Map of the brain cortex based on data from resting state fMRI scans containing 180 distinct areas per hemisphere. Shown are areas for hearing (red), touch (green), vision (blue) and opposing cognitive systems (light and dark). Image credit: Matthew Glasser and David Van Essen (CC BY-NC 2.0)
Reaching out to grab a cup of coffee may feel effortless, but it requires the brain to coordinate a precise sequence of movement commands. This control is supported by the sensorimotor cortex, the part of the brain's wrinkled outer layer involved in integrating incoming sensory feedback with outgoing movement commands.
Researchers have traditionally divided the sensorimotor cortex into separate anatomical areas based on the shapes of neurons as well as their connectivity to other brain structures. This area-based view has shaped our understanding of the sensorimotor cortex based on the idea that different areas generally serve different functions. However, it has been less clear whether these functions change sharply at the borders dividing regions or whether groups of neurons within an area might perform different functions.
Anatomical areas have long formed the basic unit for studying cortical function, yet recent work shows that movement-related activity is far more widely distributed than this view assumes. Understanding how the cortex controls movement, therefore, requires mapping the responses of large numbers of individual neurons across different areas, and determining how responses relate to known anatomical and somatotopic boundaries. Salimian, Grier and Kaufman sought to characterize how responses of single neurons are distributed across anatomically defined sensorimotor areas in the cortex of mice while they performed reach-to-grasp movements.
Salimian et al. characterized the activity of nearly 40,000 individual neurons across different areas of the sensorimotor cortex of mice using interpretable features obtained from their activity during a challenging forelimb control task. Neural response features were then organized into coherent spatial gradients spanning motor and somatosensory cortical areas.
The results showed that the spatial patterns possessed sharp transitions that closely aligned with anatomical and somatotopic borders. Clustering cells based on their features identified four unique subpopulations with characteristic response profiles, whose members were widely distributed across all recorded areas, but with different prevalence in each. The spatial distributions of these subpopulations formed overlapping zones, in which neurons from different subpopulations intermingled.
The findings of Salimian et al. suggest that a complete understanding of the sensorimotor cortex requires mapping the distributed neuronal networks, instead of just focusing on separate, specialized areas. In the future, this view could inform clinical technologies such as the placement of recording electrodes for brain-computer interfaces that help paralyzed people move, or stimulation-based maps used to guide surgical resection of brain tissue.
In addition, they may inform the design of neural networks for controlling robots, where sensory feedback must be integrated into choosing motor commands. However, realizing such benefits would first require confirming that these subpopulations exist in the human brain and elucidating their contributions to movement control.
