The Interstimulus Interval in Classical Autonomic Conditioning of Young Infants EBEN M. INGRAM Department of Psychology University of Colorado at Denver Denver, Colorado Conditioned auditory discrimination and extinction of the skin potential response were attempted in 4-month-old infants using interstimulus intervals of 1500, 3500, 5500, and 7500 msec. Half of the infants in each of the interstimulus interval groups were defined as high magnitude orienters and half were low magnitude orienters. Conditioning was successful with the 5500- and 7500-msec interstimulus intervals, but not with the 1500- and 3500-msec intervals. Analysis of individual subject data indicated that individual differences in conditionability were related to interstimulus interval and orienting response magnitude. Also, those subjects discriminating at the longer intervals tended to be high magnitude orienters. In other words, longer interstimulus intervals interacted with a high magnitude of orienting to facilitate conditioning. The results were taken as evidence that individual differences in the magnitude of the orienting response reflect different individual needs in stimulus information processing time.

During the past 15 years infant conditioning research has been concerned with specifying the conditions that influence the course of conditional reflex formation (Fitzgerald & Brackbill, 1970). Although a clearly optimal interstimulus interval (ISI) has not been identified (Gormezano & Moore, 1969), research with nonhuman and human adult subjects seems to indicate that longer ISI’s afford better conditioning performance when the response is an autonomic one (Fitzgerald & Brackbill, in press; Gerall & Woodward, 1958; Hastings & Obrist, 1967; Kakigi, 1964; Pavlov, 1927) and when discrimination conditioning rather than simple conditioning is attempted (Hartman & Grant, 1962; Hilgard, Campbell, & Sears, 1938; Jensen & Prokasy, 1973; Kimmel & Pennypacker, 1963). Most of the existing research up to this point has yielded little evidence for assuming that any specific IS1 would yield superior performance with infants (Fitzgerald & Lintz, 1967; Morgan & Morgan, 1944; Natio & Lipsitt, 1968; RendleShort, 1961; see Wenger, 1936). However, closer examination of some of the above studies indicates that differential performance does occur to different ISI’s. For instance, Lintz, Fitzgerald, and Brackbill (1967) found that an IS1 of 1000 msec

Reprint requests should be sent to Eben M . Ingram, University of Colorado at Denver, Division of Natural and Physical Science, Department of Psychology, 1100 Fourteenth Street, Denver, Colorado 80802, U.S.A. Received for publication 8 March 1976 Revised for publication 6 May 1977 Developmental Psychobiologv, 11(5):419-426 (1978) @ 1978 b y John Wiley & Sons, Inc.

0012-1630/78/0011-0419$01.OO

420

INGRAM

produced 90% conditioned responses (CR’s) in an eyeblink conditioning study. On the other hand, Natio and Lipsett (1968), in a similar study, found that a 500-msec IS1 produced a mean of 27% CR’s. However, the differences in the percent of resporiding may be attributable to patent methodologic differences. A much clearer picture of the relationships of IS1 to conditioning in infants and children is presented b!/ researches by Ohlrich and Ross (1968) and Little (1971) that demonstrate that the longer ISI’s afford better conditioning with infants and children. The above studies tend to support the contention that infants condition best at ISI’s longer than 500 msec, at least with somatic responses. Brackbill et al. (1967) failed to find any IS1 value effective in infant auditory, pupillary reflex conditioning. On the basis of the above, one can contend that the relationship between conditionality and the IS1 is complex and the formation of any developmental theory of conditionability specifying only a small po’rtion of the components of the classical paradigm is questionable. Individual differences in conditionability have been a focal issue in human adult personality research (Eysenck & Levey, 1972; Gray, 1972), and Pav1ovia.n theory of higher nervous system activity (Nebylitsyn, 1972; Teplov, 1972). Recently, individual differences in autononiic conditionability have been found to be an important dimension of early learning, particularly when the infant is exposed to a complex conditioning procedure (Brackbill & Fitzgerald, 1972; Ingram & Fitzgerald,, 1974). Thc notion of individual differences in conditionability is theoretically related to the Pavlovian idea of nervous system dynamism (Nebylitsyn, 1972) a i d mobility (Volokhov, 1972), and the contention by Sokolov (1963) that orienting response (OR) elicitation facilitates conditional reflex formation, as well as being a necessary prerequisite for conditioning. Several studies with human adults support Sokolov’s position that OR elicitation facilitates conditional reflex formation (Maltzman & Mandell, 1968; Morgenson & Martin, 1968). Kasatkin (1972) has reviewed several infant experiments that support Sokolov’s position, as well as the derivative prediction that individual differences in OR magnitude are related to individual differences in conditionability (Ingram & Fitzgerald, 1974; Zeimer & Schell, 1971). To date only one study (Ingram & Fitzgerald, 1974) has examined the Sokolovian contentions with human infants as subjects. The present study was designed (a) to investigate the relationship between IS1 and conditionability as well as the relationship between IS1 and individual differences in conditionability, (b) to examine the hypothesis that individual differences in OR magnitude are related to individual differences in conditionability, and (c) to extend the study of infant autonomic conditioning through replication of a previous at tcmpt to condition the skin potential response (SPR).

Method Subjects The subjects ranging in age at 112 days; median more parents at

were 30 full term clinically normal, home-reared human infants the beginning of experimentation from 85 t o 120 days (mean age, age, 110 days). The infants were brought to the laboratory by 1 or which time parental consent for the infants’ participation was

THE IS1 IN INFANT CONDITIONING

42 1

obtained. Three infants failed to complete the experiment due to excessive crying that persisted during each of the sessions.

Design and Procedure The experimental procedure consisted of 3 parts: adaptation, discrimination conditioning, and extinction. Each infant upon appearance at the laboratory was assigned to 1 of 4 different groups on the basis of orienting response (OR) magnitude such that each group contained 3 high magnitude OR subjects and 3 low magnitude OR subjects. The determination of OR and the cut-off value between high and low OR magnitude were based on a previous study (Ingram & Fitzgerald, 1974).

Independent and Dependent Variables The dependent variable was the SPR component of the OR. The SPR was recorded on a Grass model 7 polygraph with Beckman Bio-Potential electrodes and an NaC1-based electrolytic medium. The active electrode was placed on the medial side of the foot over the abductor hallucis muscle and midway between the first phalange and a point directly below the ankle (Edelberg, 1967); the referred electrode was placed over the tibia bone, one-eighth of the way up the shin between the ankle and knee; and the group electrode was placed on the outside of the upper thigh. The skin surface was prepared with 70% ethanol, abrading slightly the inactive site. The SPR was calculated in multivolts from the negative and positive components according to the procedure devised by Raskin, Koses, and Bever (1969). The computation for analysis consisted of determining the arithmetical difference between response magnitude to CS+ and CS-. (CS+ refers to the conditional stimulus followed by reinforcement and CS- refers to the stimulus not followed by reinforcement.) The unconditioned stimulus (US) was .36 kg/cm2 air puff delivered through .318-cm diameter surgical tubing to the infant’s right cheek, half-way between the nose and ear lobe, from a distance of 2.5 cm. The apparatus for US delivery consisted of an air tank filled with compressed air, a Hoke Regulator Valve, and a Hunter Silent Solenoid. The conditional stimuli (CS) were 400- and 800-Hz, 75-dB, square-wave tones generated by an Eico Audio Generator delivered at 60-cm distances and in a horizontal plane with the infant’s head. The US-duration was 1000 msec; CS durations were 2500, 4500, 6500, and 8500 msec for Groups 1 through 4, respectively. The CS and US terminated simultaneously. The intertrial interval was varied randomly among 10, 15, and 20 sec to avoid confounding due to temporal conditioning. All intervals and stimulus temporal parameters were electronically controlled. The different ISI’s (1500, 3500, 5500, and 7500 msec) represent the 4 experimental groups. A 5th group of infants served as a pseudoconditioning control. Each group contained 6 subjects, with half of the infants in each group receiving the 800-Hz tones as CS+, whereas the other half received the 400-Hz tone as CS+. During Day 1 (Session 1) each infant received 15 presentations of the 800-Hz tone and 15 presentations of the 400-Hz tone in mixed order for a total of 30 presentations. The adaptation IS1 varied randomly between I 5 and 20 sec with each stimulus duration being the same length as the CS duration for each respective

422

INGRAM

experimental group. On the 1st adaptation trial all subjects received the 800-Hz tone for a duration of 5 sec with subsequent trials as described above. During Days 2, 3 , and 4 (Sessions 2 , 3 , and 4) all experimental subjects received 1.5 CS+ and 1.5 CStrials for a total of 30 conditioning trials per day. On Day 5 (Session 5 ) all subjects were given 1.5 presentations of the CS- and 15 presentations of CS+ (without reinforcement) for a total of 30 extinction trials. The infants serving as pseudoconditioning controls differed from the experimental infants only during Sessions 2 , 3, and 4. The Control group received the same stimuli and stimulus presentation schedule on Sessions 1 and 5 as did the experimental groups. The basic difference between control and experimental groups was that the US was not paired with CS+ for controls at any time.

Results The responses for Session 1 were divided into 3 blocks of 10 trials each, each containing equal numbers of the 400- and 800-Hz stimuli. A significant decrement in response magnitude was found between the 1st and last blocks of trials for all subjects ( t = 3.21, df = 29, p < .Ol), but no significant differences were found between response magnitudes to the 400- and 800-Hz tones (p > .OS), suggesting that differences occurring during conditioning were due to experimental contingencies rather than preexisting subject differences. All trials from Sessions 2 , 3 , and 4 were combined, then divided into blocks o f 10 trials each, with each block of trials containing equal numbers of C S t and CS- trials. The mean response magnitudes of CSt trials were compared with the mean response magnitude of CS- trials and a difference score computed for each Trial Block. The data analyses consisted of a 3-factor (ISI, OK, and Trial Blocks) analysis of variance (Winer, 1962) with repeated measures on 1 factor (Trial Blocks), an analysis of each subject’s data using a t test for correlated measures, and an analysis of differences among the IS1 groups using Duncan’s Multiple Range Test (Duncan, 1951). Significant effects were associated with ISI (F = 5.49, df = 4/20, p < .005j, OR ( F = 10.47, df = 1/20, p < .001), and Trial Blocks (F = 9.34, d f = 81160, p < .OOl). In addition, the Trial Blocks x OR interaction was significant ( F = 15.99, df = 8/160, p < .OOl). The analysis of the individual data revealed that 7 of the 30 infants demonstrated a significant level of differential responding to CS+. Of these infants, 4 were in the 7500-msec IS1 group and 3 were in the 5500-msec IS1 group. Along with this distribution of subjects, the differences among the IS1 groups support the notion that longer ISI’s were more effective in producing conditioning. The 5.500- and 7500-nisec groups showed a rapid increase in mean difference response magnitude, whereas the 1500-msec, 3500-msec, and control sub:jects showed only slight increases in mean difference response magnitude (see Fig. 1). The control subjects showed very little change in differential responding across acquisition sessions. The performance of the high and low OR subjects, collapsed across groups, showed an initial gradual increase with a peak in responding, which was filllowed by a decrease in responding (see Fig. 2). The level of differential responding again increased for the high OK subjects but continued to decrease for the low OR subjects. In

THE IS1 IN INFANT CONDITIONING

423

D e c i l e B l o c k s of T r i a l s Adaptation

Acquisition

Extinction

Fig, 1. The mean response magnitudes and difference score magnitudes for each group of subjects. Left panel: mean response magnitude for each group of subjects during adaptation (mV/mm). Middle and right panels: mean difference score magnitudes (mV/mm) between CS+ and CS- for each group of subjects during acquisition and extinction. 0 - 0 , 1500 msec; 0- 4, 3500 msec; 0- -0, 5500 msec; o-. -0 , 7500 msec; 0 - 0 , controls.

addition to the higher level of differential responding, the high OR subjects also responded with greater frequency to CS+ over CS- than did the low OR subjects ( t = 2.36, df = 14, p < .05).

Extinction A significant decrement in responding from the 1st to last blocks of extinction trials occurred for all subjects (t = 2.79, df = 29, p < .02; see Fig. 1). However, when comparisons between groups were made no differences were revealed in rates of extinction. On the other hand, a large difference existed between high and low OR subjects across extinction trials (t = 14, df = 14, p < .05; see Fig. 2). The difference between the 2 groups may be accounted for by the existence of a large magnitude difference on the last 2 blocks of conditioning trials (Trial Blocks 8 and 9 during acquisition). High OR subjects showed a greater decrease in responding during extinction ( t = 2.47, df = 14, p < .05),suggesting faster extinction.

Oecile Blocks of Trials Adaptation

Acquisition

Extinction

Fig. 2. The mean response magnitude and difference score magnitues for high ( 0 - 0 ) and low OR subjects. Left panel: mean response magnitudes (mV/mm) for high and low OR subjects during adaptation. Middle and right panels: mean difference score magnitudes (mV/m ) between CS+ and CS- for high and low OR subjects during acquisition and extinction. (0-0)

424

INGRAM

Discussion The results of the present study indicate that (1) the interstimulus interval has a definite influence on conditionability, such that conditionability increases as the length of IS1 increases; (2) individual differences exist in OR magnitude that are related to conditionability, such that conditionability refers to the rapidity with which conditioning occurs as well as to the attainment of set criteria (a significant level of differential responding) concerning response magnitude; ( 3 ) the OR magnitude interacts with IS1 to influence conditioning, such that high magnitude orienters tend to condition better but at the longer ISI’s. One interpretation of the ISI-OR-magnitude-discriminabilityrelationship is that some minimal time is needed for young infants to process the stimulus information input before an effective discrimination can be made. Hartman and Grant (1962) have offered a similar interpretation of adult eyelid conditioning data in which the longer ISI’s are more effective in producing discrimination than shorter JSI’s. This requirement for a longer IS1 for differential conditioning implies that operations involved in responding to stimuli differentially occur during the IS1 and have definile temporal requirements. Caldwell and Werboff (1962), using rats, also found that the younger subjects required longer ISI’s for optimal conditioning performance than did older subjects. With respect to the temporal requirements for the procession of information, the superior discrimination performance of the high OR magnitude subject suggests that it utilized more optimally the time available for performing operations necessary for differential responding. The high OR magnitude subjects regardless of IS1 demonstrated a more stable system of responding which was also reflected in the superior performance of the high OR magnitude subjects at the shorter ISI’s during which the probability of a response was not yet high enough to reach significance. If individual differences in discriminability reflect differences in amount of time required for individual subjects to produce and inhibit responses under given instances of stimulation, one might expect to find certain specific differences between high and low OR magnitude subjects. One might expect to find high OR magnitude subjects showing a lower overall spontaneous response frequency during acquisition, but a higher response frequency during acquisition and a higher response frequency to CS+ to CS-. On the other hand, low OR magnitude subjects might show a higher rate of spontaneous responses with either no difference in response frequency to CS+ and CS-, or possibly a higher frequency of response to CS- than CS+. This was indeed the case in this experiment, for low OR magnitude subjects tended to respond with a greater frequency to CS- than to CS+. Zeiner and Schell (1971) reported similar findings with human adults. They found that low OR magnitude subjects had as many as 4-7 times as many failures to respond to the CS+ as did high OK magnitude subjects. Another possible explanation for an OR magnitude-discriminability relationship is that of typological differences as described by Pavlov (1927). Clearly discrimination involves inhibition. Thus, the individual differences in discriminative ability found in this study may simply be different manifestations of general inhibitory ability, and performance differences reported here may be due to factors other than differences in maturational rate or level. Specifically, they may reflect differences in the dynamic

THE IS1 IN INFANT CONDITIONING

425

balance of excitatory and inhibitory processes that are potentially predictive of adult typology (Nebylitsyn, 1972). However, before much of h s can be said with certainty, more systematic research is needed.

Notes This study is based in part on a doctoral dissertation conducted at Michigan State University and was supported in part by National Institute of Mental Health Research Grant No. MH-18655 to Dr. Hiram E. Fitzgerald.

References Brackbill, Y., and Fitzgerald, H. E. (1972). Stereotype temporal conditioning in infants. Psychophysiology, 9 : 569-571. Brackbill, T., Fitzgerald, H. E., and Lintz, L. M. (1967). A developmental study of Classical conditioning. Monogr. SOC.Res Child Dev., 38: All of No. 8. Caldwell, D. F., and Werboff, 5. (1962). Classical Conditioning in newborn rats. Science, 136: 1118-11 1 9. Duncan, D. B. (1951). Multiple range and multiple F tests. Biornetrica, 11: 1-42. Edelberg, R. (1967). Electrical properties of the skin. In C. C. Brown (Ed.), Merhods in Psychophysiology. Baltimore: Williams and Wilkins. Pp. 1-53. Eysenck, H. J., and Levey, A. (1972). Conditioning, introversion-extroversion and the strength of the nervous system. In V. D. Neblitsyn and J. A. Goray (Eds.), 3iological Basis of Individual Behavior. New York: Academic Press. Pp. 165-81. Fitzgerald, H. E., and Brackbill, Y. ( 1 974). Stimulus-response organization, state, and conditionability during early infancy. Psihologzja, 7: 47-56. Fitzgerald, H. E., and Brackbill, Y. (in press). Classical conditioning during infancy: Developmental constraints. Psychol. Bull. Gerald, A. A., and Woodward, 5. K. (1958). Conditioning of the human pupillary response as a function of the CS-UCS interval. J. Exp. PsychoL, 55: 501-507. Gormezano, I., and Moore, 5. W. (1969). Classical conditioning. In M. M. Marx (Ed.), Learning Processes New York: Macmillan, Pp. 348-363. Hartman, T. F., and Grant, D. A. (1962). Effects of pattern of reinforcement and verbal information on acquisition, extinction, and spontaneous recovery of the eyelid CR. J. Exp. PsyChOl., 63: 21 7-226, Hastings, S., and Obrist, P. A. (1967). Heartrate conditioning in humans: Effects of varying the interstimulus interval. J. Exp. Psychol., 74: 431442. Hilgard, E . R., Campbell, R. K., and Sears, W. N. (1938). Conditioning discrimination: The effect of knowledge of stimulus relationships. Am. J. Psychol., 51: 498-506. lngram E., and Fitzgerald, H. E. (1974). Individual differences in infant orienting and autonomic conditioning. Dev. Psychobiol., 7: 359-367. Jensen, W. R., and Prokasy, W. F. (1973). Differential Skin Conductances as a function of interstimulus interval. Bull. Psychonom. Soc., 2: 397-399. Kakigi, S. (1964). Classical conditioning of human pupillary constriction. Jpn. Psychol. Res, 6: 155-16 2. Kdsatkin, N. I. (1972). First conditioning responses and the beginning of the learning process in the human infant. In G. Newton and A. H. Riesen (Eds.), Advances in Psychobiology. (Volume 1.) New York: John Wiley and Sons. Pp. 213-257. Kimmel, H. P., and Pennypacker, H. S. (1963). Differential GSR conditioning as a function of the CS-USC interval. J. Exp. Psychol., 65: 559-563. Lintz, L. hl., Fitzgerald, H. E., and Brackbill, Y. (1967). Conditioning the eye-blink to sound in infants. Psychonom. Sci., 7: 405406.

426

INGRAM

Little, A. H. (1971). Eyelid conditioning in the human infant as a function of the 1st. Paper read at the meeting of the Society for Research in Child Development, Minneapolis, Minnesota. Maltzman, L., and Mandell, M. P. (1968). The orienting reflex as a predictor of performance and learning. J. Exp. Res. Person., 3: 99-106. Morgan, J. J. B., and Morgan, S. S . (1944). Infant learning as a developmental index. J. Genet. Psychol., 65: 281-289. Morgenson, D. F., and Martin, I. (1968). The orienting response as a predictor o f autonomic conditioning. J. Exp. Res. Person., 3: 89-90. Naito, T., and Lipsitt, L. P. (1968). Two attempts to condition eyelid responses in human infants. J. Exp. Child Psychol., 6 : 181-193. Nehylitsyn, V. D. (1972). Basic Properties of the Nervous System in Man. New York: Plenum Press. Ohlrich, E. A., and Ross, L. E. (1968). Acquisition and differential conditioning of the eyelid response in normal and retarded children. J. Exp. Child PsychoL, 6 : 181-193. Pavlov, 1. P. (1927). Conditional Reflexes. London: Oxford University Press. Raskin, D. C., Kotses, H., and Bever, J. (1969). Autonomic indicators of orienting and defensive reflexes. J. Exp. Psychol., 80: 423433. Rendle-Short, J. (1961). The puff test. Arch. Dis. Qzild., 36: 50-57. Sokolov, E. N. (1963). Perception and the Conditioned Reflex. New York: Pergamon Press. Teplov, B. M. (1972). The problem of types of human higher nervous activity anti methods of determining them. In V. D. Nehylitsyn and J. A. Gray (Eds.), BiologicaZ Basis of hdividual Behavior. New York: Academic Press. Pp. 1-10. Volokhov, A. A. (1970). The ontogcnetic development of higher nervous activity in animals. In W. A. Himvich (Ed.), DeveZopmental Neurobiology. Springfield, Illinois: t. C. Thomas. Pp. 5 8 1-6 10. Wenger, M. A . (1936). An investigation of conditioned responses in human infants. Univ. Iowa Stud. Child Welfare, 12: 7-90. Winer, B. J. (1963). Statistical Principles in Experimental Design. New York: Maraw-Hill. Zeiner, A. R., and Schell, A. M. (1971). Individual differences in orienting, conditionability, and skin resistance rcsponsivity. Psychophysiology, 8: 61 2-622.

The interstimulus interval in classical autonomic conditioning of young infants.

The Interstimulus Interval in Classical Autonomic Conditioning of Young Infants EBEN M. INGRAM Department of Psychology University of Colorado at Denv...
517KB Sizes 0 Downloads 0 Views