Transcriptomic Convergence in Autism Spectrum Disorder: Synaptic, Immune-Glial and RNA-Regulatory Axes in the Human Cerebral Cortex.

Autism spectrum disorder (ASD) arises from highly heterogeneous genetic and developmental liabilities, raising the question of whether this heterogeneity converges on shared molecular programmes in the human cerebral cortex. This structured review, based on systematic database searching and narrative synthesis, examined that question specifically in human post-mortem cortical transcriptomic studies. PubMed, Scopus and Europe PMC were searched from 1 January 2009 to 6 May 2026, and 43 studies met the final eligibility criteria. Across the available literature, the evidence does not support a single invariant cortical transcriptomic signature in ASD. Rather, the most consistent signal indicates non-uniform convergence on reduced neuronal and synaptic expression together with increased immune-glial programmes. A substantial additional body of evidence implicates dysregulation of transcript-regulatory processes, particularly in studies interrogating alternative splicing and related RNA-processing mechanisms. Cell-resolved datasets further suggest that these abnormalities are concentrated within defined neuronal and glial populations rather than being distributed uniformly across the cortex. By contrast, mitochondrial and broader metabolic alterations are supported less consistently and are better interpreted as conditional or secondary features of cortical pathology than as equally well-established core axes. Interpretation of these findings is constrained by the structure of the evidence base itself. Only 9 of the 43 included studies were judged to provide direct support for the central convergence question, and only 13 were based on primary independent cohorts; much of the literature relies on dataset reuse, regionally restricted sampling and heterogeneous analytical platforms. Collectively, human cortical transcriptomic studies in ASD support a model of partial, context-dependent convergence on a limited set of biological programmes, rather than a single stable molecular lesion.