Linking neural population formatting to function

Area of Science:

• Neuroscience
• Computational Neuroscience
• Cognitive Science

Background:

• Complex behaviors in animals correlate with distinct brain areas.
• Neural activity in any brain area can often decode various sensory inputs, knowledge, and actions.
• The functional specialization of distinct brain regions remains a key question in neuroscience.

Purpose of the Study:

• To investigate whether the function of a brain area relates to information formatting (how information is combined) rather than just information presence.
• To compare information processing in the middle temporal area (MT) and dorsolateral prefrontal cortex (dlPFC).

Main Methods:

• Comparative analysis of neural representations in MT and dlPFC during decision-making tasks involving visual motion and reward.
• Utilized a recurrent neural network (RNN) model to simulate and predict the effects of information formatting.
• Experimental manipulation of neural activity via electrical stimulation in MT and dlPFC.

Main Results:

• Both MT and dlPFC contained motion and reward information, but encoded them differently.
• MT encoded information separately, while dlPFC represented information jointly, reflecting decision-making processes.
• Stimulating MT biased choices based on visual motion, while dlPFC stimulation led to 'winner-take-all' decisions.

Conclusions:

• Brain area function is determined by the way neural information is formatted and combined.
• Modular brain structure enables complex behaviors through flexible information reformatting.
• This finding supports a model where distinct brain areas contribute to complex cognition by processing information in specialized ways.