Thesis defense Giorgia Bussu (Donders series 367)

20 june 2019

Promotors: prof. dr. J. Buitelaar, prof. dr. C. Beckmann
Co-promotors: dr. E. Jones (Birkbeck College, UK)

Focusing on the individual infant: classification and heterogeneity of autism spectrum disorder

To enable early targeted intervention for Autism Spectrum Disorder (ASD), it is important to identify the infants who need intervention early in life. Group differences tell us little if anything about the individual infant as there can be significant overlap between groups in individual variation. This research aimed to translate group differences to prediction of ASD outcome in the first years of life at the level of the individual infant. First, we used multi-modal multi-domain data integration for early classification of ASD. Then, we focused on the prospective investigation of its heterogeneity through unsupervised learning. Results indicate that brain data, as opposed to behavioural and developmental data, retain the highest value for early detection of ASD. In particular, we found that a general alteration in neural processing of faces at 8 months might be an indicator of later ASD. Furthermore, data-driven analyses allowed us to identify subgroups among high-risk siblings and low-risk controls, and to uncover underlying processes acting together early in development and associated to ASD outcome.