Non-Linear and Chaos-based Analysis of Electroretinogram

Electroretinogram (ERG) is well-known for direct retinal function measurement. ERG responses to flicker stimulation can cause cyclic and oscillating changes in amplitude. Flicker response analyses are mostly based on amplitude and implicit time. However, non-linear analysis can also provide valuable information about the retinal function. In this paper, we investigate the flicker response using non-linear and chaotic features such as Approximate Entropy (ApEn), Hurst Exponent (HE), and Largest Lyapunov Exponent (LLE). Flicker responses were obtained from four groups with 16 subjects in each: one group with healthy subjects and three groups with central retinal vascular occlusion (CRVO), diabetic retinopathy (DR), and retinitis pigmentosa (RP) subjects, respectively. Statistical analysis shows that these non-linear and chaosbased features can distinguish the diseases and further indicate that the ERG has more complexity in healthy subjects than retinal disease subjects.