Minimum data requirements and automated preprocessing for reliable EEG biomarkers in Rett syndrome.

Area of Science:

• Neuroscience
• Biomarker Development
• Quantitative Electroencephalography (qEEG)

Background:

• Electroencephalography (EEG) shows promise as a biomarker for Rett syndrome (RTT).
• Artifacts and long recording times challenge reliable EEG biomarker development.
• Automated preprocessing and minimum data requirements are crucial for patient-friendly RTT studies.

Purpose of the Study:

• To develop and validate an automated EEG preprocessing pipeline for Rett syndrome research.
• To determine the minimum EEG data length required for stable quantitative EEG (qEEG) features.
• To enable consistent, multisite analysis and patient-friendly data acquisition protocols.

Main Methods:

• A fully automated, correction-based preprocessing pipeline was applied to resting-state EEG from 117 participants (aged 1-18) in the R61 RTT study.
• The pipeline incorporated artifact handling, adaptive channel rejection, ASR, and ICA-based cleaning, extracting spectral power from 4-s epochs.
• Feature stability was assessed using Statistical Convergence and Model-Based Inflection approaches, validated against a rejection-based pipeline.

Main Results:

• The correction-based pipeline retained significantly more data (95.0%) compared to rejection-based methods (28.4%).
• Stable spectral power estimates were achieved with 76-136 seconds of EEG data.
• While RTT patients showed higher signal instability, minimum data requirements for stability did not differ significantly by group or age.

Conclusions:

• Approximately 3 minutes of resting-state EEG are sufficient for stable, clinically meaningful qEEG features in children with Rett syndrome using the proposed pipeline.
• These findings support shorter EEG acquisition protocols, enhancing feasibility for patients.
• The study provides a reproducible framework for data sufficiency in multisite neurodevelopmental research.