (x2 = I .70, df = I ). subjects displayed SCR recovery to stimulus Only 43 yz:Iof experimental omission, and comparisons were made between these subjects (group R, N = 26) and those not displaying recovery (group NR, N = 34). As table 4 indicates, group R displayed significantly more spontaneous fluctuations during the rest period and required significantly more stimuli to reach an habituation criterion of three consecutive responses less than 0.02 pmho than did group NR. Subjects were divided into ‘labiles’ and ‘stabiles’ on the basis of their spontaneous activity scores, and it was found that 20 labiles, but only six stabiles displayed SCR recovery (x2 = 1 I .47, df = I ; p < 0.001). A similar analysis performed on the trials to habituation data indicated that 20 of the 30 most slowly habituating subjects displayed SCR recovery, while only six of the 30 rapidly habituating subjects did so (x2 = I I .47, df = I ; p < 0.001). The extent to which labile subjects (defined in terms of spontaneous fluctuation frequency) displayed response recovery was examined further by comparing SCR amplitude on the omission trial (x= 0.0146 log umho, SD = 0.0189) with that occurring on the last training trial (2 = 0.0109 log umho, SD = 0.0250) and

287

Stimrrlrts omission and SCR and FP V response recooery

Comparison

of groups

Mean and SD group R

Variable Mean XL

during

Table 4. R and NR in terms of electrodermal

rest (logpmho)

Amplitude of first evoked SCR (log umho) Spontaneous activity during rest (no. of responses) No. of stimuli to habituation

Mean and SD group NR

0.4415 (0.1497) 0.0838 (0.0592) 28.70 (16.30) 16.35 (4.86)

activity.

Statistic

t = 1.70

0.3725 (0.1600) 0.1024 (0. I 140) 10.71 (10.54) 9.87 (6.18)

P NS

lJ=

519

NS

u=

155

< 0.001

t = 4.43

< 0.001

the control interval prior to the omission trial (x= 0.0023 log umho, SD = 0.0057). In both cases, SCR amplitude on the omission trial was significantly greater (T = 60, p < 0.05 and T = 21, p < 0.01 respectively). during

5.2. Digital vasoconstriction The number of FPV responses to the omitted and actual stimulus 21 in the various groups and the number to stimulus 20 in experimental groups are shown in table 5. It is clear that there are no differences between the experimental and control groups in terms of the number of subjects responding to stimulus 21. Since more subjects responded to stimulus 21 (both omitted and actual) in the vasomotor system than in the electrodermal, and since these subjects were more evenly distributed among the groups, a between-groups analysis of the FPV response amplitude data was made using a 2 x 2 x 2 Table 5. Number of FPV responses which occurred to stimulus 20 and the omitted stimulus 21 in experimental groups, and to the actual stimulus 21 in control groups. Intensity

Group

IS1 (set)

Experiment

12

stimulus stimulus

21

Control

90

70

21 20

7 9

10 6

stimulus stimulus

21 20

9 7

8 6

12

stimulus

21

5

7

21

stimulus

21

8

8

Stimulus

analysis of variance. No significant effects were obtained for control versus experimental groups (F= I .84, df = I,1 l2), stimulus intensity (F< I), or ISI (F < 1). There were no significant interactions. A second analysis was performed for the experimental groups only. The main factors here were stimulus intensity, ISI and trials (last training trial versus omission trial). No significant effects were obtained for stimulus intensity (F< I), ISI (F< I), or trials (F== 2.58, df = 156). There were no significant interactions. Moreover, analyses of response frequency in the experimental groups indicated no significant differences between responsiveness on the omission trial and responsiveness on trial 20. Comparisons were made between experimental subjects displaying FPV response recovery (group R, N = 34) and those displaying no recovery (group NR, N = 26). Thegroups were compared in terms of response amplitude to the first stimlilus in the training series and number of stimuli required to reach a habituation criterion of three consecutive response failures. The N‘s were slightly reduced, however, as seven subjects did not show vasoconstriction in response to the first stimulus (group R, N = 30; group NR, N = 23). There was no significant difference in terms of either response amplitude (Group R x=0.21, SD=O.ll: Group NR x=0.14, SD=O.lO; t== 1.94, df=51) or the number of stimuli required for I~abitu~ition (Group R x= 14.38, SD = 7.08;GroupNR J?= 10.92, SD=7.35; f= 1.85,df=51). An examination was made of subjects’ consistency across response systems in terms of responding to the omitted stimulus. Only 45 y(, of subjects responded in the same way in both response systems. The results of experiment 2 provide no evidence to indicate that recovery of the FPV response occurs follo~ving stin~ulus omission. Neither between- nor within-group comparisons indicated that more subjects responded to stimulus omission than under control conditions. Moreover, analyses of response amplitudes indicated no differences in responsiveness between omission and control conditions. Neither the frequency of responding, nor response amplitude to stimulus omission was influenced by the intensity of the training stimulus or the ISI. Within the electrodermal system, there was some evidence for response recovery to stimulus omission. Between-group analyses of response frequencies indicated that significantly more subjects in the 70 dB, 12 set IS1 group responded to stimulus omission than in the appropriate control condition; within-group analyses indicated that within the 90 dB. 12 set ISI group significantly more subjects responded to stimulus omission than ‘spontaneously’ during a control interval just prior to the omission trial. One factor which may have attenuated the experimental effects is the phenomenon of below-zero habituation. According to Thompson and Spencer (1966) below-zero habituation refers to a continuation of the inferred habituation process beyond the point at which zero responding has occurred.

Stimulus omission and SCR and FPV response recovery

289

Although the phenomenon has received little systematic investigation (Gardner, 1968), it is of potential importance for studies investigating OR recovery to stimulus change. Thompson and Spencer (1966) have suggested that additional habituation training beyond the point of ‘zero’ responding will result in less response recovery, after a recovery period, than will no additional training. This argument can be extended to a situation involving stimulus change rather than a recovery period, and it might be expected that less response recovery will occur when the change stimulus is presented following, rather than before below-zero habituation training. Moreover, since there are large individual differences in how quickly subjects reach ‘zero’ responding in the component systems of the OR, it is clear that with a fixed number of training trials, subjects who require relatively few stimulus presentations to reach ‘zero’ responding will, in fact, undergo below-zero habituation training. In this connection, it is interesting to note that in Corman’s (1967) study where 10 training trials were presented, subjects displaying the most frequent zero responding prior to the introduction of stimulus change also displayed smaller SCRs to the change condition. In view of the above considerations, a third experiment on SCR recovery to stimulus omission was performed in which an attempt was made to control for below-zero habituation training. In this experiment, each subject was trained to his own criterion of habituation and then stimulus omission introduced. 6. Experiment 3 6.1. Method 6.1.1. Suejects and design. The subjects were 40 undergraduate volunteers (age range 17-26 yr). Half the subjects were allocated randomly to experimental groups and the other half to control groups. Within each of these groups, half the subjects received training tones of90 dB, while the other half received tones of 70 dB (re: 0.0002 dyn/cm2 in both cases). A constant IS1 of 12 set was employed for all groups. Each experimental subject was trained to a habituation criterion of three consecutive responses of less than 0.02 pmho and a response was sought at the time at which the next stimulus should have been presented. Control subjects were also trained to a habituation criterion of three consecutive responses less than 0.02 pmho, and then received, 12 set later, a further presentation of the training stimulus. There were six males and four females in each group. The control of stimulus presentation, recording techniques, experimental procedure and scoring were the same as in experiments 1 and 2.

7. Results and discussion There were no significant differences between the four groups in terms of spontaneous activity (F < l), SCL (F = 1.32, df = 3,36), or number of stimuli required to reach the habituation criterion (F < 1).

Within the 90 dB experimental group, only one subject responded on the omission trial, while in the 70 dB experimental group, two subjects responded. The 90 and 70 dB control groups yielded two and four SCRs respectively on the control trial. Clearly, response recovery did not occur to stimulus omission in either of the experimental groups. 8. General discussion These results provide only slight evidence to support the hypothesis that complete stimulus omission leads to recovery of either the SCR or FPV response components of the OR, and are at variance with the data reported by Sokolov (1968). Moreover, the manipulation of number of training trials, stimulus intensity, ISI and the control of possible below-zero habituation effects had little or no effect. Although the original intention was to analyse the data in terms of response amplitude to stimulus omission, this proved to be impossible, as too few subjects displayed measurable responses. Instead, the data were analysed primarily in terms of frequency of responding under various experimental conditions. The only condition under which the FPV response tended to display recovery was in experiment I following 20 training trials; this result was not replicated in experiment 2. With regard to the SCR, recovery occurred in experiment 2 with both 90 and 70 dB training stimuli presented at an IS1 of 12 sec. There was also a tendency for SCR recovery to occur in experiment I after 20 presentations of a 70 dB training stimulus with an ISI of 21 sec. There seem to be several possible reasons for the largely negative results. First, in his discussion of the stimulus omission effect, Sokolov (1965) refers specifically to the EEG component of the OR. The lack of covariation between the component systems of the OR is well documented (Dykman, Reese, Galbrecht and Thomasson, 1959) and it may well be that different OR components are sensitive to different change conditions. It is possible, therefore, that the EEG component is more sensitive to stimulus omission than are autonomic components. Second, the significance of the stimulus may not have been sufficiently high for its omission to result in increased responsiveness. Although an attempt was made to attach significance to the training stimulus by using a relatively high-intensity stimulus (experiment 2), other procedures may have been more effective. For example, it is possible to attach significance or signal value to a stimulus by preliminary instructions (Luria and Homshaya, 1970) and although the question of OR habituation and recovery to signal stimuli has received little systematic investigation, it may be that response recovery occurs to the omission of ;I signal stimulus. This question is under current investigation, The third reason concerns subjects’ awareness of stimulus omission. Sokolov (1968) has stated that OR recovery occurs to a discriminable change in sensory input, and many studies in this area (e.g. Bernstein, 1969) have employed post-experimental interviews in order to establish whether subjects

Stitmrlrrs

omission

and SCR and FPV response wcorcvy

291

were aware of the stimulus change. A similar technique was employed in this study and all subjects reported that they had expected more stimulus presentations. Of course, the interesting question raised is whether OR recovery to stimulus omission is related to the strength of subjects’ expectancy of further stimulation. This question was not investigated here. The fourth possible reason for the negative findings concerns the number of training trials employed. According to Sokolov (1968) the neuronal model encodes not only information concerning stimulus characteristics, but also information such as ISI, and it was argued earlier that the precision with which IS1 information is encoded will be a positive function of the number of training trials. Sokolov provides no experimental details regarding his data on response to stimulus omission, and it may be that the number of training trials employed in this series of experiments was too small. There is a contradiction, however, between this argument and Thompson and Spencer’s (1966) statements concerning below-zero habituation. According to Thompson and Spencer’s argument, the probability of below-zero habituation training should incuense, and hence response recovery to stimulus omission decrease, with increasing training trials. For this reason, an attempt was made in experiment 3 to control for below-zero habituation training. There is little evidence concerning the effects of amount of training on response recovery to stimulus change (Geer, 1969) and clearly, the problem is worthy of further investigation. Some evidence obtained in this laboratory subsequent to the present experiments (Stephenson and Siddle, in press) has indicated that SCR amplitude to a change in the frequency of an auditory stimulus varies as a positive function of the number of below-zero habituation trials. These findings suggest that increased precision of the neuronal mode1 is of greater importance than the attenuating effects of the inferred habituation process, and would lend weight to the suggestion that the number of training trials employed in this series of experiments was too small. Finally, it might be argued that the interval during which responses to omitted stimuli were sought was too restricted. However, an examination of the records from all three experiments indicated that the results are not changed even when the duration of this interval is doubled. Conversely, it could be argued that the response interval of 8 set employed omission on trials was too liberal when compared with the 4 set interval employed on stimulation trials. Nevertheless, the point remains that even with an extended response interval, significant recovery was generally not demonstrated. The most consistent finding to emerge from the present experiments concerns the relationship between SCR recovery and electrodermal lability (Crider and Lunn, 1971), and in this respect, the data are in agreement with previous findings (Bernstein, 1969; O’Gorman et al., 1970; Siddle, 1974). It seems unlikely that the SCRs displayed by labile subjects during the omission interval were a mere manifestation of their high rate of spontaneous activity

292

D. A. T. Siddle and P. A. Heron

since this factor was controlled by utilizing within-subject comparisons. These comparisons indicated that the SCRs displayed on omission trials were significantly larger than those displayed on either the last training trial or during the control interval just prior to the omission trial. O’Gorman (1972) has suggested that SCR recovery to stimulus change is, in part, a function of the degree of inhibitory control exerted in the electrodermal system and the results of these experiments are consistent with such an interpretation. Acknowledgement

This research was supported Council.

by Grant

B/RG/44047

from the Science Research

References Allen, C. K., Hill, F. A. and Wickens, D. D. (1963). The orienting reflex as a function of the interstimulus interval of compound stimuli. Jownal of Experimental Psycholog),, 65, 309-316. Badia, P. and Defran, R. H. (1970). Orienting responses and GSR conditioning: a dilemma. Psychological Rrciew, 17, 171-181. Bernstein, A. S. (1969). To what does the orienting response respond? Psychophysiology, 6, 338-350. Bernstein, A. S., Taylor, K., Austen, B. G., Nathanson, M. and Scarpelli, A. (1971). Orienting response and apparent movement toward or away from the observer. Journal of Experimental Psychology, 87, 37-45. Corman, C. D. (1967). Stimulus generalization of habituation of the galvanic skin response. Journal of E,vperimental P.sq.chology, 74, 236-240. Crider, A. and Lunn, R. (1971). Electrodermal lability as a personality dimension. Journa/ sf Experimental Research in PersonalitJJ, 5, 145-l 50. Dykman, R. A., Reese, W. G., Galbrecht, C. R. and Thomasson, P. J. (1959). Psychophysiological reactions to novel stimuli: measurement, adaptation, and relationship of psychological and physiological variables in the normal human. Annals of the New York Academy of Science, 79,43-107. Fried, R., Welch, L., Friedman, M. and Gluck, S. (1967). Is no stimulus a stimulus? Journal of Experimental Psychology, 76, 145-146. Furedy, J. J. (1968a). Human orienting reaction as a function of electrodermal versus plethysmographic response modes and single versus alternating stimulus series. Journal of Experimental Psychology, 77, 7G-80. Furedy, J. J. (1968b). Novelty and the measurement of the GSR. Jorwnal of E.t~erimental Psycholov, 76, 501-503. Gardner, L. E. (1968). Retention and overhabituation of a dual-component response in Iamhricrrs tcrrestris. Journal of Comparatiae and Physiological Psychofog]., 66, 3 15-318. Geer, J. H. (1969). Generalization of inhibition in the orienting response. Psvchophysiology, 6, 197-201. Gliner, J. A., Harley, J. P. and Badia, P. (1971). Elicitation and habituation of the orienting response as a function of instructions, order of stimulus presentation, and omission. Joknot of Experimental Psychology, 89,414416. Graham. F. K. (1973). Habituation and dishabituation of responses innervated by the autonomic nervous system. In: Peeke, H. V. S. and Herr, M. J. (Eds.) Habituation. vol. I, Behaoioral studies. Academic Press : New York, 1633218. Grandage, A. (1958). Orthogonal coefficients for unequal intervals. Biometrics,14,287-289. Groves, P. M. and Thompson, R. F. (1970). Habituation: a dual-process theory. Psychological Rwiew, 77, 4 19-450. Harley, J. P. (1973). Temporal conditioning as a function of instructions and intertrial interval. Journal of E.wperimental Psychology, 100, 1788184.

Stimulus

omission

and SCR and FP V response recovery

293

Houck, R. L. and Mefferd, R. B. (1969). Generalization of GSR habituation to mild intramodal stimuli. Pyschophysiology, 6, 2022206. James, J. P., Daniels, K. R. and Hanson, B. (1974). Overhabituation and spontaneous recovery of the galvanic skin response. Journal of Experimental Psychology, 102, 732-134. Lacey, J. I. and I.acey, B. C. (1958). The relationship of resting autonomic activity to motor impulsivity. Research Publications, Association for Research in Nervous and Mental Disease, 36, 144-209. Lockhart, R. A. (1966). Temporal conditioning of GSR. JournalofExperimenta/Psychology~, 71,438446. Luria, A. R. and Homskaya, E. D. (1970). Frontal lobes and regulation of arousal processes. Jn: Mostofsky, D. I. (Ed.) Attention: Contemporary Theory and Analysis. AppletonCentury-Crofts: New York, 303-330. Lykken, D. T. and Venables, P. H. (1971). Direct measurement of skin conductance: a proposal for standardization. Psychophysiology, 8, 656672. Morgenson, D. F. and Martin, I. (1968). The orienting response as a predictor of autonomic conditioning. Journal of Experimental Research in Personality, 3, 89-98. O’Gorman, J. G. (1972). Electrodermal lability and recovery of habituated OR. Australian Journal of Psychology, 24,241-244. O’Gorman, J. G. (1973). Change in stimulus conditions and the orienting response. Psychophysiology, 10,465-470. O’Gorman, J. G., Mangen, G. L. and Gowen, J. A. (1970). Selective habituation of galvanic skin response component of the orientation reaction to an auditory stimulus. Psychophysiology, 6,7166721. Siddle, D. A. T. (1974). Overextinction of the evoked skin conductance resoonse: an EEG study. Psychophysiology, 11,630-638. Sokolov. E. N. (1960). Neuronal models and the orienting reflex. In: Brazier. M. A. (Ed.) The Central Nervous System and Behavior. Josiah Ma&y, Jr. Foundation: New York: 187-276. Sokolov, E. N. (1963). Perception a&the Conditioned Reflex. Pergamon Press: Oxford. Sokolov, E. N. (1966). Orienting reflex as information regulator. In: Leontiev, A., Luria, A. and Smirnov, A. (Eds.) Psychological Research in the U.S.S.R., Vol. 1. Progress Publishers: Moscow, 334-360. Sokolov, E. N. (1969). The modelling properties of the nervous system. In: Cole, M. and Maltzman, I. (Eds.) Handbook of Contemporary Soviet Psychology. Basic Books: New York, 671-704. Stephenson, D. and Siddle, D. A. T. (in press). Effects of ‘below-zero’ habituation on the electrodermal orienting response to a test stimulus. Psychophysiology. Still, A. W. (1967). Use of orthogonal polynomials with nonparametric tests. Psychological Bulleiin, 68, 327-329. Sutton, S., Tueting, P., &bin, J. and John, E. R. (1967). Information delivery and the sensory evoked potential. Science, 155, 14361439. Thompson, R. F. and Spencer, W. A. (1966). Habituation: a model phenomenon for the study of neuronal substrates of behaviour. Psychological Review, 73, 16-43. Yaremko, R. M. and Keleman, K. (1972). The orienting reflex and amount and direction of conceptual novelty. Psychonomic Science, 27, 195-196.