Publication Abstract Display
Type: Poster
Title: The Relationship Between Contextual Factors, Performance, and Validity of Smartphone-Based Mobile Cognitive Tests of Executive Function and Learning.
Authors: Campbell LM, Parrish E, Heaton A, Swendsen J, Depp CA, Moore RC
Date: 02-04-2021
Abstract:Abstract: Objective: Smartphone-based mobile cognitive tests (MCT) are new assessment tools developed to aid in longitudinal tracking of cognitive functioning and assessing cognitive performance in natural environments. However, we currently have little research examining how contextual factors such as distractions, interruptions, presence of other people, and location impacts MCT performance. Therefore, we examined how these contextual factors relate to MCT performance and convergent validity with lab-based neuropsychological tests. Participants and Methods: Eighty-four adults (54 persons with HIV, 30 HIV-negative adults) aged 50-74 (Mean age=59, 70% male, 67% non-Hispanic white) completed in-person neuropsychological testing, including a test of executive functioning (Stroop Color Word Interference Test) and verbal list-learning (Hopkins Verbal Learning Test-Revised [HVLT-R]). Participants also completed the UCSD Performance-Based Skills Assessment-Brief (UPSA-B). Following the in-person assessment, participants completed MCTs of inhibition (mobile Color Word Interference Test [mCWIT]; score=time to complete) and verbal list-learning (mobile Verbal Learning Test [mVLT]; score=total words correctly recalled over three learning trials) for 14 days. Participants also self-reported if they were distracted or interrupted (combined), if they were with other people, and their current location (home vs. not at home). Matched-pairs Wilcoxon tests were used to examine how context impacted average test performance, and Spearman correlations were used to understand how different contexts impacted convergent validity of the MCTs with lab-based tests. Results: Presence of a distraction negatively impacted mCWIT performance (no distraction: M=22.3 seconds, distraction: M=24.7 seconds, p<0.01). However, the presence of others (alone: M=22.2 seconds, with people: M=23.3 seconds, p=0.07) or being away from home (home: M=22.6 seconds, not at home: M=21.9 seconds, p=0.07) did not significantly impact mCWIT performance. On the mVLT, presence of a distraction (no distraction: M=21.2 words, distraction: M=18.6 words, p<0.01), being with others (alone: M=20.6 words, with people: M=19.3 words, p<0.01), or being away from home (at home: M=20.8 words, not at home: M=19.6 words, p<0.01) negatively impacted performance. Performance on the mCWIT was more highly correlated with the in-person Color Word Interference Test when not distracted (ρ=-0.68 vs. distracted ρ=-0.56), but was not significantly impacted by being away from home (ρ=-0.64 vs. home ρ=-0.70) or presence of others (ρ=-0.67 vs. alone ρ=-0.66). UPSA-B performance was more highly correlated with mCWIT performance when not distracted (ρ=-0.32 vs. distracted ρ=-0.12) but the correlation did not differ when home vs. not at home or alone vs. with people. The relationship between the mVLT and in-person HVLT-R was less impacted by contextual factors (no distraction ρ=0.56 vs. distraction ρ=0.52; alone ρ=0.49 vs. with others ρ=0.50; home ρ=0.55 vs. not at home ρ=0.49). Interestingly, mVTL performance was somewhat more correlated with UPSA-B performance when participants were not at home (ρ=0.38) as compared to when they were at home (ρ=0.28), but was not as impacted by distractions (ρ=0.30 vs. no distractions ρ=0.31) or presence of others (ρ=0.34 vs. alone ρ=0.29). Conclusions: Contextual factors impact MCT performance and convergent validity with in-lab neuropsychological testing, but these contextual factors vary by mobile task. Additionally, some contextual factors such as current location may provide better insight into real-world functioning.

return to publications listing