The DYNAM-O Toolbox: Characterizing Individualized Neural Signatures in Sleep EEG

Abstract

Here, we introduce the Dynamic Oscillation (DYNAM-O) Toolbox, an open-source, cross-platform (MATLAB, Python, and Rust) software package for data-driven characterization of individualized neural dynamics in sleep EEG. Conventional sleep electroencephalography (EEG) measures often rely on predefined bands, thresholds, and averages that incompletely capture transient oscillatory dynamics across the night. For example, sleep spindles, 12-16 Hz bursts of EEG activity related to memory consolidation and altered in disease states, are traditionally detected using fixed frequency and amplitude criteria. Recent studies, however, have shown that canonical spindles represent only a small subset of the tens of thousands of transient "spindle-like" oscillations occurring each night, comprising multiple event classes across a broad frequency range. Together, these classes of transient oscillations form highly individualized signatures of brain state with demonstrated utility as EEG biomarkers of disease. DYNAM-O substantially extends our transient oscillation framework and provides the first end-to-end pipeline for this analysis, from event detection to group-level statistics. It identifies transient oscillations as time-frequency peaks (TF-peaks) using a novel multi-resolution procedure that combines time- and frequency-optimized multitaper spectrograms to separate closely spaced events and refines each event's peak frequency to resolve oscillations separated by less than the spectrogram resolution. It then computes intrinsic and sleep-state-dependent extrinsic features for each event and represents the overnight distributions of TF-peaks as feature histograms spanning oscillation frequency, slow oscillation power, and slow oscillation phase, preserving continuous brain-state variation obscured by stage-based averaging. New parametric and spline-based dimensionality reduction methods distill these histograms into a small number of interpretable modes, and integrated whole-histogram statistical tests support exploratory and hypothesis-driven group analyses. As a demonstration, we analyzed sex differences in polysomnography recordings from 112 adults (C3 electrode; 67 females, 45 males; ages 20-35 years) from the Cleveland Family Study. In addition to replicating the established higher fast-spindle frequency in females, we identified greater low-alpha transient oscillatory activity in females, a novel sex difference in events ignored in traditional analyses. DYNAM-O provides an accessible and interpretable framework for studying individualized sleep physiology, with applications to longitudinal studies, group comparisons, and feature extraction for deep learning models.