Publications

Background:

Nonlinear measures like Sample Entropy which is a measure of chaos based on self-similarity extended from dynamic systems & chaos theory can potentially help to interpret time-series data obtained from EEG signals and facilitate understanding of complex proles seen in ASD. The aim of this study is to characterize EEG patterns in ASD cohorts by extracting non-linear measures (Sample Entropy) from time-frequency data in awake and sleep stages.

Methods:

Unmedicated Preschool children, ages 2 to 6 years, underwent EEG recording up to 180 mins with 64 Channel EGI-GES400 system after a sleep deprivation protocol. ASD group had 28 children (24M, 4F). The controls were those with ADHD, Specic Speech Delay, and Global Developmental Delay had 10 children (6M, 4F). Awake and sleep stage 1 (N1), 2 (N2) and 3 (N3) recordings were included. Artifact free continuous segment of 60s were selected manually. Each 10s epoch was decomposed into 6 bands (gamma high through delta) by wavelets. Sample Entropy (SE) was computed for each band per channel. Mean Differences were computed between groups in awake & sleep stages. Significance was estimated with p-values (α=0.05).

Results and conclusions:

Differences in entropies was significant (p < 0.05) in N2 & N3. In higher bands (gamma & beta), ASD was dominant in parietal & frontal particularly Pz. In N3, non-ASD had dominant delta in occipital, parietal & frontal regions. A higher entropy translates to more possible states of being. Delta band is more dynamic in non-ASD group while higher bands are more dynamic in ASD group.