Annual Philosophy of Social Science Roundtable

Elizabeth R. Blum and Rajesh K. Kana (University of Alabama at Birmingham): Conceptual Issues in Psychiatric Disorders: The Curious Case of Autism [email protected] According to the DSM-IV, autism is a developmental disorder characterized by impairments in reciprocal social interaction, language, and communication; restricted interests; and repetitive behaviors. Despite being the subject of active scientific investigation (including behavioral, clinical, genetic and neuroimaging approaches), autism remains enigmatic to researchers and families, and has consequently generated intense public interest. While much progress has been made in the characterization of autism since Kanner’s (1943) initial description, conceptual issues remain. In this paper, we investigate how the problems associated with the conceptualization of autism may affect the nature of continuing empirical research. We will discuss several issues that are relevant to questions of measurement and other areas of philosophy of social science. The argument, which is informed by empirical research from our group, is motivated by a desire to generate a conceptual framework that may be more useful in scientific practice. Spectrum: A fundamental tension can be seen in the question of whether autism ought to be characterized as a set of categories or as a continuum. On the one hand, the DSM-IV places the disorder in a categorical format under the umbrella Pervasive Developmental Disorders (PDD). On the other hand, PDD encompasses several disorders with overlapping symptoms and characteristics. For this reason, PDD disorders are often discussed as “Autism Spectrum Disorder” (ASD), which suggests a range of conditions falling on a continuum. In addition, ASD comprises disorders with huge variation in abilities and impairments, which further supports the idea of “continuum”. This tension between the “continuum” and the categories is problematic in terms of diagnosis, intervention, education and treatment, and intersects with the issues outlined below. Variability: Although there are sound reasons, including much clinical and experimental research, for treating autism spectrum disorder as a homogeneous category, autism differs from other developmental disorders in that its presentation is highly variable across individuals. This heterogeneity is surprising given the fairly concrete evidence for both a common developmental course and a strong genetic component to the disorder. Clarity in the neurobiological profile of autism has also been elusive, especially the efforts to track autism down to a modular brain abnormality. The variability is implicitly acknowledged by the highly individualized treatment approaches in autism. Yet, the variability in the manifestation of the disorder is challenging since clear definitions and categories are required when diagnosing and classifying (for instance, the absence of definitions might lead to mis-, underor overdiagnosis), and ultimately treating a patient. The variability also creates difficulty for research studies (whether behavioral or neuroimaging) that focus on data averaged across a sample. This may lead researchers to commit the “ecological fallacy”: a mistaken inference from properties of the population to properties of the individuals in it. In addition, for the obvious practical reasons, many research studies focus on high-functioning individuals, which may lead to biased results when the findings are assumed to be applicable to the entire autism population. Another problem relates to the tendency to generalize brain results across individuals, groups, or populations, despite our knowledge that brain structure is highly variable. In other words, the focus when reporting results is always toward the forest with little attention to