Distributed and dynamic storage of working memory stimulus information in extrastriate cortex.

Area of Science:

• Neuroscience
• Cognitive Science
• Neuroimaging

Background:

• The canonical model of working memory (WM) proposes information storage via stable, elevated neural activity in selective neurons, primarily in the lateral prefrontal cortex (lPFC).
• Emerging evidence suggests limitations of this model, necessitating a re-evaluation of WM neurobiology.

Purpose of the Study:

• To directly test key tenets of the canonical WM model using functional magnetic resonance imaging (fMRI).
• To investigate the neural mechanisms underlying information storage during WM tasks.

Main Methods:

• Analysis of fMRI data from participants performing a face and scene working memory task.
• Application of multivariate decoding to identify neural activity patterns encoding WM information.
• Examination of neural activity levels, selectivity, and stability over trial time.

Main Results:

• WM information was found in both extrastriate visual cortex (EC) and lPFC, but only EC showed sensory representation patterns.
• Information persisted in both regions without elevated activity, questioning the necessity of sustained high firing rates for WM storage.
• WM information was broadly distributed across neural populations in EC, not limited to highly selective neurons.
• Neural activity patterns coding for WM information were dynamic, varying over the trial course, contradicting the static storage hypothesis.

Conclusions:

• Findings challenge the canonical model by demonstrating that WM storage is not solely dependent on elevated, stable activity in highly selective neurons.
• Evidence supports a dynamic and distributed neural code for WM information, potentially involving sensory representations in extrastriate cortex.
• The canonical model requires revision to account for the dynamic and distributed nature of neural representations in working memory.