BROAD-NESS Uncovers Dual-Stream Mechanisms Underlying Predictive Coding in Auditory Memory Networks.

Area of Science:

• Neuroscience
• Cognitive Science
• Brain Network Analysis

Background:

• Understanding whole-brain network dynamics in auditory memory and predictive coding remains a challenge.
• Existing methods may not fully capture the complexity of broadband brain networks.

Purpose of the Study:

• Introduce BROAD-NESS (BROADband brain Network Estimation via Source Separation), a novel pipeline for whole-brain network extraction from MEG data.
• Characterize the spatiotemporal architecture of auditory memory networks.
• Investigate the relationship between network dynamics and auditory sequence recognition performance.

Main Methods:

• Developed and applied the BROAD-NESS pipeline to source-reconstructed MEG data.
• Utilized auditory sequence recognition tasks.
• Employed phase space and recurrence quantification analysis (RQA).
• Compared Principal Component Analysis (PCA) and Independent Component Analysis (ICA) based network extraction.

Main Results:

• BROAD-NESS identified two orthogonal networks centered on auditory cortices.
• One network processes sounds, showing subtle condition differences; the second involves hippocampus and other regions, reflecting prediction and error processing.
• More recurrent and stable network dynamics were associated with higher accuracy and faster responses.
• PCA-based networks demonstrated greater robustness and interpretability compared to ICA.

Conclusions:

• A dual-stream organization of auditory memory networks exists, functionally diverging from visual pathways.
• BROAD-NESS provides a powerful and interpretable tool for analyzing brain network spatiotemporal architecture.
• Network stability and recurrence are key indicators of auditory memory performance.