Multimodal immersive trail making-virtual reality paradigm to study cognitive-motor interactions

BACKGROUND: Neuropsychological tests of executive function have limited real-world predictive and functional relevance. An emerging solution for this limitation is to adapt the tests for implementation in virtual reality (VR). We thus developed two VR-based versions of the classic Color-Trails Test (CTT), a well-validated pencil-and-paper executive function test assessing sustained (Trails A) and divided (Trails B) attention—one for a large-scale VR system (DOME-CTT) and the other for a portable head-mount display VR system (HMD-CTT). We then evaluated construct validity, test–retest reliability, and age-related discriminant validity of the VR-based versions and explored effects on motor function. METHODS: Healthy adults (n = 147) in three age groups (young: n = 50; middle-aged: n = 80; older: n = 17) participated. All participants were administered the original CTT, some completing the DOME-CTT (14 young, 29 middle-aged) and the rest completing the HMD-CTT. Primary outcomes were Trails A and B completion times (t(A), t(B)). Spatiotemporal characteristics of upper-limb reaching movements during VR test performance were reconstructed from motion capture data. Statistics included correlations and repeated measures analysis of variance. RESULTS: Construct validity was substantiated by moderate correlations between the’gold standard’ pencil-and-paper CTT and the VR adaptations (DOME-CTT: t(A) 0.58, t(B) 0.71; HMD-CTT: t(A) 0.62, t(B) 0.69). VR versions showed relatively high test–retest reliability (intraclass correlation; VR: t(A) 0.60–0.75, t(B) 0.59–0.89; original: t(A) 0.75–0.85, t(B) 0.77–0.80) and discriminant validity (area under the curve; VR: t(A) 0.70–0.92, t(B) 0.71–0.92; original: t(A) 0.73–0.95, t(B) 0.77–0.95). VR completion times were longer than for the original pencil-and-paper test; completion times were longer with advanced age. Compared with Trails A, Trails B target-to-target VR hand trajectories were characterized by delayed, more erratic acceleration and deceleration, consistent with the greater executive function demands of divided vs. sustained attention; acceleration onset later for older participants. CONCLUSIONS: The present study demonstrates the feasibility and validity of converting a neuropsychological test from two-dimensional pencil-and-paper to three-dimensional VR-based format while preserving core neuropsychological task features. Findings on the spatiotemporal morphology of motor planning/execution during the cognitive tasks may lead to multimodal analysis methods that enrich the ecological validity of VR-based neuropsychological testing, representing a novel paradigm for studying cognitive-motor interactions. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12984-021-00849-9.