•  
  •  
 

CONVERGENT VALIDITY AND STABILITY OF A 5-MINUTE WEB CAMERA-BASED EYE-TRACKING COGNITIVE ASSESSMENT

Abstract

Joshua Gills1, Spencer Smith1, Emily Bates1, Jordan M. Glenn2, Erica N. Madero2, Nick T. Bott2, & Michelle Gray1

1University of Arkansas, Fayetteville, AR., 2Neurotrack Technologies, Inc., Redwood City, CA.

As populations age worldwide, dementia prevalence is projected to triple from current rates to 132 million by 2050. While there is no cure for Alzheimer’s disease (AD) or other forms of dementia, early detection of symptoms allow treatment to start earlier and improve outcomes. Currently, there exists a noninvasive validated 30-minute (min) eye-tracking cognitive assessment for predicting AD risk. However, the time requirements and passive nature of the paradigm creates a burden for the user. A shorter task utilizing an active paradigm would improve user experience and increase scalability of the test. PURPOSE: The purposes of this study were to 1) determine convergent validity of an active 5-min web camera-based eye-tracking task to measure visual recognition memory compared to the validated passive 30-min task and 2) determine the stability and test-retest reliability of the 5-min test. METHODS: This prospective study included 44 cognitively intact participants (n = 28 females, n = 16 males; age = 50.0 ± 27.6) who were divided into two cohorts: older adults (ages 65+ years, n = 20) and young adults (ages 18 – 46 years, n = 24). Participants reported for testing on two separate occasions. The first visit included informed consent, medical history questionnaire, Montreal Cognitive Assessment (MOCA), 30-min eye tracking test, and 5-min eye tracking test. The second testing session occurred at least 14 days later and participants were given an alternate form of the 5-min eye tracking test to minimize learning effects. RESULTS: Participants were cognitively normal based on MOCA scores (27.9 ± 1.4). A Pearson’s correlation determined the 30-min task was moderately correlated with the 5 min task at the first (r =.55; p < .001) and second (r =.58; p = .001) time points. Moreover, there was a high test re-test reliability of the 5-min test (r =.73; p < .001). CONCLUSION: The active 5-min eye-tracking assessment displayed moderate convergent validity to the passive 30-min test for assessing working memory and demonstrated strong test-retest reliability. Initial data indicate the 5-min version of the eye-tracking task may be a more scalable alternative to the original 30-min version. However further research is needed to ultimately substantiate this claim.

ACKNOWLEDGEMENTS: This study was funded by Neurotrack Technologies and The University of Arkansas’ Department of Health, Human Performance, and Recreation.

This document is currently not available here.

Share

COinS