2025-07-23 – 20:30-21:00 (Africa/Abidjan), Main Room 2
Large scale EEG analysis pipelines are a critical tool for understanding complex neurophysiology. Using Beacon Biosignal’s platform for EEG data analysis in Julia we performed one of the largest studies of chronic insomnia to date. It appears that the most prevalent EEG differences between those with and without insomnia arise during periods of wake or near wake-like (N1) activity throughout the night. Overall Julia proved to be a robust platform for EEG ingest, featurization and analysis.
Chronic insomnia is a common disorder (affecting ~10% of the population) which is diagnosed based on self-reported symptoms, but is often difficult to objectively characterize (with a high mismatch between objective measurements and symptom burden). When examined across a massive scale (thousands of individuals), quantitative EEG features can provide deep insight into insomnia severity and diagnosis.
Our post-hoc analysis across three EEG dataset found that individuals with insomnia had: 1.) a higher likelihood of waking from all sleep stages, 2.) increased wake-like brain activity according to both WESI—an aggregate data-driven measure of arousal—as well as elevated alpha and theta power during periods of wake and 3.) altered sleep spindle characteristics. We then examined one of these datasets—originating from a phase III clinical trial—to see if these same features were reversed following treatment with a dual orexin receptor antagonist (DORA). This class of drugs aims to target the orexin/hypocretin system, thought to play a key role in maintaining wakefulness and promoting arousal. We found that treatment with the DORA partially reversed insomnia-linked differences in EEG by: reducing wake-to-wake transitions and increasing transitions into sleep stages (N1, N2, and REM), reducing measures of wake-like brain activity according to both WESI during wake, reduced beta and alpha power during wake and N1, and increased delta power during wake. Therefore it appears that the most prevalent EEG differences between those with and without insomnia arise during periods of wake or near wake-like (N1) activity throughout the night.
In this talk I’ll describe these findings, the technical successes and technical challenges of performing the analyses as well as some of the strategies we have adopted at Beacon to manage massive heterogeneous datasets. Julia has proven itself to be a robust platform for managing data ingest, feature extraction and model fitting at scale. Sources of success include Julia’s ergonomics, the robust open source libraries available and the ease of integrating with existing tooling built in other languages. Sources of challenge with Julia largely stem from the immaturity of some corners of the ecosystem. Overall Julia has demonstrated its ability to perform scalable analytics to improve neurophysiological data processing.
I work at Beacon Biosignals to bring scientific insight to large scale neurophysiological datasets of EEG, identifying novel biomarkers for the treatment and stratification of neuropathology. I have a background in human auditory psychophysics and auditory cognitive neuroscience as well as machine learning and signal processing methods applied to music.