Scand. J . Psychol. 20, 37-41, 1979
Heart-rate and electroder maI orienting responses to visual stimuli differing in complexity MATS FREDRIKSON ARNE OHMAN
University of Uppsala, Sweden
Fredrikson, M. & Ohman, A.: Heart-rate and electrodermal orienting responses to visual stimuli differing in complexity. Scand. J . Psychol. 20, 3741, 1979, Orienting response (OR) theory predicts that amount of information in the stimulus, or stimulus complexity, should be an important determinant of OR elicitation and habituation, more intense and more slowly habituating ORs being expected to complex than to sinlple stimuli. This prediction was tested in an experiment where subjects were exposed to simple and complex visual stimuli in randomized order, while heart-rate and skin conductance were measured. Complex stimuli evoked a more pronounced deceleratory heartrate response than did simple stimuli. However, the two conditions did not differ in rate of habituation of this response. For skin conductance responses, on the other hand, the complex stimulus took more trials than the simple stimulus to reach habituation, whereas the two conditions did not differ in response magnitude. Thus, the hypothesis of more intense orienting to complex stimuli was supported by the heart-rate data, and that of slower habituation to complex stimuli, of the skin conductancedata. M . Fredrikson, Department of Psychology, University of Uppsala, Box 227, S-75104 Uppsala, Sweden.
Sokolov (1963, p. 12) defined the orienting response (OR) as “a complex combination of somatic and autonomic reactions forming a complete functional system”. This functional system involves sensory, muscular, autonomic, and electrocortical changes to novel or significant stimulation, or to any change in some parameter of a previously presented stimulus. With repeated stimulation, the OR components decrease in amplitude until they eventually disappear completely, i.e. habituation occurs (Graham, 1973; Sokolov, 1963). Habituation of autonomic components of the OR has been extensively studied during recent years (see review by Graham, 1973). Heart-rate responses have occasioned special interest, since they have been suggested to afford differentiation between ORs and defensive responses (DRs) (Graham & Clifton, 1966). Thus, auditory stimuli of low to moderate intensity elicit a phasic deceleration in heart rate immediately at stimulus onset, provided that the onset characteristics of the stimulus is controlled, whereas intense stimuli produce a phasic
acceleration (Graham, 1973). The perhaps most widely used autonomic OR component, skin conductance responses (SCRs), in general does not permit this differentiation, although some recent data from our laboratory appear promising for such a distinction (Ohman, et al. 1978). Psychological studies of OR habituation typically have used auditory stimuli. Some potentially interesting experimental manipulations, however, are more easily studied by visual stimulation. Sokolov (1%3; 1%9) proposed that habituation of the OR is due to the development of a “stimulus model”, which incorporates various parameters of the stimulus, becoming more complete for each stimulus presentation. A “comparator mechanism” examines this model in relation to the actual stimulus at each presentation, suppressing the OR if a match occurs, and producing an OR in case of a mismatch. For such an account, amount of information in the stimulus, or its complexity, should be an important determinant of the magnitude and habituation of the OR since more information would reScand. J . Psychol. 20
38
M . Fredrikson and A . Ohrnan
quire a more complex stimulus model, which presumably would take longer to develop. Few studies, however, have examined this variable. Berlyne and coworkers have reported data indicating that amount of information or complexity of visual stimuli indeed affects some components of the OR. Berlyne et al. (1963) found evidence of more electrodermal ORs t o complex than to simple visual stimuli and Berlyne & McDonnel (1965) found alpha-blocking in the E E G t o persist longer after complex than after simple stimuli. I n neither of the studies, however, were the effects found very impressive. In the former it was significant only when the subjects were instructed t o actively attend to the stimuli, and in the latter it was reliable only when data were collapsed over several complexity conditions. Furthermore, in neither of the studies were data on habituation reported. Such data, however, were reported by Spinks & Siddle (1976) who found stimuli containing less information to require fewer presentations to produce skin conductance response habituation, estimated as total numbers of responses through the experiment and number of trials to criterion measure, than did those containing more information. In another study O G o r m a n (1971) found a relationship between amount of information in the stimulus and number of trials t o reach a criterion of habituation of skin conductance responses. In view of the theoretical centrality of the complexity variable, and in view of the scarcity of data, we decided to examine habituation of autonomic responses to visual stimuli differing in complexity. Since heartrate data regarding this problem are lacking, and since this measure has been regarded as a particulary valid OR indicator (e.g. Graham, 1973), and as sensitive t o processing of external stimuli (e.g. Lacey & Lacey, 1974), this measure was our primary dependant variable, although skin conductance data were also collected. An additional purpose of the study was to compare various measures of heart-rate orienting. METHOD
Subjects Fifteen psychology students at the University of Uppsala were paid to participate as subjects. Eleven were females and 4 were males. The age range was 20-29 years. Apparatus and recording methods
The subject was comfortably seated in an armchair in a Tegntr sound attenuating cubicle. The physiological variScand. J . Psycliol. 20
ables were recorded on paper using a Hewlett-Packard 7700 Polygraph. Skin conductance responses were recorded by Beckman-Offner silver/silver chloride skin electrodes, 9 mm in diameter and embraced in plastic cups, through a Hagfors-type constant voltage circuit (Venables & Christie, 1973). An isotonic electrode paste was used as the electrolyte (0.58 g NaCl per 100 ml). The electrodes were applied by means of adhesive collars to the palmar side of the second phalanx of the subject’s left hand. Heart-rate responses were recorded by HewlettPackard electrodes placed approximately 10 cm apart on the lower frontal aspect of the rib cage, with an electrode on the subject’s right shoulder serving as ground. A Hewlett-Packard bioelectric aplifier (HP 881 1A) picked up the electrocardiogram and fed it into a cardiotachometer (HP rate computer 8812A), which provided an output calibrated in beats per minute (BPM). Respiration was measured by a straingauge fastened around the subject’s chest. It was used as control variable to exclude trials with respiratory irregularities from further analysis. Stimuii
Color slides, 24x36 mm, were projected from a Sawyer projector into the cubicle and on to a screen approximately 2 m in front of the subject. This arrangement resulted in a 60x95 picture on the screen. The exposure time was controlled by an electronic timer, which was trigged by relay detectors activated by a tape-recorder (Tandberg) upon which the interstimulus intervals had been recorded. The pictorial stimuli were constructed to differ in several of the complexity dimensions used by Berlyne (1958). The simple stimulus consisted of the projector light projected through a green or a yellow filter. For the complex stimulus, a complex abstract drawing of a number of irregular and heterogenous elements randomly arranged was superimposed on the filters. Each subject saw one simple and one complex stimulus of the same color. Design The design was a 2x 10 randomized block design (Kirk, 1%8), with two levels of complexity as one factor, and 10 repetitions of the stimulus as the other.
Procedure
The subject was seated in the experimental chamber and the various electrodes and transducers were applied, while a male experimenter informally explained the general purpose of the experiment. The formal instructions stressed that the subject would see two types of pictures on the screen in front of him, and that he was expected to relax and pay attention to them while the experimenter measured his physiological responses. After the application of the electrodes the subject was allowed to rest a few minutes before the experiment started. Each picture was shown for 8 sec and the interstimulus interval varied between 20 and 40 sec with a mean of 30 sec. The two pictures were presented in random order with the restriction that no more than two successive presentations of the same stimulus were allowed. There
Orienting as a function of stimulus complexity
-3L
-3 -onset
KCONDS TRIAL 1-5
Fig. 1 . Heart-rate responses to complex (0-0) and simple (0-0)visual stimuli during the first (left panel) and second (right panel) half of the habituation series.
L
I
39
t KCONDS
S - a t
were 10 presentations of each picture, a number of trials that formally has been shown sufficient for studying habituation of SC- and HR-responses (Graham, 1973).
Response definitions Skin conductance responses. A change in conductance initiated in the interval 1-4 sec after stimulus onset was scored as a response if it exceeded 0.05 micromho. Following Lykken (1972) the responses were range corrected, i.e. each response for a certain subject was divided by the maximum response emitted to a stimulus by that subject. Heart-rate responses. The output from the cardiotachometer was scored each second for 5 prestimulus and 10 poststimulus seconds. These data were transformed to poststimulus change score by subtracting each poststimulus second from the mean of the 5 prestimulus seconds, and the 10 trials were blocked in two trial blocks of 5 trials each. These data were subjected to a series of statistical analyses. First, an overall analysis of variance was performed including stimulus complexity, trial blocks, and poststimulus seconds as factors. Then various points of interest were identified (see Gatchel & Lang, 1973) and subjected to different analyses according to practices reported in the literature. An initial deceleratory component was scored as the maximum deceleration during the first 3 sec ( E D , ) (cf. Gatchel & Lang 1973, Jackson, 1974) or mean deceleration during the first 3 sec. Secondly, a deceleratory component peaking 5-8 sec after stimulus onset was scored as change from prestimulus level (B-D2; cf. Gatchel & Lang, 1973) or as change from the maximal acceleration 3-5 sec after stimulus onset (A-D,; cf. Lang & Hnatiow, 1%2). Thirdly, total deceleration was scored as the sum of the initial and secondary decelerations ((B-D,)+@-D,)) (Gatchel & Lang, 1973). Finally, heart rate variability was scored as the standard deviation (SD) during the 10 poststimulus sec (F’rigatano & Johnson, 1974).
RESULTS
Heart-rate responses The mean prestimulus level for the simple stimulus condition was 71.28 BPM and that for the complex
TRIALG-I0
stimulus was 71.59 BPM. This difference, of course, was not significant, nor were the trials and trials x complexity effects, which indicates a stable base level that did not differ between conditions. The average shapes for the heart-rate responses are shown in Fig. 1 , which depicts heart-rate changes in BPM as a function of poststimulus seconds for the simple and compIex stimulus in two blocks of 5 trials each. An overall analysis of variance confirmed the more pronounced deceleratory response to the complex stimulus, F (1, 14)=5.94,p
Heart-rate and electroder maI orienting responses to visual stimuli differing in complexity MATS FREDRIKSON ARNE OHMAN
University of Uppsala, Sweden
Fredrikson, M. & Ohman, A.: Heart-rate and electrodermal orienting responses to visual stimuli differing in complexity. Scand. J . Psychol. 20, 3741, 1979, Orienting response (OR) theory predicts that amount of information in the stimulus, or stimulus complexity, should be an important determinant of OR elicitation and habituation, more intense and more slowly habituating ORs being expected to complex than to sinlple stimuli. This prediction was tested in an experiment where subjects were exposed to simple and complex visual stimuli in randomized order, while heart-rate and skin conductance were measured. Complex stimuli evoked a more pronounced deceleratory heartrate response than did simple stimuli. However, the two conditions did not differ in rate of habituation of this response. For skin conductance responses, on the other hand, the complex stimulus took more trials than the simple stimulus to reach habituation, whereas the two conditions did not differ in response magnitude. Thus, the hypothesis of more intense orienting to complex stimuli was supported by the heart-rate data, and that of slower habituation to complex stimuli, of the skin conductancedata. M . Fredrikson, Department of Psychology, University of Uppsala, Box 227, S-75104 Uppsala, Sweden.
Sokolov (1963, p. 12) defined the orienting response (OR) as “a complex combination of somatic and autonomic reactions forming a complete functional system”. This functional system involves sensory, muscular, autonomic, and electrocortical changes to novel or significant stimulation, or to any change in some parameter of a previously presented stimulus. With repeated stimulation, the OR components decrease in amplitude until they eventually disappear completely, i.e. habituation occurs (Graham, 1973; Sokolov, 1963). Habituation of autonomic components of the OR has been extensively studied during recent years (see review by Graham, 1973). Heart-rate responses have occasioned special interest, since they have been suggested to afford differentiation between ORs and defensive responses (DRs) (Graham & Clifton, 1966). Thus, auditory stimuli of low to moderate intensity elicit a phasic deceleration in heart rate immediately at stimulus onset, provided that the onset characteristics of the stimulus is controlled, whereas intense stimuli produce a phasic
acceleration (Graham, 1973). The perhaps most widely used autonomic OR component, skin conductance responses (SCRs), in general does not permit this differentiation, although some recent data from our laboratory appear promising for such a distinction (Ohman, et al. 1978). Psychological studies of OR habituation typically have used auditory stimuli. Some potentially interesting experimental manipulations, however, are more easily studied by visual stimulation. Sokolov (1%3; 1%9) proposed that habituation of the OR is due to the development of a “stimulus model”, which incorporates various parameters of the stimulus, becoming more complete for each stimulus presentation. A “comparator mechanism” examines this model in relation to the actual stimulus at each presentation, suppressing the OR if a match occurs, and producing an OR in case of a mismatch. For such an account, amount of information in the stimulus, or its complexity, should be an important determinant of the magnitude and habituation of the OR since more information would reScand. J . Psychol. 20
38
M . Fredrikson and A . Ohrnan
quire a more complex stimulus model, which presumably would take longer to develop. Few studies, however, have examined this variable. Berlyne and coworkers have reported data indicating that amount of information or complexity of visual stimuli indeed affects some components of the OR. Berlyne et al. (1963) found evidence of more electrodermal ORs t o complex than to simple visual stimuli and Berlyne & McDonnel (1965) found alpha-blocking in the E E G t o persist longer after complex than after simple stimuli. I n neither of the studies, however, were the effects found very impressive. In the former it was significant only when the subjects were instructed t o actively attend to the stimuli, and in the latter it was reliable only when data were collapsed over several complexity conditions. Furthermore, in neither of the studies were data on habituation reported. Such data, however, were reported by Spinks & Siddle (1976) who found stimuli containing less information to require fewer presentations to produce skin conductance response habituation, estimated as total numbers of responses through the experiment and number of trials to criterion measure, than did those containing more information. In another study O G o r m a n (1971) found a relationship between amount of information in the stimulus and number of trials t o reach a criterion of habituation of skin conductance responses. In view of the theoretical centrality of the complexity variable, and in view of the scarcity of data, we decided to examine habituation of autonomic responses to visual stimuli differing in complexity. Since heartrate data regarding this problem are lacking, and since this measure has been regarded as a particulary valid OR indicator (e.g. Graham, 1973), and as sensitive t o processing of external stimuli (e.g. Lacey & Lacey, 1974), this measure was our primary dependant variable, although skin conductance data were also collected. An additional purpose of the study was to compare various measures of heart-rate orienting. METHOD
Subjects Fifteen psychology students at the University of Uppsala were paid to participate as subjects. Eleven were females and 4 were males. The age range was 20-29 years. Apparatus and recording methods
The subject was comfortably seated in an armchair in a Tegntr sound attenuating cubicle. The physiological variScand. J . Psycliol. 20
ables were recorded on paper using a Hewlett-Packard 7700 Polygraph. Skin conductance responses were recorded by Beckman-Offner silver/silver chloride skin electrodes, 9 mm in diameter and embraced in plastic cups, through a Hagfors-type constant voltage circuit (Venables & Christie, 1973). An isotonic electrode paste was used as the electrolyte (0.58 g NaCl per 100 ml). The electrodes were applied by means of adhesive collars to the palmar side of the second phalanx of the subject’s left hand. Heart-rate responses were recorded by HewlettPackard electrodes placed approximately 10 cm apart on the lower frontal aspect of the rib cage, with an electrode on the subject’s right shoulder serving as ground. A Hewlett-Packard bioelectric aplifier (HP 881 1A) picked up the electrocardiogram and fed it into a cardiotachometer (HP rate computer 8812A), which provided an output calibrated in beats per minute (BPM). Respiration was measured by a straingauge fastened around the subject’s chest. It was used as control variable to exclude trials with respiratory irregularities from further analysis. Stimuii
Color slides, 24x36 mm, were projected from a Sawyer projector into the cubicle and on to a screen approximately 2 m in front of the subject. This arrangement resulted in a 60x95 picture on the screen. The exposure time was controlled by an electronic timer, which was trigged by relay detectors activated by a tape-recorder (Tandberg) upon which the interstimulus intervals had been recorded. The pictorial stimuli were constructed to differ in several of the complexity dimensions used by Berlyne (1958). The simple stimulus consisted of the projector light projected through a green or a yellow filter. For the complex stimulus, a complex abstract drawing of a number of irregular and heterogenous elements randomly arranged was superimposed on the filters. Each subject saw one simple and one complex stimulus of the same color. Design The design was a 2x 10 randomized block design (Kirk, 1%8), with two levels of complexity as one factor, and 10 repetitions of the stimulus as the other.
Procedure
The subject was seated in the experimental chamber and the various electrodes and transducers were applied, while a male experimenter informally explained the general purpose of the experiment. The formal instructions stressed that the subject would see two types of pictures on the screen in front of him, and that he was expected to relax and pay attention to them while the experimenter measured his physiological responses. After the application of the electrodes the subject was allowed to rest a few minutes before the experiment started. Each picture was shown for 8 sec and the interstimulus interval varied between 20 and 40 sec with a mean of 30 sec. The two pictures were presented in random order with the restriction that no more than two successive presentations of the same stimulus were allowed. There
Orienting as a function of stimulus complexity
-3L
-3 -onset
KCONDS TRIAL 1-5
Fig. 1 . Heart-rate responses to complex (0-0) and simple (0-0)visual stimuli during the first (left panel) and second (right panel) half of the habituation series.
L
I
39
t KCONDS
S - a t
were 10 presentations of each picture, a number of trials that formally has been shown sufficient for studying habituation of SC- and HR-responses (Graham, 1973).
Response definitions Skin conductance responses. A change in conductance initiated in the interval 1-4 sec after stimulus onset was scored as a response if it exceeded 0.05 micromho. Following Lykken (1972) the responses were range corrected, i.e. each response for a certain subject was divided by the maximum response emitted to a stimulus by that subject. Heart-rate responses. The output from the cardiotachometer was scored each second for 5 prestimulus and 10 poststimulus seconds. These data were transformed to poststimulus change score by subtracting each poststimulus second from the mean of the 5 prestimulus seconds, and the 10 trials were blocked in two trial blocks of 5 trials each. These data were subjected to a series of statistical analyses. First, an overall analysis of variance was performed including stimulus complexity, trial blocks, and poststimulus seconds as factors. Then various points of interest were identified (see Gatchel & Lang, 1973) and subjected to different analyses according to practices reported in the literature. An initial deceleratory component was scored as the maximum deceleration during the first 3 sec ( E D , ) (cf. Gatchel & Lang 1973, Jackson, 1974) or mean deceleration during the first 3 sec. Secondly, a deceleratory component peaking 5-8 sec after stimulus onset was scored as change from prestimulus level (B-D2; cf. Gatchel & Lang, 1973) or as change from the maximal acceleration 3-5 sec after stimulus onset (A-D,; cf. Lang & Hnatiow, 1%2). Thirdly, total deceleration was scored as the sum of the initial and secondary decelerations ((B-D,)+@-D,)) (Gatchel & Lang, 1973). Finally, heart rate variability was scored as the standard deviation (SD) during the 10 poststimulus sec (F’rigatano & Johnson, 1974).
RESULTS
Heart-rate responses The mean prestimulus level for the simple stimulus condition was 71.28 BPM and that for the complex
TRIALG-I0
stimulus was 71.59 BPM. This difference, of course, was not significant, nor were the trials and trials x complexity effects, which indicates a stable base level that did not differ between conditions. The average shapes for the heart-rate responses are shown in Fig. 1 , which depicts heart-rate changes in BPM as a function of poststimulus seconds for the simple and compIex stimulus in two blocks of 5 trials each. An overall analysis of variance confirmed the more pronounced deceleratory response to the complex stimulus, F (1, 14)=5.94,p
