Rxivist logo

Validation of Remote Testing using BrainCheck, a Computerized Neurocognitive Test

By Siao Ye, Bin Huang, Kevin Sun, Huy Phi, Reza Hosseini Ghomi

Posted 03 Jun 2020
medRxiv DOI: 10.1101/2020.06.01.20119289

Remote computerized neurocognitive testing (CNT) is a promising solution to have these assessments more accessible to a population facing a global pandemic and increased aging. BrainCheck (BC) is a CNT software available on iPhone, iPad, and computer browser, designed to fit the need for remote testing. Consistent measures across these varying platforms are necessary to ensure users have consistent cognitive assessments and results. We aimed to assess BC across all administration platforms and interactions to observe any differences in assessment performance. 75 participants were enrolled in the study and were divided into two groups: participants who took BC across multiple platforms and participants who took BC in both an administered and self-administered fashion. Here we found Stroop, Digit Symbol, and Trail A/B had significantly different performance across the platforms, while Flanker, Coordination, Matrix, Immediate and Delayed Recognitions did not. Also, we found that the test metrics did not show significant differences in performance between being administered and self-administering the test. We did observe quicker completion times during the second instance of the test when taken in quick succession (within a day apart, which would not be a typical clinical pattern) and despite this, composite scores did not change reflecting the resilience of BC to practice effects. In conclusion, our results demonstrate BC may be a robust, self-administered CNT solution with an appropriate adjustment for the platform used.

Download data

  • Downloaded 615 times
  • Download rankings, all-time:
    • Site-wide: 67,981
    • In neurology: 209
  • Year to date:
    • Site-wide: 80,727
  • Since beginning of last month:
    • Site-wide: 70,702

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide