Drifting codes within a stable coding scheme for working memory.

Area of Science:

• Cognitive Neuroscience
• Neuroscience

Background:

• Working memory (WM) is crucial for short-term information retention, enabling stable mental states amidst dynamic brain activity.
• The inherent dynamism of neural processes poses a challenge for maintaining stable representations in WM.

Purpose of the Study:

• To investigate the interplay between working memory stability and neural dynamics.
• To understand how the brain maintains information despite neural fluctuations.

Main Methods:

• Electroencephalography (EEG) was employed to record neural responses to impulse stimuli during WM delays.
• Multivariate pattern analysis (MVPA) was utilized to decode neural representations of memorized information.

Main Results:

• Neural representations in WM exhibit both stable and dynamic characteristics.
• Distinct neural states were observed across different time points, indicating dynamic changes.
• A stable coding scheme for memorized orientations was identified within the dynamic neural system.
• Memory representations showed drift even within the stable coding subspace over time.

Conclusions:

• A stable subcomponent within WM facilitates information maintenance within a dynamic neural environment.
• Stable coding simplifies the readout of WM information for behavior.
• The dynamic component may encode temporal information.
• Drift in memory representations contributes to performance degradation and error distribution width in WM.