INS NYC 2024 Program

Poster	Poster Session 04 Program Schedule 02/15/2024 12:00 pm - 01:15 pm Room: Shubert Complex (Posters 1-60) Poster Session 04: Neuroimaging \| Neurostimulation/Neuromodulation \| Teleneuropsychology/Technology Final Abstract #33 Can Virtual Reality Driving Enhance the Prediction of Real-World Unsafe Driving? Kathryn Devlin, Drexel University, Philadelphia, United States Rachel Lyons, Drexel University, Philadelphia, United States Molly Split, Drexel University, Philadelphia, United States Jocelyn Ang, Drexel University, Philadelphia, United States Maria Schultheis, Drexel University, Philadelphia, United States Category: Teleneuropsychology/ Technology Keyword 1: driving Keyword 2: technology Keyword 3: ecological validity Objective: Clinical driving evaluations typically include neuropsychological tests, which provide valuable information about the cognitive capacities that relate to driving risk, and a behind-the-wheel evaluation. Virtual reality driving simulation (VRDS) can complement such measures by providing information that is objective, sensitive, challenging, and functionally relevant. While VRDS metrics have been associated with driver records and self/informant reports, research on their relation to directly observed naturalistic driving behaviors is in its infancy. Furthermore, its added value over traditional driving evaluation methods is unknown. The present study examined the ability of virtual reality driving simulator (VRDS) performance to predict naturalistic risky driving behaviors in healthy adults, and its added value beyond neuropsychological measures. Participants and Methods: Twenty-five neurologically healthy drivers (ages 23-61, 68% women) were recruited from the general community. They completed neuropsychological testing and a VRDS drive, followed by 28 days of naturalistic driving with an in-vehicle video telematics platform that detected unsafe driving behaviors. The neuropsychological battery measured driving-relevant domains, namely basic attention, complex attention, processing speed, executive function, and visuospatial memory. The VRDS drive included driving on a straight rural road with and without an external distractor task, and four driving challenge scenarios: following a truck on a highway, stopping for a school bus, reacting to a child running into a residential street to retrieve a ball, and reacting to a car pulling into a commercial street. The naturalistic driving measure of interest was the number of unsafe driving behaviors per hour driven. Hierarchical linear regression analyses were conducted to examine the independent and shared contributions of neuropsychological and VRDS measures in relation to naturalistic unsafe driving behaviors. Results: VRDS metrics alone, namely, straight rural road lane position and truck-following speed, significantly predicted real-world unsafe driving behaviors, explaining a large proportion of their variance (p = .007, adjusted R-squared = 32%). The same was true of neuropsychological measures (p = .004, adjusted R-squared = 36%), namely, tests of psychomotor speed/executive function (Trail Making Test B) and immediate visuospatial recall (10/36 Spatial Recall Test). Together, VRDS and neuropsychological measures explained 41% of the variance in driving behaviors (p=.006), 27% of which was shared. VRDS metrics (p=.177) uniquely explained 5%, while neuropsychological measures uniquely explained 9% (p=.097). Conclusions: Findings indicate that both virtual-reality-simulated driving performance and neuropsychological measures provide unique, ecologically valid information about real-world unsafe driving behaviors. While statistical power is limited by the small sample size, results suggest that VR-simulated performance may enhance predictive value beyond traditional neuropsychological measures. These novel findings provide preliminary support for incorporating VRDS assessment when evaluating driving capacity. Future research will expand these investigations to larger samples and clinical populations, such as acquired brain injury. Identifying novel approaches for improving prediction of driving capacity can optimize the balance of safety and independence in people affected by brain injury or illness.

Poster

02/15/2024
12:00 pm - 01:15 pm
Room: Shubert Complex (Posters 1-60)

Poster Session 04: Neuroimaging | Neurostimulation/Neuromodulation | Teleneuropsychology/Technology

Final Abstract #33

Can Virtual Reality Driving Enhance the Prediction of Real-World Unsafe Driving?

Kathryn Devlin, Drexel University, Philadelphia, United States
Rachel Lyons, Drexel University, Philadelphia, United States
Molly Split, Drexel University, Philadelphia, United States
Jocelyn Ang, Drexel University, Philadelphia, United States
Maria Schultheis, Drexel University, Philadelphia, United States

Category: Teleneuropsychology/ Technology

Keyword 1: driving
Keyword 2: technology
Keyword 3: ecological validity

Objective:

Clinical driving evaluations typically include neuropsychological tests, which provide valuable information about the cognitive capacities that relate to driving risk, and a behind-the-wheel evaluation. Virtual reality driving simulation (VRDS) can complement such measures by providing information that is objective, sensitive, challenging, and functionally relevant. While VRDS metrics have been associated with driver records and self/informant reports, research on their relation to directly observed naturalistic driving behaviors is in its infancy. Furthermore, its added value over traditional driving evaluation methods is unknown. The present study examined the ability of virtual reality driving simulator (VRDS) performance to predict naturalistic risky driving behaviors in healthy adults, and its added value beyond neuropsychological measures.

Participants and Methods:

Twenty-five neurologically healthy drivers (ages 23-61, 68% women) were recruited from the general community. They completed neuropsychological testing and a VRDS drive, followed by 28 days of naturalistic driving with an in-vehicle video telematics platform that detected unsafe driving behaviors. The neuropsychological battery measured driving-relevant domains, namely basic attention, complex attention, processing speed, executive function, and visuospatial memory. The VRDS drive included driving on a straight rural road with and without an external distractor task, and four driving challenge scenarios: following a truck on a highway, stopping for a school bus, reacting to a child running into a residential street to retrieve a ball, and reacting to a car pulling into a commercial street. The naturalistic driving measure of interest was the number of unsafe driving behaviors per hour driven. Hierarchical linear regression analyses were conducted to examine the independent and shared contributions of neuropsychological and VRDS measures in relation to naturalistic unsafe driving behaviors.

Results:

VRDS metrics alone, namely, straight rural road lane position and truck-following speed, significantly predicted real-world unsafe driving behaviors, explaining a large proportion of their variance (p = .007, adjusted R-squared = 32%). The same was true of neuropsychological measures (p = .004, adjusted R-squared = 36%), namely, tests of psychomotor speed/executive function (Trail Making Test B) and immediate visuospatial recall (10/36 Spatial Recall Test). Together, VRDS and neuropsychological measures explained 41% of the variance in driving behaviors (p=.006), 27% of which was shared. VRDS metrics (p=.177) uniquely explained 5%, while neuropsychological measures uniquely explained 9% (p=.097).

Conclusions:

Findings indicate that both virtual-reality-simulated driving performance and neuropsychological measures provide unique, ecologically valid information about real-world unsafe driving behaviors. While statistical power is limited by the small sample size, results suggest that VR-simulated performance may enhance predictive value beyond traditional neuropsychological measures. These novel findings provide preliminary support for incorporating VRDS assessment when evaluating driving capacity. Future research will expand these investigations to larger samples and clinical populations, such as acquired brain injury. Identifying novel approaches for improving prediction of driving capacity can optimize the balance of safety and independence in people affected by brain injury or illness.