<ul><li>A novel imputation framework is proposed using Self-Attention-based Imputation networks for time series data, specifically addressing missing data in smooth pursuit eye movements common in biomedical sequences.</li><li>The approach leverages deep learning and self-attention mechanisms to impute missing data, further refining the imputed data using a custom autoencoder tailored for smooth pursuit eye movement sequences.</li><li>Implemented on sequences from 172 Parkinsonian patients and healthy controls, the method showed significant improvement in reconstructing eye movement sequences, reducing common time domain error metrics and preserving frequency domain characteristics.</li><li>This method provides a robust solution for handling missing data in time series, enhancing the reliability of smooth pursuit analysis for screening and monitoring neurodegenerative disorders.</li></ul>

Imputation of Missing Data in Smooth Pursuit Eye Movements Using a Self-Attention-based Deep Learning Approach

Discover more