This article explores the potential of using non-calibrated eye-tracking technology to enhance stroke diagnosis. The study investigates the need for calibration to measure eye movement symmetry in healthy controls and the potential of eye movement symmetry to distinguish between healthy controls and patients with neuro-ocular abnormalities. The calibration procedure posed an obstacle to the clinical translation of the eye tracker to diagnose stroke, as patients may have difficulty fixating on a target or following a moving object.
The study found that implementing off-the-shelf eye-tracking technology without calibration is feasible in healthy adults and those with neurological injuries. However, none of the enrolled patients could complete the calibration procedure, which is concerning as current eye-tracking research excludes patients unable to perform a calibration procedure. This opens the door to utilizing a machine learning tool to learn relevant features of normal and abnormal eye movements. Machine learning techniques could potentially be used to assess and diagnose neuro-ocular abnormalities without a calibration procedure, as long as the analysis is based on relative and correlative eye measurements.
The preliminary findings suggest that machine learning techniques could be used to assess and diagnose neuro-ocular abnormalities without needing a calibration procedure. If successful, this approach could provide an alternative method for assessing patients who are unable to perform traditional calibration procedures due to neurological deficits or other factors. In conclusion, this study provides evidence that non-calibrated eye-tracking technology may have the potential to enhance stroke diagnosis. If successful, this approach could provide an alternative method to augment stroke diagnosis as part of clinical and pre-clinical care for patients with ocular abnormalities.