Th predominant view in the field has been that there is early cerebral overgrowth followed by either normalization or regression of brain volume in adolescence. However, this conclusion is based on cross-sectional comparisons. This means, they look at different people at different ages, and make inferences on developmental trajectories based on these cross-sectional, age-related patterns. While this is often a starting point, cross-sectional research can suffer from sampling biases.

A huge weakness in autism neuroimaging research is the un-representativeness of their samples. Nearly a third of individuals with autism have severe intellectual impairments (IQ's < 70) yet represent less than 1% of neuroimaging samples. Individuals with other immense behavioral, sensory and language challenges are also rarely make it through the rigors of imaging protocols.

A rare exception has been imaging research that performed brain imaging in very young children during natural sleep. and thus can hold still enough for quality MRI images to be acquired. This has allowed imaging of autistic children aged 2-6 years to include autism over a whole range of severities, challenges, and intellectual abilities.

This presents a problem though. The research that suggested there is brain overgrowth in early childhood sampled from a wide range of autism phenotypes and severities, while the normalization evidence in adolescents and adults came from autistic participants with normal ranged IQs and less severe challenges, a clear cross-sectional sampling bias that threatens the validity of the overgrowth normalization story. Moreover, research indicated that disproportionate brain size in autism was associated with slower intellectual improvements with development.

I and my colleagues thus hypothesized that the discrepancy would be removed if we can follow the same children from childhood into adolescence longitudinally. Using a number of behavioral techniques and a lots of care and dedication, our team managed to acquire brain data in a broad spectrum of autism phenotypes and severity levels from early childhood into early adolescence.

We reported the results of our study in Biological Psychiatry Lee in 2021. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8089123/ (open access).

Our conclusion: Longitudinal evidence does not support the notion that early brain overgrowth is followed by volumetric regression, at least from early to late childhood.