Abstract Details

Title Comparison of Calibration vs Non-calibration Techniques in the Automated Capture Of Eye Movement Data: Initial Validation of the ROADIE Device

Topic Neuro-ophthalmology/Neuro-otology

Presentation(s) Neuro-ophthalmology/Neuro-otology Posters (7:00 AM-5:00 PM)

Poster/Presentation
Number 021

Background Abnormal extraocular movements, including dysconjugacies and nystagmus, are common signs of neurologic disease and difficult to assess for both neurologists and non-neurologists, particularly in the emergent setting.

Objective Develop and validate an automated screening tool to detect abnormal eye movements using machine learning and computer vision analysis.

Design/Methods As an initial step, we built a portable platform called RoADIE (Rolling Apparatus to Detect Impairment of the Eyes), equipped with eye-tracking software and camera visualization to acquire gaze data from patients presenting to the emergency department with acute vestibular syndrome. We first performed a validation study in 19 healthy controls comparing calibration (C) vs non-calibration (NC) techniques. Eye movements were captured using a standard H-Test exam. The NC condition was performed first for each volunteer to avoid a learning effect. Correlation between C and NC tests was determined using the Spearman coefficient (r).

Results Conjugate gaze (i.e. ability of eyes to track in unison) demonstrated strong correlation along the horizontal [r = 0.976 (C), 0.922 (NC)] and vertical axes [r= 0.866 (C), 0.881(NC)]. Smooth pursuit (i.e. ability of each eye to track a moving target) also showed strong correlation in the horizontal plane: right eye [r = 0.945 (C), 0.946 (NC), left eye [r=0.945 (C), 0.943 (NC)]. Vertical tracking showed moderate correlation: right eye [r=0.652 (C), 0.575 (NC), left eye [r=0.678 (C), 0.550 (NC)].

Conclusions In this initial validation test of a portable eye tracking platform, we demonstrated strong correlation for both conjugate gaze and smooth pursuit in the horizontal axis, and moderate correlation for eye tracking in the vertical axis. Given similar performance in non-calibrated tests, a calibration procedure may not be necessary for future data capture. Next steps will include validation of the eye tracking device in prospective patients presenting with acute vestibular syndrome.

Authors/Disclosures
Andrew M. Southerland, MD, FAAN (University of Virginia, Dept of Neurology) PRESENTER	Dr. Southerland has received personal compensation in the range of $10,000-$49,999 for serving as an Expert Witness for Plaintiffs and Defense Cases related to Stroke and Vascular Neurology.. The institution of Dr. Southerland has received research support from American Heart Association. The institution of Dr. Southerland has received research support from NIH. The institution of Dr. Southerland has received research support from Abbvie Pharmaceuticals, Inc.. Dr. Southerland has received intellectual property interests from a discovery or technology relating to health care. Dr. Southerland has received intellectual property interests from a discovery or technology relating to health care.
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file