Abstract

Abstract: Four-month-old infants were screened (N = 433) for temperamental patterns thought to predict behavioral inhibition, including motor reactivity and the expression of negative affect. Those selected (N = 153) were assessed at multiple age points across the first 4 years of life for behavioral signs of inhibition as well as psychophysiological markers of frontal electroencephalogram (EEG) asymmetry. Four-month temperament was modestly predictive of behavioral inhibition over the first 2 years of life and of behavioral reticence at age 4. Those infants who remained continuously inhibited displayed right frontal EEG asymmetry as early as 9 months of age while those who changed from inhibited to noninhibited did not. Change in behavioral inhibition was related to experience of nonparental care. A second group of infants, selected at 4 months of age for patterns of behavior thought to predict temperamental exuberance, displayed a high degree of continuity over time in these behaviors. Article: INTRODUCTION There are multiple reports in the research literature of the behavioral and physiological correlates of the temperamental pattern known as behavioral inhibition. Inhibited toddlers and preschool children are characterized as displaying vigilant behaviors and motor quieting when confronted with novelty. They are unlikely to approach unfamiliar adults In addition, there have been a number of studies that have described the antecedents and developmental trajectories of behaviorally inhibited children. For example, In an attempt to examine infant predictors of behavioral inhibition, Kagan and colleagues selected infants at 4 months of age who displayed either high or low levels of motor arousal and negative affect in response to a series of novel visual and auditory stimuli One of the issues in the study of the continuity of inhibition is the characterization of its form across development. In his Handbook chapter, Caspi (1998) describes five different types of continuity of personality to be considered in the study of developmental change. Four of these (differential, absolute, structural, ipsative) refer to homotypic continuity, which is continuity of similar behaviors or phenotypic attributes over time. The fifth, heterotypic continuity, refers to continuity of an inferred genotypic attribute presumed to underlie diverse phenotypic behaviors. It is this latter type of continuity that characterizes developmental change in behavioral inhibition. The term behavioral inhibition is used to describe temperamental differences in infants' and young children's initial reactions to a range of novel stimuli including people, objects, contexts, and challenging situations Fox and Rubin have used the term social reticence to reflect behavioral inhibition in social situations. They and their colleagues have conducted detailed observations of children's behaviors during free play, speech-making, and cooperative tasks with unfamiliar peers in order to study social reticence (e.g., A number of studies using parent report measures of temperament to identify inhibited children have also found moderate continuity in this trait. In each of the aforementioned studies of inhibition, the emphasis has been on continuity. Yet given the modest correlations reported, there are clearly many children who do not remain inhibited across the early years of life. We know little about these children or the factors that may contribute to the changes in their behavior over time. One factor that may contribute to the continuity of inhibition over time is the physiological disposition of infants to express negative affect and withdrawal in response to novelty. A variety of data suggests that the pattern of the resting electroencephalogram (EEG) may reflect an individual bias to respond with positive or negative affect to stressful situations. Davidson and colleagues In addition to being related to specific affective states, Fox and his colleagues have found relations between EEG asymmetry and shyness in adults and reticence in children. Specifically, adults who rated themselves as high in shyness displayed right frontal EEG asymmetry The following study reports on two cohorts of infants, some of whom were selected at four months of age for characteristics thought to predict behavioral inhibition. These infants were high in negative reactivity. A second group of infants who were high in positive reactivity, as well as a low reactive group, were also recruited. All infants were seen at 9, 14, 24, and 48 months of age in the laboratory. Measures of brain electrical activity were recorded at each of these ages and behavioral observations were made at 14, 24, and 48 months. At 14 and 24 months, children were seen in a standard behavioral inhibition paradigm. At 48 months of age, children were observed in same-gender, unfamiliar peer quartets, and measures of reticence and social play were coded. In this paper, we examined the degree to which infant reactivity patterns predicted later behavioral inhibition and reticence. We also compared the degree of continuity in inhibition based on patterns of infant reactivity. In addition, we examined relations among observed behavior and EEG to determine if frontal EEG asymmetry was related to patterns of continuity in behavioral inhibition. The group of infants displaying high motor reactivity and high positive affect in response to novel stimuli were unlike Kagan's low reactive group in that these infants displayed high motor reactivity but positive affect in their responses. We predicted that this group would present with its own unique developmental trajectory and pattern of physiological response including lack of fear or inhibition, high sociability, and left frontal EEG asymmetry. It is important to distinguish this group from noninhibited children who may simply show less fear or inhibition in response to challenge. Indeed, we expected that exuberant, uninhibited children would be different from noninhibited controls. The data in the current paper reflect two cohorts of children. Children in each cohort were selected at 4 months of age and each cohort was seen at the identical age points through age 4. Data on the patterns of EEG activity at 9 months and behavioral inhibition at 14 months of age for the first cohort have previously been reported in METHOD Participants To select infants thought likely to display behavioral inhibition later in infancy and early childhood, a total of 433 4-month-old infants were screened for motor reactivity and emotional reactivity in response to novel sights and sounds. The infants recruited came from two separate cohorts with identical screening procedures. Details of the screening procedure have been described previously Families with young infants were initially contacted by mail using commercially available lists of names and addresses compiled from the birth records of area hospitals. Interested parents were asked to complete a brief background survey. Families were excluded from further participation if any of the following were true: (1) the infant was preterm (less than 36 weeks gestation), (2) the infant had experienced any serious illnesses or problems in development since birth, (3) the infant was on any long-term medi cation, or (4) either of the parents were left-handed. Families who did not meet any of these exclusion criteria were contacted by phone and given more information regarding the study. Home visits were scheduled for interested families (N = 433). Home visits took place when the infant was 4 months of age (±14 days). While in an infant seat and in a quiet and alert state, the infant was presented with two sets of novel visual (brightly colored mobiles) and auditory (taped sentences and nonsense syllables) stimuli. Each set of stimuli consisted of a series of visual presentations followed by a series of auditory presentations. The first series of visual stimuli consisted of three mobiles differing in the number of hanging characters (1, 3, or 6 characters). Each mobile was presented for 20 s, and the presentations were separated by a 10-s intertrial interval. The series of three mobiles was repeated three times for a total of nine trials. Each mobile was displayed at the infant's eye level and approximately 12 inches from the face. The first series of auditory stimuli consisted of 8 short sentences. Each sentence was approximately 6 s in duration, followed by a 2-s intertrial interval. The sentences were presented in pairs. The pairs differed in the number of voices speaking and, as a result, the volume of the presentation. The first pair was spoken by a single voice, the second pair by two voices, the third pair by three voices, and the fourth pair by four voices. The second set of novel stimuli was similar to the first except that the characters on the mobiles were different, and the auditory stimuli were nonsense syllables (ma, ga, pa) rather than sentences. The series of three mobiles was presented in exactly the same fashion as in the first set, for a total of nine trials. Each nonsense syllable was presented in three consecutive 10-s trials, with 5-s intertrial intervals. The stimuli were presented to the infants in an identical order (mobiles 1, auditory 1, mobiles 2, auditory 2). Infants who began to cry during an episode were allowed to cry for a continuous period of no more than 20 s after which the mother was asked to intervene and calm her infant. Once sufficient calm was restored, the session was continued. If an infant was unable to continue with a session, scores were prorated for the amount of time (number of episodes) that the infant missed. All sessions were videotaped, allowing for the later coding of infant reactivity. Infants from both cohorts were selected based on the amount of motor reactivity as well as positive and negative affect expressed during the presentation of the novel sights and sounds. The methods of coding and quantifying reactivity varied slightly between the two cohorts. In the first cohort, coding was based on the procedures described previously by Infants who were extreme on the dimensions of motor activity and affect were selected for participation in the study. The criteria for selection were established based on the reactions of the first 25% of the screened sample. Three groups were selected: (1) those above the mean on motor activity and negative affect (High Negative), (2) those above the mean on motor activity and positive affect (High Positive), and (3) those below the mean on motor activity, positive, and negative affect (Low Reactive). Of the 208 infants screened in the first cohort, 29 were identified as High Negative, 22 as High Positive, and 30 as Low Reactive. Consistent with the selection criteria, a MANOVA comparing the three temperament groups on the three reactivity dimensions was significant, p < .001. Specifically, the High Negative group had significantly higher negative affect scores than both the High Positive and Low Reactive groups, F(2, 78) = 19.24, p <.001; Tukey's HSD, both ps <.001. The High Positive group had significantly higher positive affect scores compared to both the High Negative and the Low Reactive groups, F(2, 78) = 37.40, p < .001; Tukey's HSD, both ps < .001. The Low Reactive group had significantly lower motor activity scores than both the High Negative and the High Positive groups, F(2, 78) = 11.38, p < .001; Tukey's HSD, p < .001 and p < .01, respectively. In the second cohort, coders used 7-point Likert-type scales to rate the infant's motor, positive, and negative reactions to each of the visual and auditory presentations. Thus, each infant had a total of three ratings on each session (visual 1, auditory 1, visual 2, auditory 2) for a total of 12 ratings. On the Motor Scale, a score of 7 indicated intense gross motor activity including back arching, body twisting, and hyperextensions of the arms and legs. A score of 1 indicated very little or no motor activity. A score of 7 on the Positive Affect Scale was used to describe infants who responded with many positive vocalizations, and many instances of gurgling, cooing, and smiling (at either the stimulus or the experimenter). Neutral vocalizations were also considered in scoring positive affect; however, neutral vocalizations on their own were not sufficient for a high score on the Positive Affect Scale. A score of 1 on the Positive Affect Scale was reserved for infants who did not smile at all or make more than two neutral or low-intensity positive vocalizations during the presentation. On the Negative Affect Scale, a score of 7 indicated a high degree of intense negative affect and described infants who cried or fussed continuously during the majority of the stimulus presentations. A 1 on the scale indicated an absence of negative affect across the different stimulus events. Three coders rated the tapes. Estimates of inter-rater reliability were computed for pairs of coders, based on 20% of the sample. Pearson correlations ranged from .62 to .95, with a mean correlation of .68 on the motor scale, .69 on the Positive Affect Scale, and .95 on the Negative Affect Scale. Classifications into each of the three temperament groups (High Negative, High Positive, Low Reactive) were made based on the frequency of -high‖ scores across the four stimulus conditions. To be classified as High Negative, an infant had to receive a score of 4 or more on Motor Activity for both the visual and auditory presentations in either the first or second set of presentations. In addition, the infant had to have a score of 4 or more for Negative Affect during the same set of stimuli presentations. To be classified as High Positive, an infant had to have a score of 4 or more on Motor Activity for both the visual and auditory presentations in either the first or second set, as well as a score of 4 or more on Positive Affect during the same presentations. The Low Reactive group was comprised of infants who had scores of 3 or less on Motor Reactivity, Positive Affect, and Negative Affect during both the first and second sets of presentations. Also, an infant had to have seven scores (out of a possible 12) of 1 to be identified as low reactive. Of the 225 infants screened in the second cohort, 27 were classified as High Negative, 23 as High Positive, and 22 as Low Reactive. Consistent with the selection criteria, these three groups differed significantly on each of the three reactivity dimensions. Specifically, the High Negative group was rated as significantly higher in negative affect compared to the other two groups, F(2, 67) = 63.29, p < .001; Tukey's HSD, both ps < .001. The High Positive group was scored as having significantly more positive affect than the other two groups, F(2, 68) = 23.83, p < .001; Tukey's HSD, both ps < .001. The Low Reactive group was rated as displaying significantly less motor activity than the other two groups, F(2, 69) = 23.83, p < .001; Tukey's HSD, both ps < .001. To examine the comparability of the two selection systems, we double coded the cohort originally selected using the 7-point scale system utilized by As a further check, we examined whether the temperament groups within the two cohorts differed on any of the major outcome measures (behavioral inhibition, social reticence, EEG asymmetry). They did not. Procedure Following the initial selection, families were invited to the laboratory when their infants were 9, 14, 24, and 48 months of age. When the infants were 9 months of age, brain electrical activity (EEG) was recorded and mothers completed a temperament survey. At each of the subsequent ages, EEG was recorded, mothers completed age-appropriate temperament surveys, and children's reactions to unfamiliar stimuli were observed in the laboratory. Measures EEG data collection. Several procedures were used during EEG data collection to help maximize the chances that the infants would sit quietly and maintain a steady, attentive state. At 9, 14, and 24 months, infants were seated in front of a table. At 9 months, infants sat on their mothers' laps to help minimize fussing and movement, and at 14 and 24 months infants sat alone in an infant seat with their mothers just off to one side. At each of these ages, a metal bingo wheel was placed on the table directly in front of the infant. An experimenter placed different numbers of brightly colored balls (1, 3, or 7) in the wheel and spun the wheel for a series of six trials each lasting 20 s. These trials were separated by 10-s intervals in which the experimenter tapped the balls on the outside of the bingo wheel to keep the infant's attention between trials. EEG was recorded for the entire 3-min period. When the children were 4 years of age, the EEG collection room was decorated to resemble a space shuttle (see Prior to the recording of EEG from each subject, a 50 µV 10 Hz signal was input into each of the channels and this amplified signal was recorded. This signal of known frequency and amplitude was later used for calibration purposes. At each age, the experimenter began preparing for EEG collection by measuring the circumference of the child's head in order to select a Lycra stretch cap of appropriate size. The stretch caps have electrodes for EEG recording sewn into the fabric according to the 10-20 system of electrode placement A small amount of abrasive gel (Omni-prep) was inserted into each of the six active sites on the cap (F3, F4, P3, P4, O1, and O2) as well as the reference site at the vertex (CZ). Using the blunt end of a Q-tip, each site was gently abraded. Following abrasion, a small amount of electrolyte gel was inserted in each site. The blunt end of a Q-tip was again used to ensure that the gel was making contact with the scalp in the area below each electrode site. Impedances were measured at each site and were considered acceptable if they were at or below 5 K ohms. One channel of EOG was recorded from the right eye using two Beckman mini-electrodes, one placed at the outer canthus and the second placed at the supra orbit position. The EEG and EOG were amplified by separate Grass AC bio-amplifiers (7p511) with the high pass setting at 1 Hz and the low pass at 100 Hz. The data were digitized online using a HEM A/D board and acquisition software. The digitized data were stored for later analyses. During EEG recording, a second experimenter pressed a button switch at the onset and offset of each stimulus condition. The output of the button switch went to one A/D channel, and was used to synchronize the stimulus times with the EEG. EEG data reduction. The EEG data were digitized at a rate of 512 Hz. The EEG data were then re-referenced via software so that the data could be analyzed with an average reference configuration. The digitized EEG data were then displayed graphically for artifact scoring. Portions of the EEG record marked by eye movement or motor movement artifact were removed from all channels of the EEG record prior to subsequent analysis. The re-referenced, artifact-scored EEG data were submitted to a discrete Fourier transform analysis that utilized a Hanning window with 50% overlap. The result of this analysis was to produce power in picowatt ohms (or microvolts squared) for each channel, for each of the stimulus conditions. Spectral power data in single Hz frequency bins from 1 to 12 Hz were computed for each of the stimulus conditions at each of the collection sites. At 9 months of age, power in the 4-6 Hz frequency band was computed for each site by summing the single hertz bins in these three frequencies for each of the stimulus conditions (1, 3, and 7 balls). At 14 and 24 months of age, power was computed in the 6-8 Hz band for each site by summing the single hertz bins in these three frequencies for each of the stimulus conditions (1, 3, and 7 balls). At 4 years of age, power in the 6-8 Hz band was computed separately for the eyes-closed and eyesopened conditions. Again, power in the 6-8 Hz band was computed by summing the single hertz bins in the three frequencies. The use of different frequency bands at the different ages reflects observed developmental changes in the EEG. Specifically, at 9 months of age the majority of power was localized to the 4-6 Hz frequency range, while at older ages there was a clear shift in the locus of power to 6-8 Hz. Previous research with human infants has documented this shift in spectral power across the first years of life Observed behavioral inhibition. At 14 and 24 months, the infant's reactions to unfamiliar stimuli in the laboratory were coded to provide an index of behavioral inhibition (see At the beginning of the visit, the infant and mother entered an unfamiliar playroom with some toys on the floor. Mothers were instructed to work on questionnaires and to let infants play on their own. They were told to respond as they normally would if the infants attempted to get their attention. The free-play period lasted 5 min. Following the free-play session, the toys were removed from the room and an unfamiliar female research assistant entered the room with a toy dump truck and some blocks. The stranger sat quietly for 1 min, played with the truck for 1 min, and then (if the child had not yet approached) invited the child to join her in play for 1 min. After the third minute, the stranger took the truck out of the room and returned with an electronic robot. The battery-operated robot was approximately 18 inches in height, had flashing lights, made loud noises, emitted smoke, and moved around the room. The research assistant left the robot running in the room for 2 min. At 24 months, the observations continued when the experimenter returned to the room with an inflatable tunnel that she encouraged the child to crawl through. After she left, another female experimenter entered the room dressed as a clown. The clown was silent for 30 s, invited the child to approach for 1 min, and then removed enough of her costume for the child to realize that she was another experimenter whom he or she had met before. At 14 months, an index of inhibition was computed based on the infant's reactions to these unfamiliar stimuli. Standardized scores of the following measures were summed to create a single summary index of inhibition: (1) latency to first touch a toy during free-play, (2) latency to vocalize during free-play, (3) time spent in proximity (within arms length) to mother during freeplay, (4) latency to vocalize to the stranger, (5) latency to approach the stranger, (6) time spent in proximity to mother while the stranger presented the truck, (7) latency to vocalize to the robot, (8) latency to approach the robot, (9) time spent in proximity to mother during the robot episode. The summed index was standardized and scores on the index of inhibition ranged from -1.85 to 3.00. Intercoder reliability was computed for 15% of the sample using percent agreement given that all measures were based on recordings of time. Pearson correlations between pairs of coders on the individual measures ranged from .85 to 1.00. At 24 months, a single composite measure of inhibition was computed by summing standardized scores on (1) time spent in proximity (within arms length) to mother during free-play, (2) time spent in proximity to mother during the truck episode, (3) time spent in proximity to mother during the robot episode, (4) time spent in proximity to mother during the tunnel episode, (5) latency to approach the stranger and/or touch the truck, (6) latency to approach and/ or touch the robot, and (7) latency to pass through the tunnel. Intercoder reliability was computed for 24% of the sample using percent agreement given that all measures were based on recording of time. Pearson correlations between pairs of coders on the individual measures ranged from .77 to .97. The summed index of inhibition was standardized and scores on the index of inhibition ranged from -2.30 to 2.56.