Nystagmus and Fixations in Eye-Tracker Data
Georg Heller · 2024
Analyzing the waveforms of nystagmus from eye-tracking records is essential for the clinical interpretation of this abnormal eye movement. A significant challenge in automating this analysis lies in the interference of natural eye movements and eye blink artifacts with the signals of interest. We propose a method utilizing Convolutional Dictionary Learning to effectively distinguish between nystagmus waveforms and natural motion. Our simulations demonstrate that this approach can enhance pattern recovery, and we provide clinical examples to showcase its performance. Furthermore, the assessment of nystagmus in patients with neurological disorders is vital for accurate differential diagnosis, as highlighted by numerous studies in neuro-otology and neuro-ophthalmology, which focus on various types of nystagmus like vestibular, positional, and optokinetic. Regular updates on the diagnostic and therapeutic relevance of these findings are necessary. Critical aspects of clinical practice include a detailed description of nystagmus features such as its three-dimensional beating direction, triggers, and duration. The differential diagnosis for conditions like downbeat nystagmus is extensive, encompassing acute intoxications, neurodegenerative disorders, and cerebrovascular issues, while accurately differentiating between common benign and rare but serious central causes in positional nystagmus cases is crucial.