Maintenance and transformation of representational formats during working memory prioritization

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Visual working memory (VWM) relies on ventral visual stream (VVS) for stimulus maintenance and prefrontal cortex (PFC) for executive control.
• The precise mechanisms by which executive control prioritizes VWM content and its impact on representational formats remain unclear.

Purpose of the Study:

• To investigate the representational format of prioritized VWM content using intracranial EEG (iEEG) recordings.
• To explore how Deep Neural Network (DNN) architectures can model neural representations in VVS and PFC during prioritized VWM.

Main Methods:

• Analysis of iEEG data from epilepsy patients with electrodes in VVS and PFC during a multi-item VWM task with a retro-cue.
• Application of Representational Similarity Analysis (RSA) using various DNN architectures to characterize neural representations.

Main Results:

• Recurrent DNN representations in the beta band (15-29 Hz) after the retro-cue matched PFC representations.
• DNN representations in a lower frequency range (3-14 Hz) towards the end of the maintenance period corresponded to VVS representations.

Conclusions:

• Prioritized VWM content is represented differently in the VVS and PFC.
• Distinct coding schemes and representational formats are employed by VVS and PFC for prioritized information, as modeled by DNNs.