Authors :
Presenting Author: Noshin Tasnia, BS – University Of Alabama at Birmingham
Siddharth Gupta, PhD – Kennedy Krieger Institute
Aida Doucoure, MA – The University Of Virginia
Constance Smith-Hicks, MD, PhD – Center for Synaptic Disorders, Rett and Related Disorders Clinic, Kennedy Krieger Institute
Rachel Smith, PhD, MS, BS – University of Alabama at Birmingham
Rationale:
Electroencephalography (EEG) is often used in the assessment of neurodevelopmental disorders to characterize brain network dysfunction and quantify brain network changes as a function of time or after introduction of specific therapies. SYNGAP1-Related Disorder (SYNGAP1-RD) is an autosomal dominant neurodevelopmental disorder resulting from pathogenic variants in SYNGAP1. We aimed to identify EEG features that distinguish SYNGAP1-RD from neurotypical control subjects, highlighting potential disease relevant biomarkers that may modulate with successful therapeutic interventions.
Methods:
Because the SYNGAP1 gene regulates synaptic function, variants in this gene lead to dysfunction in neuronal communication, altering global brain network dynamics. To assess this, we have performed a cross-correlation based functional connectivity analysis in nine SYNGAP1-RD patients and eight neurotypical control subjects. EEG data were collected using 31 electrodes, placed according to the 10-20 placement system and recorded at 1000 Hz sampling rate. The recorded data were first cleaned of artifactual time periods, re-referenced to the common average, and broadband filtered between 0.5-55 Hz using a 4th order butterworth filter. The data were divided into 1-second windows, and we identified maximum correlation values within ±200ms lags, with a Fisher z-transformation to normalize correlation strengths. Nonparametric permutation resampling was used to determine significance of individual connections. Volume conduction effects were controlled for by removing observations with a maximum correlation at zero lag, and partial correlations were used to reject observations of strong connectivity due to re-referencing. We averaged across individual one-second windows to determine the overall connectivity for each patient.
Results:
We averaged connectivity matrices across subjects to compare consistent connectivity differences between SYNGAP1-RD and control subjects (Figure 1). We found significantly stronger diffuse network connectivity in SYNGAP1-RD, with a median connection strength of 0.080 (IQR: 0.095), while the median connection strength in Control subjects was 0.055 (IQR: 0.097). This difference in network strength was significant (Wilcoxon rank-sum test, p< 0.001) (Figure 2). We also found that 74.0% of connections are stronger in SYNGAP1-RD patients when compared to controls, though specific connections were much stronger in the Control subjects, indicating physiological connections that are much weaker in the SYNGAP1-RD subjects.