772-784 Herbert [0318]

Social and emotional processing uses neural systems involving structures ranging from the brain stem to the associational cortex. Neuroimaging research has attempted to identify abnormalities in components of these systems that would underlie the behavioral abnormalities seen in disorders of social and emotional processing, notably autism spectrum disorders, the focus of this review. However, the findings have been variable. The most replicated anatomic finding (a tendency toward large brains) is not modular, and metabolic imaging and functional imaging (although showing substantial atypicality in activation) are not consistent regarding specific anatomic sites. Moreover, autism spectrum disorder demonstrates substantial heterogeneity on multiple levels. Here evidence is marshaled from a review of neuroimaging data to support the claim that abnormalities in social and emotional processing on the autism spectrum are a consequence of systems disruptions in which the behaviors are a final common pathway and the focal findings can be variable, downstream of other pathogenetic mechanisms, and downstream of more pervasive abnormalities. (J Child Neurol 2004;19:772–784). Received March 19, 2004. Received revised June 17, 2004. Accepted for publication June 17, 2004. From the Center for Morphometric Analysis, Division of Pediatric Neurology, Department of Neurology, Massachusetts General Hospital-East, Neurosciences Center, Harvard Medical School, Charlestown, MA. This work was supported by National Institute of Neurological Disorders and Stroke grant NS02126 and the Cure Autism Now Foundation. Address correspondence to Dr Martha R. Herbert, Center for Morphometric Analysis, Division of Pediatric Neurology, Department of Neurology, Massachusetts General Hospital-East, Neurosciences Center, Harvard Medical School, 149 13th Street Room 6012, Charlestown, MA 02129. Tel: 617-724-5920; fax: 617-812-6334; e-mail: [email protected]. common belief in the field that the variability of measures is a consequence of differences in methodology, whether of subject selection or investigative tools. This notion derives from the assumption that the common behavioral features in these syndromes should rest on a common underlying biology and that were methods and subjects more consistent, it would be easier to identify specific underlying commonalities. However, there is another possibility, which is that some of the variability might be an irreducible intrinsic reflection of the multiple pathways from biology to behavioral phenotype,8 with the autistic syndrome being not a specific biologic entity but a “final common pathway.” There is another type of variability in these disorders, namely, a difference in behaviors over time in individual subjects. These differences can occur on a day-to-day basis, with “good” and “bad” days or behaviors interspersed among each other. It can also occur over longer intervals, in those cases in which the results of therapeutic interventions lead to more enduring improvements in the level of functioning of affected individuals. These types of variability suggest that there are at least some aspects of the underlying abnormalities that are not strictly “hardwired” but rather are subject to change in various ways. A further type of variability can well be changing the rates of the disorder. A growing number of studies suggest an increase in incidence.9,10 Although this remains the subject of controversy, if it is even partly true, then the environmental component of the gene-environment interactions underlying the disorder might be changing.11,12 The study of brain abnormalities in the autism spectrum has largely aimed to identify brain-behavior correlations. This requires identifying regions of the brain that normally subserve the behaviors that are disordered; however, for the core behavioral features of the spectrum other than language, the relevant normal regional specializations are poorly understood. Even for language, the changes in regional specialization over developmental time, which are highly relevant for developmental disorders, are not clearly known; and regions involved in language domains especially impaired in autism, such as pragmatics, are poorly characterized. Brain-behavior correlation also involves characterizing the neuroanatomic and neurofunctional abnormalities underlying the neurobehavioral impairments. However, this correlational project is challenged by the diversity and inconsistency of the findings among studies and among individuals, some of which are documented below. In this article, I argue that the data support a shift of interpretive emphasis. Rather than trying to eliminate the variability, I pose the question of just how it can be that such a presumably heterogeneous set of underlying biologic abnormalities can eventuate in the same syndrome of behavioral abnormalities. How can we characterize the cognitive neuroscience of the phenomenon of a final common pathway? I hypothesize that autism is a disorder of systems alteration or disruption and that systems can be impacted in multiple ways and yet yield a similar syndrome of behaviors. From this vantage point, the variability in biomarkers can be transformed from noise into signal,13 yielding insight into the multiple systems vulnerabilities that can lead to this type of disordered social-emotional functioning. We also need to characterize the limits to the variability of systems impact; that is, under what circumstances can systems disruption lead to a lesser or different syndrome than autism? Formulating the research challenge in these terms can allow a more parsimonious integration not only of variability in biologic findings but also of variability in the gene-environment interactions involved in underlying mechanisms. ANATOMY OF SOCIAL AND EMOTIONAL PROCESSING In broad terms, social and emotional behaviors are evolved modes of regulation of individual and group dynamics. At the neuroanatomic level, they involve systems ranging from core somatic regulatory circuitry to circuits modulating action and interaction with objects and beings in the world. They coordinate physiologic adaptations with responses to safety and danger. In more social organisms with more complex intelligences, the environment becomes more functionally differentiated for species with an expanded repertoire of perception and response capabilities.14 In highly social species, the social environment becomes ever more central to survival, and the evaluation of the intentions, emotions, and thoughts of conspecifics comes to occupy a major place in mental life and in survival. In such organisms, both the social environment and the ability to perceive it strongly affect developmental outcomes. The evolving adaptive and associative intelligence that enables this increasing social complexity cannot be entirely separated from that enabling increasing complexity in mental operations used for material adaptive strategies. Visceral regulation is an important component of emotional processing and is affected by social experience. Dysfunction in visceral regulation systems can be a component of both dysphoric states and emotional disorders, as well as a consequence of social misfortune. A three-stage model of phylogenetic stages of autonomic nervous system development has been proposed.15–17 The most primitive stage generates an unmyelinated vegetative vagal system that fosters digestion and that responds to threat by cardiac output reduction and immobilization. A later emerging spinal sympathetic nervous system inhibits the earlier vagal system’s influence on the gut and increases the metabolic rate to mobilize for “fight or flight.” A third stage, unique to mammals with a myelinated vagal system, is related to regulating cardiac output in relation to engagement and disengagement with the environment. This third system is derived from the branchial or primitive gill arches and thus relates to cranial nerves V, VII, IX, X, and XI. Owing to the functions of these cranial nerves, it is thus related to regulation of components of social engagement that include hearing via middle ear muscles, facial expression via facial musculature, the coordination of breathing with vocalizing via the laryngeal and pharyngeal muscles, and orientation via headturning muscles. Via rapid vagal regulation of the heart and bronchi, it also modulates rapid social engagement and disengagement. This level of nervous regulation is also associated with fostering early mother-child interactions. Vagal regulation modulates the release of oxytocin and vasopressin, two neuropeptides involved in social bonding and isolation and with responses to safety and fear. Abnormalities in this level of neural systems related to social and emotional behavior have been associated with the autism spectrum, and it is argued that they contribute to the autism. These include brainstem abnormalities,18–20 abnormalities of oxytocin,21 and autonomic dysregulation.22,23 Neuroimaging in Social Emotional Functioning Disorders / Herbert 773