Flexible Visualization of Eye Tracking Data to Aid Assessment of Visuospatial Neglect

Abstract

Traditional pen-and-paper tests used for assessing visuospatial neglect, a condition which may occur in stroke patients, provide limited amounts of data, often leaving it up to the clinician to fill in the gaps.
There is growing interest in applying VR to these situations by adapting traditional pen‑and‑paper tests for virtual environments. Combined with eye tracking, such systems generate far richer datasets, but also demand new tools to structure and visualize the increased data.
This article explores various visualization techniques of eye tracking data through a custom built system, and evaluates its utility with respect to neglect assessment by consulting experts within the field. Overall, the system was deemed highly useful for clinicians.
Static visualizations were found useful for quickly identifying patterns and areas of reduced attention by providing an overview of the attention allocation, while dynamic visualizations provide real-time insight into visual behavior, making them useful for investigating the temporal aspects of visuospatial neglect.
However, due to neglect assessment being a highly specialized field, the number of experts giving statements are limited, and further research is needed to ensure reliable and more broadly generalizable results regarding the evaluation of the usefulness of the visualization system.