Closed-loop phase-targeted stimulation during sleep: Open-source benchmarking of methods and a novel algorithm for...

Area of Science:

• Neuroscience
• Sleep Science
• Computational Neuroscience

Background:

• Phase-targeted auditory stimulation (PTAS) enhances sleep slow oscillations (SOs) and memory consolidation.
• Clinical application of PTAS is limited by real-time phase estimation challenges, especially in epilepsy.
• Existing methods struggle with pathological EEG artifacts like interictal epileptiform discharges (IEDs).

Purpose of the Study:

• To develop a robust real-time phase estimation algorithm for PTAS in clinical populations, particularly those with epilepsy.
• To address the limitations of current PTAS methods in handling pathological EEG features.

Main Methods:

• Developed TWave, a real-time algorithm integrating wavelet-based phase estimation, predictive modeling, and multi-feature validation.
• Designed TWave to maintain SO phase estimation accuracy while rejecting EEG artifacts like IEDs.
• Validated TWave on healthy adult and pediatric epilepsy EEG recordings.

Main Results:

• TWave demonstrated high accuracy and precision in phase estimation for both healthy and epilepsy EEG data.
• The algorithm successfully rejected 83% of IEDs while preserving SO sensitivity.
• Benchmarking showed TWave outperformed existing algorithms in phase estimation precision across different EEG types.

Conclusions:

• TWave provides a validated method for real-time phase estimation, facilitating clinical translation of PTAS.
• The open-source TWave toolbox promotes reproducibility in sleep modulation research.
• This advancement is crucial for developing effective PTAS therapies for neurological conditions like epilepsy.