Symposium
Chair: Daniel S. Messinger
Discussant: Alan Fogel
This symposium showcases new methodological perspectives that are> producing insights into how infants attend to and communicate with their> caregivers. Topics range from the identification of different patterns of> infant interest, through a new event-based perspective on temporal> coordination in infant communication, to the prediction of infant> attachment style from changing patterns of dyadic interaction. Methods> ranging from hierarchical linear modeling to Monte Carlo bootstrapping are> united by a principled approach to finding patterns of consistency in the> actions and development of infants. > > Messinger, Fogel, and Dickson ask why 1- to 6-month-old infants gaze where> they do during face-to-face interaction. Earlier studies found no> association between the duration of one infant gaze at mother's face and> the next, suggesting that infants look toward that part of the environment> that seems most interesting at the moment (an ahistoric process).> Motivated by dynamic systems perspectives, Messinger, et al. use> hierarchical linear modeling to show that individual infants tend to> maintain visual interest for periods of time that span three consecutive> gazes at and away from mother's face. > > Lavelli and Fogel integrate facial expressions and body movements into> attention-related interest expressions involving gazing at mother.> Configural frequency analysis allows them to identify patterns such as> concentrated attention composed of actions (gazing at mother's face with> brows furrowed) which co-occur at greater than chance levels. Hierarchical> linear modeling is then used to distinguish more active forms of> attention, which rise through three months of age, from simply attending> to mother, which rises through two months of age and is then supplanted by> the more active and emotionally positive forms of attention to mother.> > Yale, Messinger, and Cobo-Lewis use a cutting-edge Monte Carlo> bootstrapping approach to identify how infants temporally coordinate> gazing at mother's face, facial expressions, and vocalizations. Infants> tend to temporally embed vocalizations in facial expressions at greater> than chance levels, creating a communicative sandwich in which the> vocalization may highlight the impact of the facial expression which> precedes and follows it. Facial expressions were coordinated not only with> vocalizations but with gazing at mother, but vocalizations and gazing at> mother were not coordinated with each other. The findings give new> empirical support to the view that facial displays of affect are a> central, coordinating topic of infant communication. > > Moore, Cohn, and Campbell expand their attention to the dyad,> calculating not only infant and caregiver positive expressivity, but the> responsivity of each partner to the expressions of the other (proportion> of variance explained by the partner's prior expression). Guided by a> theory-driven image of attachment developing over time, these> characteristics were measured repeatedly in the first six months of life.> Consistent but moderate levels of positive expression and expressive> responsivity on the part of the infant (and mirrored by similar maternal> patterns) predicted secure attachment. The findings suggest the importance> of (consistent) developmental processes in the formation of positive> outcomes. They also reflect a more general theme of the symposium as a> whole: The use of innovative methodological approaches to illuminate> consistent patterns in complex behavioral and developmental processes.>
Details of individual items:
paper
During face-to-face interaction, infants visually attend to (gaze at)mother's face and then attend elsewhere. Infant gazes at mother's face arethought to signal increased engagement in social interaction while infantgazes away from mother's face are thought to modulate infant's arousallevels. Cycles of gazing at (and away from) mother's face are thought to bea precursor of turn-taking in later play and conversation. Previous attemptsto predict how long infants gaze at and away from their mothers' faces havebeen unsuccessful, suggesting an ahistoric process in which one period ofgazing is unrelated to the next.To investigate infant visual attention to mother's face, 13 infants andmothers were observed during 208 weekly, five minute, face-to-faceinteractions in the first six months of life. Gazing at and away from motherwas reliably coded (90% agreement, Kappa .77). Hierarchical linear models(HLM) were used to predict the duration of individual gazes away frommother's face (i.e. time to a gaze at mother) and the duration of individualgazes at mother's face based on the strength of each infant's individualassociations for each infant individually. Log transformations of thedurations of gazes at and away from mother (which were positively skewed)were used in these analyses. In the first six months of life, the duration of infants' gazes at mother'sface became briefer while the duration of gazes away from mothers facebecame more lengthy, ps < .001. This may reflect infant's increasingfamiliarity with mother's face and increasing interest in other features ofthe environment. While controlling for infant age, the duration of aninfant's gaze at mother's face was positively associated with the durationof both the infant's previous gaze at mother's face (lag 1) and the durationof the gaze at mother's face previous to that (lag 2), ps < .001 and .01,respectively. The duration of infant's gazes away from mother's face werepositively associated with the duration of both the infant's previous gazeaway from mother's face (lag 1) and the duration of the gaze away frommother's face previous to that (lag 2), ps < .001 and .025, respectively.The duration of a gaze away from mother's face was not a robust predictor ofthe immediately subsequent gaze at mother's face, and vice-versa. Micro-analytic coding of infant gaze, lag transforms which reduced theimpact of outliers, and hierarchical linear modeling which capitalized onassociations that were robust across infants produced a new perspective oninfant communication in a face-to-face setting. The lag results indicatethat infant gazes at and away from mother's face are not ahistoric events.Instead infants tend to follow a longer gaze with a longer gaze and ashorter gaze with a shorter gaze for up three instances of gazing atmother's face and, relatively independently, three instances of gazing awayfrom mother's face. Sequences of longer and shorter gazes appear to reflectinfants' periods of sustained visual interest in mother's face and,relatively independently, other features of the environment.
paper
Literature on early social interaction has shown the important role> that the infant gaze plays in regulating perceptual input as well as> internal physiological state. This study reports findings on expressive> configurations of facial actions and body movements associated with the> infant gazing at the human face. We also document whether and how these> attention related expressive configurations change during the first three> months of life, i.e., across the key developmental transition at two> months related to the acquisition of exogenous control.> 09> Sixteen primiparous mother-infant dyads were videotaped during a> naturalistic face-to-face interaction at their home weekly, from age 1> to14 weeks. Mother-infant dyads were selected on the basis of no> particular medication during pregnancy and delivery, spontaneous full-term> delivery, normal birth weight, APGAR > 7 at 1 min and > 8 at 5 min,> absence of particular problems on the part of the mothers. Both lower and> middle SES, and different educational levels are represented in the> sample. Each week, dyads were videotaped when the infant was in alert> state, for at least 6 minutes. Infant gaze direction was coded as 'At the> mother's face' and 'Elsewhere'. Using the FACS coding system (Ekman &> Friesen, 1978), we selected three groups of actions: Upper Face Actions,> Lower Face Actions, and Body Actions (motor quieting vs.motor activity).> Microanalysis of the co-occurrences of these actions allowed us to> identify 10 different attention-related expressive configurations shown by> the infants in the context of face-to-face communication with their> mother. The identified expressive configurations were coded by two> independent coders with an intercoder reliability (kappa) of .81. > > Configural Frequency Analysis (von Eye, 1990) was then applied to> assess the significance of co-occurrences between infant gazing direction> and infant expressive configurations. Six expressive configurations,> labeled 'Simple Attention', 'Concentrated Attention' (AU3), 'Excited> Attention' (strong motor activity with AU1+2 and/or AU3), 'Astonished> Attention' (AU1+2 with possible AU5), 'Attentive Smile' (AU12+brow> relaxation), 'Cooing Expression', were found to co-occur significantly> with 'Gazing at the mother's face'. Sequential analysis showed that> 'Concentrated Attention' - identified by brow knitting - significantly> precedeed 'Smile' and 'Cooing Expression'. These findings support the> hypothesis (Oster, 1978; Sroufe, 1979; Sroufe & Waters, 1976) that brow> knitting may reflect emotional and cognitive processes related to an> effort to assimilate the complex, dynamic pattern of stimulation produced> by the mother during face-to-face communication. Developmental analysis> showed a main qualitative change from the dominance of 'Simple Attention'> to the dominance of expressive configurations (particularly 'Concentrated> Attention') that show a more active engagement by the infant, toward the> end of the second month. After 2 months, only 'Smile' and 'Cooing> Expression' continue to increase significantly. This suggests, in> agreement with Trevarthen's observations (1998), that the third month is> characterized in terms of more playful/approaching expressive> configurations, while the second month shows more 'serious'/receptive> expressions.
paper
This study examined the coordination of early infant communicativebehaviors from an event-based perspective. Previous studies have typicallyanalyzed the coordination of infant behaviors with a time-based approach inwhich the unit of coding and analysis is a fixed time interval (e.g., onesecond). In contrast, an event-based approach focuses on the temporalsequencing of whole actions from different communicative modalities thathave some temporal overlap. An event-based approach suggests that when twobehaviors from different modalities temporally overlap, they create acommunicative signal that may be more effective than if each modality wasexpressed alone. Following up recent work showing that 3- and 6-month old infantscoordinated facial expressions and vocalizations at greater than chancelevels, the present explored the coordination of three communicativemodalities using an event-based approach. Forty full-term, typicallydeveloping infants were observed at 3 and 6 months of age while engaged amodified face-to-face/still-face interaction with their mothers. Theinfants' facial expressions, vocalizations and gaze direction were codedand the coordination between (1) vocalizations and facial expressions; (2)facial expressions and gaze direction; and (3) gaze direction andvocalizations were analyzed.Coordinated events in the framework of this study included behaviors fromtwo communicative modalities that had some temporal overlap. Abootstrapping/Markov model was used to determine the expected frequenciesof the various coordinated and non-coordinated events. The bootstrappingprocedure used the actual frequencies, durations and sequential transitionprobabilities of facial expressions, vocalizations and gaze directionseparately in each session as raw data. It then constructed 1,999simulated sessions for each infant using each infants' facial expressions,vocalizations and gaze direction individually. The frequency of each typeof coordinated event expected by chance was then computed. Thebootstrapping procedure yielded z-scores that quantified the frequency ofcoordinated events observed in the data relative to the frequency ofcoordinated events expected by chance.The infants showed a systematic tendency to coordinate actions across thevocal and facial domains and across the facial and gaze domains in that thenumber of observed coordinated events occurred significantly more thanexpected by chance. The infants did not coordinate vocalizations and gazedirection at better than chance levels. Further examination of thecoordinated vocalization-facial expression sequences and the facialexpression-gaze sequences revealed that the infants favored certainsequence patterns over others. For example, in the facial and vocalmodalities, infants tended to embed vocalizations in facial expressions.However, in the facial and gaze modalities, infants tended to gaze at theirmothers' face before producing a facial expression but then look awaybefore ending the facial expression. This study provides clear evidence that infants display coordinatedbehaviors across communicative modalities in the first six months of life,confirms results from other time-based and event-based coordination studiesand applies new methods for the analysis of event-based data toinfant-caregiver interactions.
paper
Prior research on the antecedents of infant attachment has focussed on responsive parenting, usually measured at one point in time prior to the assessment of attachment. This approach has paid insufficient attention to affective characteristics of infants and to aspects of the parent-infant relationship over time. To study the development of parent-infant affective relationships, we examined patterns of infant positive affect and emotion responsiveness toward mother from 2 to 6 months and relations between these developmental profiles and later attachment relationships. In the current research, 107 infants of middle-class, primiparous, married, European-American mothers were observed in face-to-face interaction with their mothers at 2, 4, and 6 months. Infants' and mothers' affective behaviors were coded on a 1-s time base and percentages of positive, neutral, and negative affect were computed. To assess emotion responsiveness to each other, time series regression analyses were used to compute proportions of variation in infant's and mother's emotion expressions that were due to partner's behaviors. At 12 months, infants were observed in the Strange Situation paradigm and classified as securely or insecurely attached.Cluster analysis using only infant affective behaviors during face-to-face interaction from 2 to 6 months distinguished four groups of infants that differed across the three ages in level of positive affect (F(3,103) 3D 24.82, p<.001) and in emotion responsiveness (F(6,206) 3D 16.97, p<.001) to mother. Mothers of infants in these groups followed developmental profiles that were similar to the infants. At 12 months, the four groups differed in patterns of attachment (X^2(3) 3D 11.19, p<.05). Infants who were most positive and most responsive at 2 and 4 months, but whose positive responsiveness decreased to the level of other infants by 6 months, were most likely to be classified as insecurely attached. Their mothers showed a similar profile of high positive responsiveness at 2 and 4 months with a decline to the level of other mothers by 6 months. In contrast, infants who were moderately positive and moderately responsive relative to other infants, were most likely to be classified as securely attached. Mothers of these infants also showed moderate and very consistent levels of positive responsiveness from 2 to 6 months. Initial high levels of positive responsiveness at 2 and 4 months in infants and mothers were related to a decline in these measures by 6 months and to later insecure attachment. Consistent with some earlier research, these results suggest that high levels of emotion responsiveness may be experienced as intrusive by infants. Results also suggest that consistency of affective behavior is an important predictor of secure attachment relationships. Thus, we emphasize attachment as an index of the affective relationship between parent and child and reinforce the importance of examining attachment as a developmental process between infants and their parents.