Overview
June, 2010
Infants are intensely emotional. They move into and out of states of desperate crying and joyous exuberance. Much of infant emotional life takes place during interactions with their caregivers. I investigate infants’ development of fundamental emotional and social competencies, and their impact on a variety of outcomes. My theoretical orientation is dynamic systems theory (Messinger & Fogel, 2007; Messinger, Fogel, & Dickson, 1997). I am specifically interested in identifying the emergence of basic competencies in ongoing interactions that enable later developmental achievements. This work is conducted from a developmental psychopathology perspective. I ask how typical developmental processes are disrupted in psychopathology, and then address what these disruptions reveal about typical development. An example of both the dynamic systems and the developmental psychopathology orientation concerns infants at risk for autism. Among these infants, early flexibility in allocating attention between social and nonsocial foci during interaction appears to predict later levels of referential joint attention.
My work involves scientific collaborations with experts on behavioral development, measurement technologies, environmental risk, and autism. It also reflects my mentorship and ongoing collaborations with multiple graduate students and postdoctoral associates. My research has been well-supported with multiple grants from both the National Institutes of Health and the National Science Foundation. Additional support has come from the University of Miami and Autism Speaks. I am the director of the Miami Marino Autism Research Institute, which administers seed grants to junior investigators, and a member of the NIH/Autism Speaks Baby Sibs Research Consortium, which coordinates studies of infants at risk for autism. I am associate editor of the journal Emotion, was a member of the editorial board of Infancy, and am a frequent ad hoc reviewer for the NIH, the NSF, and Autism Speaks. I am committed to active dissemination of new research technologies and scientific findings (see http://measurement.psy.miami.edu). My research has received coverage in both the scientific and popular press, including write-ups in nationally distributed parenting magazines and local television coverage.
Research
There are five strands to my research on early social and emotional development. The first strand of research focuses on the dynamics of infant expressions of positive and negative emotion; the second investigates the development of communicative responsivity during interaction. The third strand of research involves the development of referential communication, and the referential sharing of positive affect. These strands of research are infused with an interest in risk guided by a developmental psychopathology perspective. My fourth strand of research concerns the development of autism symptomatology in at-risk infants, while the fifth strand is related to risk and resilience among low birth weight infants and infants prenatally exposed to cocaine. Future directions involve interfacing precise behavioral measurements with genetically-determined risk indices, and using neurophysiological measurements and genetics to better understand healthy and disturbed developmental processes.
Early Emotion Dynamics
The first strand of research examines infant expressive behaviors during interaction to understand emotions and their development. My student’s research indicated that infant facial expressions were temporally coordinated with gazes at the parent and with infant vocalizations (Yale, Messinger, & Cobo-Lewis, 2003; Yale, Messinger, Cobo-Lewis, Oller, & Eilers, 1999). The centrality of infant facial expressions to communication provided an empirical basis for studying their meaning and development. My colleagues and I described the emergence of surprisingly mature smiling among neonates and traced the social development of specific forms of smiling during peak periods of dyadic interaction (Dondi, et al., 2006; Messinger, et al., 2002; Messinger, Fogel, & Dickson, 1999, 2001). We next asked ourselves whether infant facial expressions were qualitatively distinct, or whether they varied along a continuum of affective intensity (Messinger & Fogel, 2007).
Initial work suggested qualitative differences between different forms of smiling. The Duchenne smile involving eye constriction appears, for example, to be specifically involved in the reciprocation of positive affect in infants (Messinger, et al., 2001). However, contextual situations that elicited smiles with eye constriction also elicited smiling without eye constriction, though to a lesser degree. This suggests that although all infant smiling indexes positive emotion, specific forms of smiling are more positive than others. Rating studies supported this view of smiling as communicating gradations in positive emotion. Observers rated all smiles as emotionally positive, but rated smiles involving eye constriction and mouth opening as more positive than smiles without these features (Bolzani-Dinehart, et al., 2005; Messinger, Cassel, Acosta, Ambadar, & Cohn, 2008).
These rating studies also revealed an interesting similarity between infants’ smiles and negative cry-face expressions. When smiles were accompanied by actions such as eye constriction, they were perceived as more emotionally positive; when cry-faces were accompanied by a greater degree of eye constriction, they were perceived as more emotionally negative. The finding that the same facial actions are associated with both intense positive and with negative expressions suggests that the form of facial expressions is not entirely arbitrary, but instead has a certain grammar. Illustrating a central systems principle, when an 'intensifying' facial action like eye constriction (which has little intrinsic emotional meaning) occurs with a smile or a frown, the emotional significance of the expression considered as a whole is greater than the sum of its parts (Messinger, 2002; Messinger, et al., 1997).
These investigations of emotional expression prompted the development of novel measurement approaches that, in turn, offered new insight into infant emotion and communication. In collaboration with Jeff Cohn, I conducted the first automated measurements of the intensity of anatomically–based facial expression in dyadic interaction (Messinger, Mahoor, Chow, & Cohn, 2009). These measurements indicated a single dimension of positive emotional expression that was simultaneously indexed by the intensity of smiling and the intensity of eye constriction. This result challenged adult-oriented theories which dichotomize smiles as emotional or non-emotional, suggesting instead that smiling, and perhaps positive emotion, is a question of degree,
Our automated measurements—involving active appearance modeling and machine learning—showed good reliability with human coders. We used a novel approach to determine that the automated measurements were also directly related to perceived emotion. Non-expert observers were asked to make continuous ratings of affective valence using a joystick interface. The premise of this approach is that human beings are expert observers of other human beings. Armed with a simple description of the construct of interest, raters produce a continuous stream of observations that are aggregated to create a time-series of perceived qualities such as emotional valence, parental sensitive structuring, and familial negativity. I next applied the continuous ratings and automated objective measurements to understanding early interaction (Messinger, et al., in press).
Infant-Parent Interaction
I am concerned with the development of communicative influence during interaction in the first six months of life. I employ both naturalistic face-to-face interactions and the still-face procedure. Measurements include manual coding of infant and parent expressive behavior, as well as the automated coding and continuous ratings described above. Data-analytic approaches range from chi-squares of co-occurrences to sophisticated time-series modeling of multiple dyads to machine learning modeling of information flow between parent and mother. An invited article, for example, documents provides the first documentation that there is a developmental increase in infant-mother turn-taking in the realm of smiling (Messinger, Ruvolo, Ekas, & Fogel, under review). Another study formally demonstrated that infant-to-parent emotional influence was more central to interaction structure than was parent-to-infant influence (Chow, Haltigan, & Messinger, 2010).
The aim of this work is to uncover the basic structure of early social interaction and understand its development. Our automated analyses of infant and mother facial expression produced a detailed portrait of ebbs and flows in joint smiling intensity (Messinger, et al., 2009). In fact, changes in dyadic synchrony—defined as local periods of high correlation—were evident within the interaction of individual dyads. Following up on this finding, we used continuous ratings of affective valence to formally document the existence of changing infant-mother emotional dynamics during the course of interaction (Chow, et al., 2010). This suggests that parent and infant vary over time in their emotional responsivity to one another, a plausible but unexplored feature of (early) interaction.
The Development of Referential Communication
Through six months of age, infants engage in complex sequences of communicative behavior but there is little evidence that they intentionally communicate. My work in the development of referential communication is aimed at revealing how infants move from expecting a partner’s action to intentionally communicating in order to elicit that action. I examine this transition both in the infant’s interaction with a parent, and in the examiner-administered Early Social Communication Scales.
Beginning at about eight months, infants initiate joint attention by using gaze and gesture to share an experience of an event or object. My early work indicated that initiating joint attention was an approach-based behavior frequently associated with smiling (Messinger & Fogel, 1998). Adamson and Bakeman had hypothesized that positive affect might motivate infants to coordinate joint attention with an adult. Armed with these insights, I distinguished sequences in which infants gaze at an interesting toy and then gaze at the examiner who created the spectacle, as if sharing the smile. Anticipatory smiles emerged between 8 and 10 months, and seem to be an early communication relating the examiner and toy. Additional research indicated that early levels of smiling in interaction with mother are associated with later levels of anticipatory smiling with an examiner, suggesting that high levels of positive affect in interaction enable the sharing of that affect in a referential situation. In this sense, positive affect appeared to motivate the infant’s referential communication in real time and in developmental time. Moreover, anticipatory smiling appeared to index an early pro-social orientation that was manifested in parent-reported social competence at two and a half years of age.
Autism Risk
The study of autism risk has been central to my investigation of interaction and referential communication. The infant siblings of children with Autism Spectrum Disorder (ASD-sibs) are at risk for disruptions of early emotion and communication that may be related to later social and behavioral symptomatology. ASD-Sibs present a broad continuum of behavior and interaction that encompasses typical, subtly disturbed, and diagnosable levels of social functioning. Autism-related research has integrated my work into a fast-paced scientific community, and deepened my understanding of early expressivity and the development of communication.
My research has demonstrated that infant siblings of children with ASDs exhibited a dual pattern of lower levels of rated positivity in the still-face combined with lower levels of emotional fluctuation during and after the still-face. ASD-sibs also exhibit difficulties flexibly shifting attention between social and nonsocial foci over the course of the face-to-face/still-face procedure (Ibanez, Messinger, Newell, Sheskin, & Lambert, 2008). In continuing work, I will follow up these results with an expanded array of measurement approaches and larger samples to relate early differences among at-risk infants to later communication difficulties.
Initiating joint attention is a deficit in children with Autism Spectrum Disorders and is thought to be a critical risk area for their younger siblings. In initial work, we found that ASD-Sibs exhibited subtle deficits both in requesting objects from an examiner and in initiating joint attention. Understanding the balance of these behaviors and their associated functions—requesting an object versus sharing attention to an object—remains a theme of my work in understanding the development of autism symptomatology and its treatment. Work under review indicates that among these at-risk infants, flexibility of attention at 6 months predicts initiating joint attention at 18 months, suggesting a pathway to communicative deficits for these children.
My thinking about the development of ASDs is guided by constructivist models such as Dawson’s, which suggest that environmental factors may impact the development of symptom severity. We found support for this premise examining mothers’ sensitive structuring of their 15-month-olds’ engagement with the environment during a free-play interaction. Greater sensitive structuring was associated with increases in expressive language between two and three years of age (Baker, Messinger, Lyons, & Grantz, 2010). These findings suggest promising responsivity to interaction, even among children developing an ASD. I am following up these observations by collaborating on an intervention study to assess the effects of promoting parental sensitive structuring for young children with ASD symptoms. Findings from a randomized control trial currently under review indicates improvements in parental responsivity to child communication and moderated effects on child communication that were dependent on initial child object interest.
Perinatal Risk and Development
My fifth strand of research concerns the developmental trajectories of children facing risks such as low birth weight (Messinger, Dolcourt, King, Bodnar, & Beck, 1996; Messinger, Lambert, et al., in press) and prenatal drug exposure. This work builds on my participation in a multi-site study of some 1,200 infants, half of whom were prenatally exposed to cocaine and/or opiates. My colleagues and I discovered that the mothers of exposed infants tended to be somewhat less engaged with their one-month-olds during feeding interactions and less engaged with their four-month-olds infants during the still-face procedure (LaGasse, et al., 2003; Tronick, et al., 2005). Infant social and emotional differences associated with cocaine exposure were, however, sparse between 1 and 18 months (Brunner, Messinger, & Bauer, 2005).
In a similar fashion, prenatal cocaine exposure was not associated with lower levels of mental and motor functioning after accounting for indices of poverty. Instead, poverty, a more prevalent risk factor, was associated with deficits in mental functioning that became more pronounced as children grew older. Current work with a low birth weight sample indicates that behavior ratings by an examiner administering a standardized assessment at 18 months predict mental and psychomotor performance at 30 months (Messinger, Lambert, et al., in press). In other words, infant behavior during the assessment provides clues to future performance above and beyond previous performance. This finding has clear clinical implications in the assessment of at-risk infants in community settings, and builds on my interest in using behavioral engagement to predict developmental outcomes. Uniting this work with at-risk infants and my investigations of social and emotional functioning is a concern with change processes rooted in my background in dynamic systems theory.
Future Directions
In broad strokes, I have two future directions. First, I will continue to adapt and employ robust automated measurement systems that can provide efficient, continuous descriptions of facial, vocal, and gestural communication. As technologies improve, automated measurement will become a fertile ground for testing hypotheses concerning interaction and development. I am also beginning to use machine learning approaches to model development. A recent example stems from an interdisciplinary project funded by NSF. We plan to test Vygotsky’s hypothesis that responses to infant reaching are required for the development of requesting. Results will both provide insights into infant development and inform the developmental programming of an infant robot.
Second, I am expanding my understanding of social and emotional functioning in funded collaborations investigating neural functioning and genetics. These investigations involve research on the neural correlates of behavioral measures of approach activation and smile perception utilizing free-running EEG and event-related potentials. I am also broadening my understanding of behavioral development by investigating genetic indices of autism risk and genetic indices of the functioning of neurotransmitter systems. I intend to pursue both future directions from a developmental psychopathology perspective, focused most immediately on the development of autism. It is my hope that weaving together broader neural and genetic strands of research with deeper behavioral strands will provide a more complete understanding of how development occurs, goes awry, and can be facilitated.
Baker, J., Messinger, D., Lyons, K., & Grantz, C. (2010). A Pilot Study of Maternal Sensitivity in the Context of Emergent Autism. Journal of Autism and Developmental Disorders.
Bolzani-Dinehart, L., Messinger, D. S., Acosta, S., Cassel, T., Ambadar, Z., & Cohn, J. (2005). Adult perceptions of positive and negative infant emotional expressions. Infancy, 8(3), 279–303.
Brunner, S. M., Messinger, D. S., & Bauer, C. (2005). Cocaine exposure and mother-toddler social play. Infant Behavior & Development, 28, 62-73.
Chow, S., Haltigan, J. D., & Messinger, D. S. (2010). Dynamic Affect Coupling between Infants and Parents during Face-to-Face and Still-Face Paradigm: Inter- and Intra-Dyad Differences Emotion, 10, 101-114.
Dondi, M., Messinger, D., Colle, M., Tabasso, A., Simion, F., & Fogel, A. (2006). A new look at neonatal smiling: Differences between the judgments of expert coders and naïve observers. under review.
Ibanez, L., Messinger, D., Newell, L., Sheskin, M., & Lambert, B. (2008). Visual disengagement in the infant siblings of children with an Autism Spectrum Disorder (ASD). Autism: International Journal of Research and Practice, 12, 523-535.
LaGasse, L., Messinger, D., Lester, B. M., Seifer, R., Bauer, C. R., Shankaran, S., Bada, H., et al. (2003). Prenatal drug exposure and maternal and infant feeding behavior. Archives of Disease in Childhood: Fetal Neonatal Edition, 88, F391–F399.
Messinger, D. (2002). Positive and negative: Infant facial expressions and emotions. Current Directions in Psychological Science, 11(1), 1-6.
Messinger, D., Cassel, T., Acosta, S., Ambadar, Z., & Cohn, J. (2008). Infant Smiling Dynamics and Perceived Positive Emotion. Journal of Nonverbal Behavior, 32, 133-155.
Messinger, D., Dolcourt, J., King, J., Bodnar, A., & Beck, D. (1996). The survival and developmental outcome of extremely low birthweight infants. Infant Mental Health Journal, 17(4), 375-385.
Messinger, D., Dondi, M., Nelson-Goens, G. C., Beghi, A., Fogel, A., & Simion, F. (2002). How sleeping neonates smile. Developmental Science, 5(1), 48-54.
Messinger, D., & Fogel, A. (1998). Give and take: The development of conventional infant gestures. Merrill-Palmer Quarterly, 44(4), 566-590.
Messinger, D., & Fogel, A. (2007). The interactive development of social smiling. In R. Kail (Ed.), Advances in Child Development and Behavior, 35 (pp. 328-366). Oxford: Elsevier.
Messinger, D., Fogel, A., & Dickson, K. (1999). What's in a smile? Developmental Psychology, 35(3), 701-708.
Messinger, D., Fogel, A., & Dickson, K. (2001). All smiles are positive, but some smiles are more positive than others. Developmental Psychology, 37(5), 642-653.
Messinger, D., Fogel, A., & Dickson, K. L. (1997). A dynamic systems approach to infant facial action. In J. A. Russell & F. M. Dols (Eds.), The psychology of facial expression (pp. 205-226). New York: Cambridge University Press.
Messinger, D., Lambert, B., Bauer, C., Bann, C., Hamlin-Smith, K., & Das, A. (in press). The Relationship between Behavior Ratings and Concurrent and Subsequent Mental and Motor Performance in Toddlers Born at Extremely Low Birth Weight. Journal of Early Intervention.
Messinger, D., Mahoor, M., Chow, S., & Cohn, J. F. (2009). Automated Measurement of Facial Expression in Infant-Mother Interaction: A Pilot Study. Infancy, 14, 285-305. NIHMS99269.
Messinger, D., Mahoor, M., Chow, S., Haltigan, J. D., Cadavid, S., & Cohn, J. F. (in press). Early Emotional Communication: Novel Approaches to Interaction. In J. Gratch & S. Marsella (Eds.), Social emotions in nature and artifact: Emotions in human and human-computer interaction (Vol. 14): Oxford University Press.
Messinger, D., Ruvolo, P., Ekas, N., & Fogel, A. (under review). Applying Machine Learning to Infant Interaction: The Development is in the Details. Neural Networks.
Tronick, E. Z., Messinger, D., Weinberg, K. M., Lester, B. M., LaGasse, L., Seifer, R., Bauer, C., et al. (2005). Cocaine Exposure Compromises Infant and Caregiver Social Emotional Behavior and Dyadic Interactive Features in the Face-to-Face Still-Face Paradigm. Developmental Psychology, 41(5), 711-722.
Yale, M. E., Messinger, D. S., & Cobo-Lewis, A. B. (2003). The temporal coordination of early infant communication. Developmental Psychology, 39(5), 815-824.
Yale, M. E., Messinger, D. S., Cobo-Lewis, A. B., Oller, D. K., & Eilers, R. E. (1999). An event-based analysis of the coordination of early infant vocalizations and facial actions. Developmental Psychology, 35(2), 505-513.