The availability of extensive transcriptome data of the human brain has greatly expanded the possibilities to study the molecular differentiation of brain regions. A previously proposed method enabled the spatially comprehensive prediction of gene expression in the brain based on discrete microarray data. The resulting data was employed in the current work in order to derive a parcellation of the human cerebral cortex. To this end, the transcriptome dataset comprising normalized expression of 18,686 genes was used for agglomerative hierarchical clustering with Pearson correlation distance and average linkage. The optimal number of clusters indicated by the Bayesian Information Criterion was k=33. The transcriptome based parcellation was able to reproduce several well-established boundaries between cortical regions, such as primary sensory and motor areas, while revealing novel insights into their hierarchical organization.

Hierarchical clustering of the human cerebral cortex based on comprehensive transcriptome data

Scan to connect with one of our mobile apps

Coinbase Wallet app

Coinbase app

Or try the Coinbase Wallet browser extension