Analysis of electrooculogram (EOG) signals in studying myasthenia gravis

Myasthenia Gravis (MG) is a neuromuscular disorder that induces muscle weakness and fatigue which can be fatal. A common precursor for severe form of MG is ocular MG. In this thesis, we explored signal processing methodologies for early stage detection of MG using electrooculogram (EOG) signals. An EOG signal database consisting of 62 control and 16 MG (mild to moderate) subjects were analyzed for eye movement characteristics and EOG signal morphologies using time domain and wavelet domain techniques. A linear discriminant analysis (LDA) based classifier was used to quantify the ability of features in separating MG from control samples. Average overall classification accuracy achieved by the proposed method for the best time domain feature (average rise rate) and best wavelet feature (scale band energy) was 82.5% (P<0.01, AUC=0.887) and 83.8% (P<0.01, AUC=0.893), respectively. The obtained results suggest EOG based analysis is a viable, non-invasive alternative MG screening method.