Abstract:Recent advances in computational modeling techniques have facilitated a more nuanced understanding of sleep neural dynamics across the lifespan. In this study, we tensorize multiscale multimodal electroencephalogram (EEG), electromyogram (EMG), and electrooculogram (EOG) signals and apply Generalized Autoregressive Conditional Heteroskedasticity (GARCH) modeling to quantify interactions between age scales and the use of pharmacological sleep aids on sleep stage transitions. Our cohort consists of 22 subjects in a crossover design study, where each subject received both a sleep aid and a placebo in different sessions. To understand these effects across the lifespan, three evenly distributed age groups were formed: 18-29, 30-49, and 50-66 years. The methodological framework implemented here employs tensor-based machine learning techniques to compute continuous wavelet transform time-frequency features and utilizes a GARCH model to quantify sleep signal volatility across age scales. Support Vector Machines are used for feature ranking, and our analysis captures interactions between signal entropy, age, and sleep aid status across frequency bands, sleep transitions, and sleep stages. GARCH model results reveal statistically significant volatility clustering in EEG, EMG, and EOG signals, particularly during transitions between REM and non-REM sleep. Notably, volatility was higher in the 50-66 age group compared to the 18-29 age group, with marked fluctuations during transitions from deep sleep to REM sleep (standard deviation of 0.35 in the older group vs. 0.30 in the 18-29 age group, p < 0.05). Statistical comparisons of volatility across frequency bands, age scales, and sleep stages highlight significant differences attributable to sleep aid use. Mean conditional volatility parameterization of the GARCH model reveals directional influences, with a causality index of 0.75 from frontal to occipital regions during REM sleep transition periods. Our methodological framework identifies distinct neural behavior patterns across age groups associated with each sleep stage and transition, offering insights into the development of targeted interventions for sleep regularity across the lifespan.

Conditional Spectral Analysis of Replicated Multiple Time Series With Application to Nocturnal Physiology

Adaptive Bayesian Spectral Analysis of Nonstationary Biomedical Time Series

Canonical correlation analysis between time series and static outcomes, with application to the spectral analysis of heart rate variability

The Temporal Dynamics of Aperiodic Neural Activity Track Changes in Sleep Architecture

Multiscale neural dynamics in sleep transition volatility across age scales: a multimodal EEG-EMG-EOG analysis of temazepam effects

Corticothalamic modelling of sleep neurophysiology with applications to mobile EEG

Estimation of Time-Varying Spectral Peaks and Decomposition of EEG Spectrograms

0279 Intermittent Analysis of Obstructive Sleep Apnea Dynamics

Multiple Intrinsic Timescales Govern Distinct Brain States in Human Sleep

Discriminant Analysis of Time Series in the Presence of Within-Group Spectral Variability

Mixed effect Poisson log-linear models for clinical and epidemiological sleep hypnogram data

Koopman Spectral Analysis of Intermittent Dynamics in Complex Systems: A Case Study in Pathophysiological Processes of Obstructive Sleep Apnea

Normative characterization of age-related periodic and aperiodic activity in resting-state clinical EEG recordings

Mortality Risk Assessment Using Deep Learning-Based Frequency Analysis of EEG and EOG in Sleep

Seeking for a fingerprint: analysis of point processes in actigraphy recording

Spectra in low-rank localized layers (SpeLLL) for interpretable time-frequency analysis

A novel sparse linear mixed model for multi-source mixed-frequency data fusion in telemedicine

Spectral Graph Sample Weighting for Interpretable Sub-cohort Analysis in Predictive Models for Neuroimaging

: a nonparametric model for neural power spectra decomposition

Scaling behaviour of heartbeat intervals obtained by wavelet-based time-series analysis

Sleep electroencephalography and heart rate variability interdependence amongst healthy subjects and insomnia/schizophrenia patients