Transfreq: A Python package for computing the theta-to-alpha transition frequency from resting state...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biomedical Engineering

Background:

• Estimating individual theta-to-alpha transition frequency (TF) typically requires two electroencephalographic (EEG) recordings, involving lengthy sessions and potential issues with incomplete alpha rhythm desynchronization.
• These limitations pose challenges for patient studies and accurate TF estimation.

Purpose of the Study:

• To introduce transfreq, a novel, publicly available Python library for computing TF from resting-state EEG data.
• To demonstrate the effectiveness and robustness of transfreq across various experimental conditions and its potential as a clinical marker.

Main Methods:

• The transfreq library employs spectral profile clustering of EEG channels based on theta and alpha band content.
• The algorithm's architecture and core functionalities are detailed, with validation performed on public and in-house EEG datasets.

Main Results:

• Transfreq successfully computes TF from resting-state data, even in scenarios where traditional methods fail.
• The study demonstrates transfreq TF's stronger correlation with the Mini Mental State Examination score compared to other EEG features, highlighting its potential as a clinical biomarker.

Conclusions:

• Transfreq offers a simplified and robust method for TF estimation from resting-state EEG.
• This tool shows promise for discovering novel biomarkers for neurodegenerative diseases, improving upon existing EEG-based markers.