Cognitive Workload and Emergency Response during Partially Automated Driving in Preclinical Alzheimer's Disease

Research Objectives

Our aim was to compare driving performance and cognitive workload (CW) between drivers with preclinical Alzheimer's Disease (AD) and control drivers during (1) partially 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

Forty participants (28 females) with normal cognition (CDR: 0) were included. Nineteen participants [age 74.16 ±4.78; MOCA 26.42 ±2.52] had preclinical AD characterized by increased cortical beta-amyloid accumulation in PET scanning; and 21 controls [age 73.81 ±5.62; MOCA 28.24 ±1.67].

Interventions

Not applicable.

Main Outcome Measures

Participants completed two driving scenarios. Scenario 1 included partially automated driving with a sudden emergency event that required manual take-over; scenario 2 comprised the same events but with an additional cognitive distractor task (2-back test). Driving performance measures included brake response to the emergency event. CW was indexed using the Index of Cognitive Activity (ICA), a measure based on moment-to-moment changes in pupillary size. Two sample independent t-test, and Mann-Whitney U tests were used to compare differences between the groups.

Results

No differences were found between drivers with preclinical AD and controls on ICA during partially automated driving without 2 back (0.34 ±0.08 vs 0.33 ±0.17; p=0.74), and with 2-back (0.30 ±0.07 vs 0.29 ±0.17; p=0.93). Likewise, there were no differences in emergency response times between drivers with preclinical AD and controls during the manual take-over without 2-back (20.85 ±1.08 vs 20.52 ±3.18; p=0.83), and with 2-back (14.83 ±7.37 vs 13.45 ±10.43; p=0.92).

Conclusions

Our results indicated that there is no differences in CW and delayed response times to hazardous situations in partially automated vehicles between drivers with preclinical AD and controls. Our findings may have been influenced by intensity of the cognitive tasks, sample size, or use of a driving simulator, hence, future studies are warranted.

Author(s) Disclosures

Nothing to disclose.