Strategy-dependent causal relationships between large-scale cortical activity and perceptual decisions

Animals employ various internal strategies when tackling decision-making tasks, yet the specific contributions of cortical regions to decisions under differing strategies remain unclear. To unravel this mystery, researchers examined three datasets consisting of either cortical widefield imaging or optogenetic inactivation of numerous cortical locations in 54 mice, all engaging in the same sensory-evidence-accumulation task.

By employing Generalized Linear Model-Hidden Markov Models (GLM-HMMs), scientists discovered that the mice adhered to one of three fundamental strategies: an engaged, stimulus-driven state and two disengaged states marked by significant bias towards one side.

The impact of inactivating cortical regions varied depending on the behavioral strategy: during engaged states, inactivating cortical areas often amplified side bias, whereas in biased states, the reverse effect was observed. For engaged states, these effects were particularly pronounced when frontal regions were inactivated. Furthermore, neural decoding analyses conducted on cortex-wide activity demonstrated that frontal regions were more adept at encoding the animals' choices during engaged states compared to biased states.

These findings suggest that distinct patterns of large-scale cortical dynamics underpin decisions under different task strategies and highlight the frontal cortex as a crucial component in controlling engaged-strategy execution.