Spatial cognition versus language: What best differentiates autistic and neurotypical individuals?

Overview

While Autism Spectrum Disorder (ASD) is clinically defined by social communication differences and restricted, repetitive behaviours, cognitive abilities across the spectrum remain highly heterogeneous. By re-analysing the data of a comprehensive battery of tasks, this study investigated how sub-domains of spatial cognition and language differentiate autistic and neurotypical individuals using data-driven Elastic Net regression modelling. Identified spatial verbal memory as a key differentiating predictor and explored cross-system cognitive integration differences in ASD.

Core Findings

Both spatial cognition and language contribute to differentiating autism versus NT status.

Spatial verbal memory is the strongest differentiating factor. The most differentiating skills often involve later-developing capacities or those that rely on efficient cross-system integration.

Reflection

I’m more than happy to work on a Neurodiversity study! My experience on this project was multifaceted. Working alongside a statistics peer student expanded my analytical toolkit through exposure to unsupervised machine learning algorithms (Elastic Net regression model), while discussions with clinical practitioners offered valuable real-world context on childhood autism diagnosis and current ADOS-2 implementation.

Most strikingly, I came to appreciate that behind every data point lies an individual whose performance is deeply intertwined with their identity and profile—a nuanced connection that was less evident in my earlier work.