LATERAL CONTROL IN A DRIVING SIMULATOR: CORRELATIONS WITH NEUROPSYCHOLOGICAL TESTS AND ON-ROAD SAFETY ERRORS.

Driving simulators provide precise information on vehicular position at high capture rates. To analyze such data, we have previously proposed a time series model that reduces lateral position data into several parameters for measuring lateral control, and have shown that these parameters can detect differences between neurologically impaired and healthy drivers (Dawson et al, 2010a). In this paper, we focus on the "re-centering" parameter of this model, and test whether the parameter estimates are associated with off-road neuropsychological tests and/or with on-road safety errors. We assessed such correlations in 127 neurologically healthy drivers, ages 40 to 89. We found that our re-centering parameter had significant correlations with five neuropsychological tests: Judgment of Line Orientation (r = 0.38), Block Design (r = 0.27), Contrast Sensitivity (r = 0.31), Near Visual Acuity (r = -0.26), and Grooved Pegboard (r = -0.25). We also found that our re-centering parameter was associated with on-road safety errors at stop signs (r = -0.34) and on-road safety errors during turns (r = -0.22). These results suggest that our re-centering parameter may be a useful tool for measuring and monitoring ability to maintain vehicular lateral control. As GPS-based technology continues to improve in precision and reliability to measure vehicular positioning, our time-series model may potentially be applied as an automated index of driver performance in real world settings that is sensitive to cognitive decline. This work was supported by NIH/NIA awards AG17177, AG15071, and NS044930, and by a scholarship from Nissan Motor Company.