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

## TL;DR

This paper introduces a new protocol to define neural frequency bands using clustering and validation techniques, improving accuracy over traditional fixed bands.

## Contribution

A novel protocol for data-informed frequency band definition using hierarchical clustering and MVPA validation in neural data analysis.

## Key findings

- Hierarchical clustering identifies functionally distinct frequency groupings in ECoG data.
- MVPA validation confirms the functional relevance of derived frequency bands.
- The protocol enhances time-series decoding accuracy compared to fixed band divisions.

## Abstract

Traditional fixed frequency band divisions often limit neural data analysis accuracy. Here, we present a protocol for assisting frequency band definition for multichannel neural data using macaque electrocorticography (ECoG) data. We describe steps for performing time-frequency analysis on preprocessed signals and applying hierarchical clustering to frequency power profiles to identify data-informed groupings. We then detail procedures for defining frequency bands guided by these clusters and using multivariate pattern analysis (MVPA) on the derived bands for functional validation via time-series decoding.

For complete details on the use and execution of this protocol, please refer to Tanigawa et al.1

•Steps for defining frequency bands via hierarchical clustering•Procedures for time-frequency analysis and data clustering•Instructions for MVPA validation of derived frequency bands•Guidance on identifying functionally distinct neural sub-bands

Steps for defining frequency bands via hierarchical clustering

Procedures for time-frequency analysis and data clustering

Instructions for MVPA validation of derived frequency bands

Guidance on identifying functionally distinct neural sub-bands

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Traditional fixed frequency band divisions often limit neural data analysis accuracy. Here, we present a protocol for assisting frequency band definition for multichannel neural data using macaque electrocorticography (ECoG) data. We describe steps for performing time-frequency analysis on preprocessed signals and applying hierarchical clustering to frequency power profiles to identify data-informed groupings. We then detail procedures for defining frequency bands guided by these clusters and using multivariate pattern analysis (MVPA) on the derived bands for functional validation via time-series decoding.