Protocol for assisting frequency band definition and decoding neural dynamics using hierarchical clustering and multivariate pattern analysis

Chengpeng Li ¹, Isao Hasegawa ², Hisashi Tanigawa ¹

⁰Interdisciplinary Institute of Neuroscience and Technology, School of Brain Science and Brain Medicine, Zhejiang University, Hangzhou, China; College of Biomedical Engineering and Instrument Science, Zhejiang University, Hangzhou, China; MOE Frontier Science Center for Brain Science and Brain-Machine Integration, School of Brain Science and Brain Medicine, Zhejiang University, Hangzhou, China; National Key Laboratory of Brain and Computer Intelligence, Zhejiang University, Hangzhou, China.

Star Protocols +

|

June 4, 2025

View abstract on PubMed

Summary

This study introduces a new protocol for defining frequency bands in neural data analysis, improving accuracy by using data-driven clustering of macaque electrocorticography (ECoG) signals.

Area Of Science

Neuroscience
Computational Neuroscience
Signal Processing

Background

Traditional fixed frequency band divisions in neural data analysis can limit accuracy.
Electrocorticography (ECoG) provides high-resolution neural signals but requires precise analysis methods.

Purpose Of The Study

To present a novel protocol for data-driven frequency band definition in multichannel neural data.
To enhance the accuracy of neural data analysis by moving beyond arbitrary frequency divisions.

Main Methods

Preprocessing of multichannel macaque ECoG data.
Performing time-frequency analysis to obtain signal power profiles.
Applying hierarchical clustering to group similar frequency power profiles.
Defining frequency bands based on identified data clusters.
Utilizing multivariate pattern analysis (MVPA) for functional validation through time-series decoding.

Main Results

Identification of data-informed frequency groupings through hierarchical clustering.
Successful definition of new frequency bands guided by the clustering results.
Demonstration of functional relevance of the derived bands using MVPA and time-series decoding.

Conclusions

The proposed protocol offers a more accurate and objective method for defining frequency bands in neural data.
This data-driven approach can improve the insights gained from analyzing neural signals like ECoG.
The protocol facilitates enhanced functional validation of neural signal components.

Keywords:

Biotechnology and bioengineering Cognitive Neuroscience Neuroscience