1D Convolutional Neural Networks for Detecting Nystagmus

Vertigo is a type of dizziness characterised by the subjective feeling movement despite being stationary. One in four individuals community experience symptoms at any given time, and it can be challenging for clinicians to diagnose underlying cause. When result malfunction inner-ear, eyes flicker this called nystagmus. In article we describe first use Deep Neural Network architectures applied detecting The data used these experiments was gathered during clinical investigation novel medical device recording head eye movements. We methods training networks using very limited amounts data, with an average 11 mins nystagmus across subjects, less than 24 hours total, per subject. Our work replicating modifying existing samples generate new data. cross-fold validation experiment, achieve F1 score 0.59 (SD = 0.24) all folds, showing that employed are capable identifying periods modest degree accuracy. Notably, were also able identify pathological produced patient acute attack Ménière's Disease, network on induced different means.