Declines in executive function and dual-task performance have been related to falls in older adults, and recent research suggests that older adults at risk for falls also show impairments on real-world tasks, such as crossing a street. The present study examined whether falls risk was associated with driving performance in a high-fidelity simulator. Participants were classified as high or low falls risk using the Physiological Profile Assessment and completed a number of challenging simulated driving assessments in which they responded quickly to unexpected events. High falls risk drivers had slower response times (~2.1 seconds) to unexpected events compared to low falls risk drivers (~1.7 seconds). Furthermore, when asked to perform a concurrent cognitive task while driving, high falls risk drivers showed greater costs to secondary task performance than did low falls risk drivers, and low falls risk older adults also outperformed high falls risk older adults on a computer-based measure of dual-task performance. Our results suggest that attentional differences between high and low falls risk older adults extend to simulated driving performance.
Per mile driven, adults over age 65 are more likely to be involved in motor vehicle collisions than are younger experienced drivers [
Older adults often have increased difficulty when multitasking, including paradigms that involve balancing or walking [
Differences in multitasking ability that have been associated with falls risk are theorized to result from declines in executive control, the functions which select, schedule, and coordinate task processes. Low falls risk older adults outperform high falls risk adults on tasks theorized to index executive control abilities [
Executive control is also important for other real-world tasks, such as driving. Drivers must attend to several areas of the environment and plan and execute responses to avoid collisions. Indeed, poorer performance on executive control tasks is predictive of retrospective crashes in a sample of older male drivers [
The goal of the present study was to explore the relationship between falls risk and driving in older adults in greater detail using a high-fidelity driving simulator, which allowed us to place drivers in potentially dangerous situations and to collect objective performance measures. We also included a battery of cognitive tasks to examine the relationship between falls risk, cognition, and simulated driving. Given previous findings of heightened crash risk, we predicted that low falls risk drivers would outperform high falls risk drivers on our simulator driving assessments. We further predicted that high falls risk drivers would show the greatest performance decrements in simulated driving performance under high multitasking load (i.e., when responding to unexpected events). Finally, we predicted that low falls risk older adults would outperform high falls risk adults on a desktop computer dual-task paradigm.
36 independent-living older adults were recruited from the Urbana-Champaign community and paid $8 per hour for participating. All participants demonstrated normal or corrected-to-normal visual acuity (20/30 or better using a Snellen chart) and normal color vision (Ishihara Color Vision Test) and scored above 28 (of 30) on the Folstein minimental state exam. All participants had valid drivers’ licenses and drove regularly. Mobility and balance were assessed using the Timed Up and Go test (TUAG) [
Demographic and cognitive measures.
Measure | High falls risk | Low falls risk |
---|---|---|
( |
( |
|
Age (years) | 75.8 (3.3) | 74.4 (5.5) |
Physiological Profile Assessment score** | 1.67 (.64) | .33 (.26) |
Timed up and go (seconds)** | 13.94 (2.6) | 10.18 (2.2) |
Activities balance confidence score (of 16)* | 14.17 (1.4) | 15.43 (.42) |
Miles driven per week | 55.36 (9.9) | 61.42 (7.0) |
Years licensed | 58.86 (5.3) | 57.86 (3.2) |
Crashes in last 12 months | 3 | 2 |
FFOV Accuracy (% Correct) | 44.29 (19.9) | 47.13 (18.7) |
Flicker CD RT (s) | 7.40 (1.13) | 7.70 (1.28) |
Flicker CD Accuracy (% Correct) | 53.7 (10.3) | 55.39 (8.4) |
Computer dual-task cost (ms)* | 572.7 (207.2) | 354.8 (207.8) |
Collisions | 8 | 6 |
Data expressed as mean (SD).
*
FFOV: functional field of view.
CD: change detection.
RT: response time.
Dual-task cost = dual RT − single RT.
The Beckman Institute Driving Simulator at the University of Illinois (
Participants completed a falls history questionnaire (i.e., “have you fallen in the last 6 months?” “how many times?”). Only three individuals reported falling in the previous 6 months. Thus, we classified participants as high or low falls risk based on scores from the Physiological Profile Assessment (PPA), as described by Lord and colleagues [
Participants performed two tasks both separately and simultaneously. For one task, participants determined whether a letter was an A or B and pressed corresponding keys with their right hand. In the second task, participants determined whether a number was a 2 or 3 and pressed a corresponding key with their left hand. On single-task trials (50%), participants performed only one task. On dual-task trials (50%), they performed both tasks. The primary performance measure was reaction time. Participants completed single-task and dual-task practice trials, followed by a block of forty intermixed single- and dual-task test trials.
Participants searched for a white triangle within a circle among square distracters in a briefly (44 ms) presented display. Targets were presented with equal probability on one of 8 radial spokes at eccentricities of 10°, 20°, and 30° from fixation. The search display was followed by a 100 ms mask consisting of random black and white lines and shapes. Participants then clicked with the mouse on the spoke where the target appeared. The percentage of targets correctly localized was the critical measure of performance. This task is similar to the peripheral localization subtask of the Useful Field of View [
Participants performed a flicker change detection task [
Drivers followed a lead vehicle (LV) along a straight, two-lane highway for approximately 15 minutes. Participants were instructed to maintain a 5-second gap from the LV, which traveled at 45 mph. During the practice drive, participants received auditory feedback to help visualize the 5-second gap. At 20 random times during the test drives, the LV’s brake lights illuminated and its speed decreased. Drivers were instructed to brake as soon they detected LV slowing. When the driver pressed the brake, the LV accelerated back to 45 mph. Performance measures included response time to LV braking, following distance, and lane keeping.
Drivers responded to potentially hazardous events as they drove along a straight, two-lane urban road for approximately 15 minutes. Ambient traffic and pedestrians were randomly generated such that there was a constant stream of traffic in the opposite lane, and the sidewalks were crowded with pedestrians. Participants were instructed to maintain a speed of 35 mph. There were a total of 20 randomly spaced potential hazards in each drive. Hazards comprised pedestrians crossing the roadway and cars on the right shoulder beginning to pull out and stopping (Figure
Examples of potential hazards in the hazard driving task. In (a), a pedestrian crosses the street in front of the driver. In (b), a parked vehicle starts to pull out in front of the driver.
Participants completed two versions of two separate simulated driving tasks. The order of the task conditions was counterbalanced across participants. In the drive-only condition, participants drove without secondary task distraction. In the drive + 1-Back condition, participants performed cognitively demanding secondary task, a continuous 1-Back task where they heard a letter every 3 seconds and indicated whether the letter was the same as or different from the previous letter via buttons on the steering wheel, while driving. Accuracy was considered the primary measure of secondary task performance.
Participants completed three 1.5-hour sessions. Session 1 consisted of a screening drive for simulator sickness, descriptive measures, and falls risk assessment (6 potential participants showed signs of simulator sickness and were not included in the study). In sessions 2 and 3, participants completed the three computer-based cognitive tasks, followed by practice with the secondary task and two driving assessments in the simulator. The order of the cognitive tasks and driving task conditions was counterbalanced across subjects.
Three participants (2 high falls risk and 1 low falls risk) did not complete the cognitive battery (due to technical issues) and were not included in analyses of the cognitive tasks. Dual-task cost was calculated by subtracting the single-task reaction time from the dual-task reaction time. High falls risk participants had a significantly higher dual-task cost compared to the low falls risk group,
Analyses were performed separately for each driving task. Driving measures were entered into an ANOVA with falls risk group (high versus low) as a between-subjects factor and task condition (drive-only versus drive + 1-Back) as a within-subjects factor. Two participants (1 high falls risk and 1 low falls risk) who had passed the screening drive showed signs of simulator sickness during the experimental drives, did not complete the study, and were excluded from analyses.
Collisions were infrequent in both driving tasks (Table
Brake response times. Mean brake response time in seconds for the high and low falls risk driver groups in drive-only and drive + 1-Back task conditions in the hazard response and car following paradigms. Error bars represent one standard error of the mean. *
We ran separate analyses using hand reaction time, contrast sensitivity, and the combination of hand reaction time and contrast sensitivity as covariates to investigate the impact of specific components of the PPA [
High and low falls risk drivers did not differ in average velocity or lane keeping performance (all
To examine whether performance on the computer dual-task paradigm predicted simulated driving, we computed the correlation between dual-task cost on the computer paradigm and RT in the driving tasks. Participants with a lower computer dual-task cost responded faster to both LV braking events (
Driving response times and dual-task performance. Response time in seconds in the hazard (a) and (c) and following (b) and (d) driving tasks plotted against single-task and dual-task reaction time in milliseconds on the computer dual-task paradigm. *
We compared accuracy on the auditory 1-Back task during 1-Back only (during the last half of practice) and 1-Back + driving performances to examine whether there were costs to secondary task performance when driving (Figure
1-Back accuracy. Accuracy on the 1-Back task in critical and noncritical segments of the hazard and following drives and in the single-task (1-Back only) condition. Error bars represent one standard error of the mean. *
The current study compared the driving performance of high and low falls risk older adults in a high-fidelity driving simulator. Of greatest importance is the finding that high falls risk drivers responded approximately 400 ms slower than did low falls risk drivers to critical events. High falls risk drivers responded slower than did low falls risk drivers to both central lead vehicle braking events and to peripheral hazards. Such slower responses may be a contributing factor to heightened crash rates for high falls risk older adults reported elsewhere [
Our data extend the literature that examines multitasking performance in older adults at high and low risk for falls [
Our results indicate that contrast sensitivity and response time were the most important components of the PPA relating to simulated driving RT. Previous work has found that contrast sensitivity and response time are important abilities in responding to driving hazards [
We failed to find differences between high and low falls risk drivers on other simulator driving performance measures such as lane keeping. Hazard responses posed the highest multitasking demand in our driving assessments. Previous research has shown that multitasking differences in older adults and differences between high and low falls risk older adults arise primarily at the highest levels of task demand [
Future work should explore the contribution of different components of executive control (e.g., switching and inhibition) to deficits in real-world tasks such as walking and driving. Eye tracking techniques could inform as to whether high and low falls risk drivers differ in the way they deploy attention within a driving scene. Research should also explore other driving tasks where older adults are differentially involved in crashes, such as busy intersections [
In summary, our results demonstrate that high falls risk older drivers respond slower than do low falls risk drivers when responding to potential dangers in a driving simulator. A multidimensional approach that includes falls risk may be useful in more accurately assessing older driver impairment.
The authors would like to thank the Illinois Simulator Lab for their support and J. C. and K. L. for subject running. A subset of this data was presented at the Cognitive Aging Conference in Atlanta, in April 2012.