Computerized eye-tracker assessments have demonstrated moderate to excellent reliability in distinguishing between healthy individuals, patients with mild traumatic brain injury, and those with persistent post-concussion syndrome, according to a recent study.
Published in Scientific Reports, study investigators evaluated computerized eye-tracker assessments (CEAs) among 55 healthy adults, 20 patients with mild traumatic brain injury (mTBI), and 40 patients with persistent post-concussion syndrome (PPCS). Researchers examined oculomotor function, visual attention, and selective attention through tests such as egocentric localization, fixation stability, smooth pursuit, saccades, Stroop tests, and the vestibulo-ocular reflex (VOR).
The CEAs demonstrated moderate to excellent test-retest reliability, with intraclass correlation coefficients ranging from 0.50 to 0.98 (P < .05). Machine learning models trained on CEA data achieved balanced accuracy scores of 0.83 for controls, 0.66 for mTBI, and 0.76 for PPCS, with an area under the curve of 0.82. Key diagnostic outcomes included VOR gaze stability, fixation stability (vertical error), and smooth pursuit metrics (total error, vertical gain, and number of saccades).
The researchers concluded that CEAs are reliable and effective in differentiating between healthy individuals, patients with mTBI, and those with PPCS. They recommend further validation with larger datasets to enhance diagnostic accuracy for clinical application.
The authors declared no competing interests.