NIH Research Festival
Electroencephalogram (EEG) can capture brain oscillatory activities during sleep as a form of electrophysiological signals. Unique brain wave activities may be present in patients with neurological disorders such as Down Syndrome (DS) when compared to healthy controls. In this study, we analyzed 97 full night in-laboratory polysomnography EEG recordings of DS subjects along with age- and sex-matched controls. We aimed to identify EEG features that are distinctive to patients with DS.
From each subject‚Äôs EEG, we extracted the relative power at six frequency bands (delta, theta, alpha, slow spindle, fast spindle, and beta) for each of the five sleep stages (N3, N2, N1, R, and W) and six channels (frontal F3 and F4, central C3 and C4, and occipital O1 and O2) ‚Äì 180 features in all. We applied statistical and machine learning methods to the data including paired t-tests and Extreme Gradient Boosting (XGBoost).
We showed that EEG features identified by XGBoost can distinguish between DS patients and healthy controls. Our results also revealed that, in N1 sleep, DS patients had significantly lower power in the alpha band (8-10.5 Hz) but higher power in the delta band (0.25-4.5 Hz). We also noticed that, during N1 sleep, the power difference between DS patients and matched controls gradually increased with age in both the alpha and the delta bands. Our findings suggest that unique EEG features can be identified in DS patients and that those features may be used as markers for disease management.
Scientific Focus Area: Neuroscience
This page was last updated on Monday, September 25, 2023