Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder characterized by deficits in social, communication and restricted and repetitive behaviors. A significant challenge in understanding ASD lies in its heterogeneity, with up to 70% of patients reporting comorbid psychiatric, medical or genetic conditions, as well as vast differences in neuroimaging, genetic and immune factors. This variability has hindered biomarker isolation, possibly due to widely used case-control experimental designs combining patients varying in behavioral, genetic and/or clinical profiles into one group. The global aim of this thesis work is to better characterize autistic patients, which is vital in the advancement of appropriate treatments and therapies. To achieve this, we used statistical methods and multidimensional data. We show that case-control comparisons in autistic populations fail to elicit consistent and meaningful results, and that a combination of dimensional and subgrouping approaches proves most valuable in the understanding of ASD. In the process, we isolated important and consistent behavioral and neuroimaging autistic traits subgroups.