History-dependent variability in population dynamics during evidence accumulation in cortex.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• The posterior parietal cortex is crucial for integrating sensory information and guiding behavior.
• Understanding how neural populations encode accumulating evidence is key to deciphering decision-making processes.

Purpose of the Study:

• To investigate how the posterior parietal cortex integrates new information with ongoing neural dynamics during evidence accumulation.
• To elucidate the mechanisms underlying short-term memory formation for sequential information processing in a virtual navigation task.

Main Methods:

• Analysis of population neural activity on single trials during a virtual navigation task in mice.
• Development of novel methods to track dynamic transitions between neural activity patterns.
• Quantification of trial-event-induced modifications to neural transition probabilities.

Main Results:

• Neural activity in the posterior parietal cortex transitions rapidly between distinct neuronal ensembles.
• Evidence cues and behavioral choices induce lasting modifications in neural transition dynamics, establishing short-term memory.
• A sequence of cues generates a cascade of these modifications, forming a signal for accumulated evidence.

Conclusions:

• Evidence accumulation can emerge implicitly from the dynamical properties of neural networks, specifically through short-term memory mechanisms.
• Explicit neural competition, as proposed by winner-take-all models, may not be necessary for evidence accumulation.
• The findings suggest a flexible neural representation where distinct activity patterns can encode similar levels of accumulated evidence.