Fuzzy Categorical Nature of ASD

Within recent models of the disorder, autism spectrum disorder (ASD) is conceptualized as a condition that combines categorical and dimensional attributes. In such models, ASD is considered to be separated from typical development and other psychiatric disorders in a graded manner.

Inspired by in-depth philosophical discussions on the nature of mental disorders, we introduce a novel conceptualization of ASD, so-called fuzzy categorical model of ASD, that can reconcile current dimensional and almost historic categorical accounts. This latent model improves upon “categorical vs. dimensional” dichotomy. As the name implies, a fuzzy structure lacks sharp boundaries and is neither a discrete/bounded category nor a perfect continuum. It allows for the presence of intermediate cases between those who are affected and those unaffected. We develop computational tools to facilitate testing of this hypothesis and apply these tools to better elaborate the nature of ASD.

Illustration of the underlying main hypothesis. (a) The separation between ASD and non-ASD, as illustrated in a two-dimensional feature space spanned by features f1 and f2 (e.g., social communication and motor skills), is assumed to be fuzzy in nature. That is, there is a continuous valued dimension of autism severity (as compared to a binary valued label), with presence of a transition region between ASD and non-ASD consisting of individuals who are neither clearly ASD nor clearly non-ASD. (b) In a classification framework (ASD vs. non-ASD), a significant portion of heterogeneity can be captured by measuring the distance between an individual and the classification boundary (the dashed line). This leads to the hypothesis that individuals who are close to the classification boundary (within the ellipse) are more likely to change diagnostic category.

Source Codes
PUNCH: Population Characterization of Heterogeneity

Tunç B., Pandey J., John T. St., Meera S., Maldarelli J. E., Zwaigenbaum L., Hazlett H. C., Dager S. R., Botteron K. N., Girault J. B., McKinstry R. C., Verma R., Elison J. T., Pruett Jr. J. R., Piven J., Estes A. M., Schultz R. T., Diagnostic shifts in autism spectrum disorder can be linked to the fuzzy nature of the diagnostic boundary: A data-driven approach, , , 2020
Tunç B, Yankowitz L D, Parker D, Alappatt J A, Pandey J, Schultz R T, Verma R, Deviation from normative brain development is associated with symptom severity in autism spectrum disorder, Molecular autism, 10:46, , 2019
Tunç B, Ghanbari Y, Smith A R, Pandey J, Browne A, Schultz R T, Verma R, PUNCH: Population Characterization of Heterogeneity, NeuroImage, vol:98, 50-60, 2014