High-level visual and motor regions are doubly dissociated in controlled tasks but only at specific times during natural behaviors

Recent research has shed light on the neural mechanisms underlying sensory and motor information processing, primarily through controlled laboratory settings. However, these findings do not fully capture the complexity of natural behaviors, which entail more intricate visual cues and a broader spectrum of movements. It is widely believed that during natural behaviors, movement information influences visual cortex, while visual inputs impact motor cortex.

Alternatively, this co-mingling might merely reflect correlations between sensory inputs and movements observed in everyday activities. To clarify this matter, scientists conducted wireless recordings from high-level visual (IT) and motor regions (PMv/PFC) in primates while they engaged in both natural and controlled tasks. In each task, they analyzed visual and movement features to determine their influence on neural activity.

In the fixation task, a remarkable double dissociation was observed: visual features (but not movement) forecasted activity in IT, whereas motor features (not visual) predicted activity in PMv & PFC. This dissociation was notably weaker in the natural task, albeit still present at specific moments when visual and motor features exhibited reduced correlation.

Thus, high-level sensory and motor regions in the brain demonstrate a double dissociation under controlled conditions, yet this pattern persists at particular instances during natural behavior.