Alpha phase coding supports feature binding during working memory maintenance

Area of Science:

• Cognitive Neuroscience
• Neuroscience

Background:

• Working memory (WM) relies on binding object features into unified representations.
• The precise neural mechanisms for feature binding and the cause of binding errors remain largely unknown.
• Binding errors, where features are incorrectly associated, highlight limitations in WM capacity.

Purpose of the Study:

• To investigate the role of neural phase synchrony in working memory feature binding.
• To test the hypothesis that binding errors stem from perturbations in neural synchrony.
• To elucidate the neural dynamics underlying feature binding and WM capacity limitations.

Main Methods:

• Utilized magnetoencephalography (MEG) to record brain activity in human subjects.
• Employed a cognitive task specifically designed to elicit binding (swap) errors in working memory.
• Analyzed phase-locked oscillatory activity and phase coding variability during memory retention.

Main Results:

• Swap errors were associated with reduced phase-locked oscillatory activity during WM retention.
• This reduction aligns with predictions from attractor models of spiking neural networks.
• Increased phase coding variability in the alpha-band over sensorimotor areas characterized swap errors.

Conclusions:

• Neural phase synchrony is a critical mechanism for feature binding in working memory.
• Binding errors arise from disruptions and increased variability in phase coding dynamics.
• These findings suggest that feature binding emerges from neural competition within a distributed network.