Safety and Usefulness of Automated Driving for Older Adults with Cognitive Impairment

Research Objectives: The objective of our research was to compare the driving performance and cognitive workload (CW) of drivers with cognitive impairment (CI) to control drivers during two driving scenarios: (1) conditional automated driving and (2) when manual take-over is required due to automation failure.

Design: Cross-sectional, observational study.

Setting: University of Kansas Medical Center.

Participants: Out of the 35 participants which were recruited to the study, 28 (70%) were women. 21 were characterized as healthy controls based on their PET scan results, and 14 categorized as CI. Participants were on average 73.37 ± 5.64 years old and scored on average 26.31 ± 4.1 on the MOCA.

Interventions: Not applicable.

Main Outcome Measures: Participants completed two driving scenarios. Scenario-1 included conditional automated driving with a sudden emergency event that required manual take-over, while scenario-2 had the same events but with an additional cognitive task (2-back test). Response time to the emergency event and cognitive workload was the outcomes. The alteration of the pupil's size is transformed into an Index of Cognitive Activity (ICA) to measure cognitive workload. Two sample-independent t-tests, Chi-square, and Mann-Whitney U tests were used to compare differences between groups.

Results: Drivers with CI took longer to respond to the emergency event during conditional automated driving with 2-back compared to controls (13.45s ± 10.43 vs. 17.2s ± 7.04), (p=0.046). No statistical differences were found in emergency response times between drivers with CI compared to the controls during the manual take-over without the 2-back (20.52s ± 3.18 vs. 21.73s ± 11.2, p = 0.31). Likewise, no statistical differences were found in CW.

Conclusions: Our pilot study found that drivers with CI may not have the necessary skills to safely take control of the car after automation failure, especially when distracted. We suggest that future studies are necessary to identify the best predictors of conditional automated driving in older adults with CI.

Author(s) Disclosures: Dr. Akinwuntan and Dr. Devos receive royalties from the sales of the portable driving simulator. All other authors do not report conflicts of interest.