Share this article on:

Aging and Time-Sharing in Highway Driving


Original Article

Purpose. Proper time-sharing—visual attention allocation—between the road view ahead and other targets is an essential requirement for safe driving, along with other visual and attentional performance. Earlier on-road research has shown that neurologic problems (Alzheimer disease, brain injury) impair time-sharing during in-car tasks. This study analyzed age effects on time-sharing performance.

Methods. Thirty participants in three age groups (mean age 22, 34, and 67 years) drove an instrumented car a trip of 350 km and performed an in-car visual search task with either a motor (keying) or vocal response. The frequency and duration of glances at the in-car targets, total time eyes off the road during task, speed, and lateral displacement of the car were recorded. The participants were also tested on a battery of cognitive tasks during the midway break.

Results. The elderly used a longer total time looking at the in-car display and they traveled a longer distance with eyes away from the road. The number of long (>2 sec) glances and the car’s lateral displacement on the road were larger among the elderly than the young drivers. The difference between the older and younger participants was larger when a motor (keying) response was required. The age effects were mediated by cognitive performance (best by the Trail Making A test) rather than by vision parameters.

Conclusion. Older drivers have difficulties in time-sharing in highway driving already at the age of 65 to 70 years.

Traffic Research Unit, Department of Psychology, University of Helsinki, Helsinki, Finland

Received December 1, 2004; accepted February 16, 2005.

Driving vision has traditionally been associated with basic visual functions like static and dynamic visual acuity, static visual field, and contrast sensitivity.1–3 More recently, functional field of vision and distractibility have increasingly been used in driving ability assessment.4 This means a step toward taking visual attention as an essential part of visual performance testing, which appears to be highly relevant for vision screening among aging drivers especially.5 Less attention has been devoted to the fact that drivers often search for information that is located at a substantial eccentricity from their driving path such as road signs, navigation information, intersecting vehicles, and pedestrians. Drivers also voluntarily share visual attention between in-car tasks and driving. Proper time-sharing is a necessary precondition of safe driving. To update traffic information often enough during in-car tasks, the driver should glance at the roadway at least every 1.5 seconds.6, 7

In time-sharing, it is the properly timed attention shift from one target to another that is critical for safety. As expected, our earlier on-road studies showed that drivers with Alzheimer disease8 and frontal brain lesion9 often fail in returning the gaze back from the in-car task in due time, indicating impaired control of action and major narrowing of the attention. These drivers simply forget the primary (driving) task and get stuck on secondary (in-car) tasks. The patients with more posterior brain damage and some patients with Alzheimer disease rather had difficulties in accomplishing a task, and they needed many short glances to complete it. In comparison with healthy control subjects, both strategies imply slower total performance and driving a longer time with eyes off the road, which means impaired lane keeping and traffic control.9

Aging has repeatedly been reported to lead to impaired performance in dual tasks resulting from general slowing in the speed of information processing,10, 11 possibly combined with impaired control processes needed for coordination and integration of multiple tasks.12–17 Rapid switching of attention from one task to another tends to impair with age.13, 18 Research on prospective memory also indicates that remembering to remember to act deteriorates with age, especially when the reliance on external cues is limited.19, 20 We can predict therefore that aging makes the timed switching from one task to another less reliable. General slowing with age also predicts that in-car tasks in general require more time and set additional requirements for proper time-sharing.

This study was aimed at analyzing age effects on time-sharing in highway driving. The focus was on switching of visual attention from roadway to target and back, keeping the additional tasks as simple and constant as possible.

Back to Top | Article Outline



Sixteen males and 14 women with a valid driver’s license participated. They were taken from the participant pool originally collected through newspaper announcements but partly completed from different sources. The inclusion criteria consisted of active driving, no known health or other problems to prevent driving, and willingness to prolonged highway driving. The participants were divided into three age groups, 20 to 24 years, 26 to 44 years, and 57 to 73 years. (Note: eight of 11 older drivers were within 65 and 69 years.) They had held the license between 2 and 49 years, very much depending on age of course, whereas the recent driving experience (last 12 mo) was fairly similar in each group (12, 16, and 13 thousands of km; see Table 1). The basic education was a little shorter in the older group (9.6 years vs. 11.17 years in the two younger groups) as a result of the change of the Finnish school system in the 1970s. The main results of cognitive tests are given in Table 2 and 3. No one was excluded based on cognitive test results. All participants had a visual field of at least 140° and corrected binocular decimal visual acuity of 0.5 or higher, both corresponding to the EU requirements. All the participants gave informed consent in writing, according to the Helsinki Declaration.

Back to Top | Article Outline


An instrumented midsized car equipped with four video cameras was used in this experiment. One video camera recorded the face of the driver, two cameras recorded the road in front using different focusing lengths, and one camera recorded the road behind the vehicle. The pictures from all four cameras, together with data from the vehicle sensors, were mixed onto the same video screen and stored on a videotape, whereas the sensor data (car controls, distance, speed, accelerations) were also stored on a computer file at 10 Hz. The car was equipped with dual brakes.

We used a specific digit display installed in the midconsole, 200 × 227 mm in size, midpoint located 30 cm down from the windscreen and at the average eccentricity of 36° to 37° from the vanishing point in horizon with no essential difference between males and females. There were eight separate digit displays (consisting of seven red light-emitting diode [LED] segments) arranged circularly as shown in Figure 1, 13 mm in height and located in round 2-mm deep hollows 25 mm in diameter. Each hollow also worked as a mechanical pushbutton with clear haptic feedback. The background was matt black. The digits could be seen in a visual angle of 1.05° in average, approximately 1.10° for males and 1.00° for females. The visual angle for the display area (the circle formed by outermost LEDs) was approximately 10° and 11° for males and females, respectively.

Back to Top | Article Outline

Additional Tasks

Two tasks were used, both using the same stimuli. In the first one, numbers 1 through 8 were randomly assigned to the eight number displays, and participants had to push all buttons consecutively in the ascending numeric order. As the participant had pushed the last number (8), the same numbers were immediately rearranged, and he or she continued pushing numbers consecutively. This sequence (later trial) was repeated until the experimenter stopped the task. The second task was similar but with a vocal response. The participants only had to read out all the eight numbers from left to right and from top down.

The pushbutton task simulated multifunction touch-screen displays, which are increasingly emerging into cars, whereas the reading task was intended to control arm and finger motor activity. The tasks were of varied mapping type designed not to allow development of automated mapping and blind routines. Therefore, they required continuous visual attention while being simple and easy to learn and remember. The task therefore forced participants to time-sharing between roadway and the visual display. The large eccentricity of the display also minimized the contribution of ambient vision to keeping the car in lane and, even to a large degree, to exploring hazards in front.21–23

Back to Top | Article Outline


After the preparations and vision tests at the department, the participants drove the instrumented car a trip of 350 km consisting of divided four-lane and undivided two-lane sections of main highway 7 and highway 140, east of Helsinki. Besides the type, the road sections differed in pavement width such that the divided motorway section had a hard shoulder of 2.5 m against 1.5 m of the undivided highway, and a speed limit of 120 km/h against 80 km/h of the undivided highway. The lane width was equal (3.75 m).

After 10 km of driving, the participants were given the in-car tasks, first on a divided motorway and thereafter on a two-lane highway. After performing the tasks, during a midway break, the participants were tested on a battery of cognitive tasks by a psychologist (ASW). The tests were performed after a 15-min rest in a quiet secluded room rented for this purpose.

All experiments were performed during the snow-free time of year. The participants began their trip at approximately the same time, at 10 to 11 am in the morning, so traffic was approximately similar and the route planned such that the position of the sun did not disturb the driver. The experimenter avoided giving tasks in demanding traffic situations. In cases in which these could not be avoided beforehand (e.g., when passing slower vehicles), these trials were excluded from the analysis.

Fifty trials of in-car tasks were done on average, in blocks of eight keying, 10 reading, and eight keying, both on the divided and undivided section. In some cases, however, oncoming traffic on two-lane roads reduced the number of trials, especially among participants with slower performance. Therefore, the average number of acceptable trials amounted to 56.8 in the young group, 51.8 in the middle-aged group, and 43.9 in the older group.

Back to Top | Article Outline

Cognitive Tests and Vision Parameters

All participants were tested on the same battery of cognitive tasks. It included the Mini-Mental State Examination (MMSE),24 Trail-Making A (TMT),25 and the subtests Digit Span, Block Design, and Digit Symbol from Wechsler Adult Intelligence Scale–Revised (WAIS-R).26 The tests were selected as being widely used and easy to administer paper-and-pencil tests, also used in assessing driving ability.27–30

Vision parameters included horizontal field of view, high-contrast visual acuity, and contrast sensitivity determined with LH symbols31 at a distance of 2, 4 and 8 m (6, 12, and 24 cycles/deg).

Back to Top | Article Outline


The direction and duration of the glances were analyzed frame by frame from the time-coded videotape. A forward-looking video picture was used to compute the lateral displacement during off-road glances.

The glance parameters determined per trial (pushing or reading numbers 1 through 8) included the average glance length away from road, the number of glances, total time eyes off the road, and the number of long offroad glances (>2 sec). The driving speed and lateral displacement of the car during each glance were also computed. Both the average and the maximum lateral displacement in lane—the worst case—were determined for each task and road type.

Back to Top | Article Outline


An analysis of variance with repeated measures was computed on each of the dependent variables, with age (three levels) and gender as between-subject factors and task (two levels) and road type (two levels) as within-subjects factors. Additionally, a multivariate analysis of variance (MANOVA) was applied, which showed that the overall effect of age on time-sharing parameters was significant (Wilks Lambda, F10,40 = 2.76; p = 0.011).

Figure 2 shows that, as expected, participants had somewhat different strategies to perform the task. Some of them used a few glances that tended to be longer, whereas the others segmented the task into many shorter glances. Different strategies imply that the driver groups do not clearly differ in mean glance length (p = 0.127) or in the number of glances (p = 0.058). It is easy to see, however, that younger drivers used less time with eyes at the display independently of the strategy.

Figure 3a shows the average total time the participants were looking off the road, accumulated from each separate glance, from the moment they turned the gaze from the roadway to the display until the moment the gaze had returned back to the road. There was a significant age main effect (F2,24 = 9.03; p = 0.001) such that the elderly differed from both the young and middle-aged drivers (Scheffé post hoc test, p < 0.001 and p = 0.022). The motor (keying) response almost doubled the time (and the number of glances needed), in comparison to reading out response, and there was a significant age times task interaction (F2,24 = 8,62; p = 0.002): in comparison to the younger groups, elderly drivers needed relatively more time for keying than for only reading out. The gender had no main effect or any significant interaction.

There was a slight but not significant difference in time of eyes off the road between the two roads (p = 0.082) in expected direction, the total time being a little longer on the wider divided motorway. Figure 3b shows, however, that as a result of the speed difference (100.2 vs. 77.9 km/h), the distance traveled during the tasks was much longer on the wider road, whereas Figures 3a (in terms of time) and 3b (in terms of distance) show very similar results in other respects.

Figure 4 indicates that the older participants did not substantially compensate their slowness by lower speed. There is no main effect of age on speed (F2,24 = 0.157) but a significant task times road interaction (F2,24 = 11.78; p = 0.002), task times age interaction (F2,24 = 3.43; p = 0.049), and task times road times age times gender interaction (F2,24 = 5.179; p = 0.013). The keying task reduced the speed on the narrow highway but not on the wide motorway, except for elderly females.

Although the mean glance time away from the road did not differ by age, as a result of different strategies used by participants, the number of long glances did (Fig. 5). Although younger drivers had practically no long glances at all, elderly drivers looked for more than 2 seconds at the display almost once per trial in the keying task on the motorway on average and every two trials on the two-lane highway (F2,24 = 3.85, p = 0.035 for the age main effect and F1,24 = 8.60, p = 0.007 for the task main effect). There was much variance in the elderly group, however. Figure 6, for the individual values averaged over all conditions, reveals that six of 11 of the older participants deviated substantially from the level of younger participants (exact two-tailed p for the group difference = 0.012), whereas the rest performed at the level of younger participants.

Along with long glances off the road, the lateral displacement of the car during in-car tasks provides a good criterion of the primary task control. Figure 7 shows a significant age effect (F2,24 = 5.70, p = 0.009) such that the oldest group differs from the other two (Scheffé, p = 0.022 resp. 0.015), as well as the task main effect (F1,24 = 8.09, p = 0.009) and the road main effect (F1,24 = 5.76, p = 0.024). This average of the absolute values of the displacement was approximately 50% to 60% higher in older drivers. Nine of the 10 largest displacements in the data—more than 1 m—were the result of older drivers, and the maximum of three older drivers reached 1.5 m.

As also shown in Table 2, both visual and cognitive performance deteriorates in older age. To search for variables to mediate age effects on time-sharing measures, the procedure proposed by Baron and Kenny32 was used based on a series of regression models. It appeared that the mediators were cognitive rather than vision-related parameters. Figure 8 A shows that the age effect on total time eyes off the road is almost fully mediated by the abilities measured with Trail Making A (TMT). In the first phase, age explains both TMT and total time separately (regression coefficients 1 and 2, respectively, 0.684 and 0.677). When both age (3a) and TMT (3b) are included in the third model as predictors to explain the total time, age has no more explanatory power over the TMA (0.216 against 0.674). In marked contrast, although visual acuity (r = 0.68) and contrast sensitivity measures (r = 0.68 to 0.70) alone correlated with the total time, their regression coefficient remained negligible (-0.18 to 0.07; not significant) when total time was regressed on both vision parameters and age.

A similar analysis (Fig. 8B) shows that the TMT best mediated the age effects on frequent long glances, too, and that vision parameters did not have any bigger role than in the case of total time.

Finally, secondary task performance was analyzed for the keying task by checking errors in the keyed sequences. It appeared that there were plenty of missing data in the recorded data, obviously not only skipped as a result of memory or other errors, but as a result of defective contact (too weak a press), which was not noticed by the experimenter who rather was concentrating on safety during experiments. One or more numbers were missing in as many as 47% of trials, with no significant differences between the groups or genders in either parametric (MANOVA) or nonparametric (Kruskal-Wallis) tests. Additionally, there were one or more extra numbers in 3% of trials. In trials with exactly eight numbers, one or more numbers were wrong in 1% and two adjacent numbers were exchanged in 0.4% of trials. The total number (percentage) of errors and any separate error category did not differ significantly between the groups. Additionally, the error rate did not correlate with any of the time-sharing parameters.

Back to Top | Article Outline


These open-road data show a marked age effect on drivers’ time-sharing performance. Older drivers have to allocate more visual attention to complete the task in terms of total time eyes off the road. Quite often they also remain looking at the target for more than 2 seconds. This is a long interruption of traffic information inflow and exceeds the generally acceptable limit for safe driving.6, 7 It also implies impaired lane-keeping control in this study in terms of larger lateral displacement during additional tasks.

The age effect was quite consistent in terms of total time needed for the task. There was only one participant in the older driver group who used less time than the slowest young driver. This indicates general slowness among the elderly.10, 11 Approximately half of our older drivers also displayed frequent long glances away from the road, clearly indicating risky behavior. This behavior can also be the result of slowness. The older participants simply need more time to search for the target button next in turn and push it if keying response is required. Even if they try to segment the task into smaller chunks to keep time eyes off the road short, one button may be too much. However, long glances may partly reflect impaired control of action such that a participant continues searching for the second, third, and fourth targets during the same glance, not being fully aware that it takes too much time from him or her, and that it already would be time to look at the road. Such problems were obvious in patients with Alzheimer disease and frontal brain injuries in our earlier studies.8, 9

These data showed a 50% to 60% larger lateral displacement in older drivers during the keying task, on the average, in comparison to the younger participants. At first sight, these additional 5 to 6 cm seem not very big. It should be noted that it refers to the mean of the absolute values of displacement, and quite often the car continues a straight line even when not visually guided, depending among others on the transverse slope of the road. However, this average displacement grows exponentially with time eyes off the road,33 and long glances away from the road represent big potential hazards. Accordingly, older drivers in these data were responsible for the largest displacements and a few lane deviations. It is to be noted, too, that many older participants performed very well when long glances away from the road are concerned (see Fig. 6). On the other hand, our on-road results show that independently of the strategy, the older participants are slower and need more time for secondary tasks. It is a risk factor in traffic.

The results showed that the age effect, both on total time and long glances, were best mediated by the Trail Making A test, which measures visual scanning, motor speed, and attention abilities.34 Vision parameters (sensory sensitivity) instead did not have an essential role in on-road time sharing performance.

It should be noted that our additional task was in fact very similar to the Trail Making A test, which requires connecting numbered circles consecutively with a pencil. Therefore, it is not surprising that TMT predicts the total time in our on-road task so well and mediates the age effect. However, TMT also mediates the age effect on frequent long glances. This supports the general conclusion that the time-sharing ability in driving is dependent on cognitive performance level, impairs with age, and can be predicted with the Trail Making A test.

The results showed that the older drivers slowed down a few km/h while doing the keying task in comparison to only reading numbers from the display. The results also suggest that the older men reduced speed during the keying task on the narrow highway only, but the older women did it also on the wide four-lane motorway. It should be noted that this either indicates compensatory behavior or the fact that speed control tends to deteriorate when an additional task demands plenty of efforts.9 The relatively modest compensation, if any, may be partly the result of the experimental setting, which even might influence the time-sharing behavior. Although an on-road study always involves real risks of running off the road and colliding with other vehicles, in contrast to simulator studies, the presence of the experimenter and the prescribed task may get participants to trust on him and go beyond the limit they would not exceed alone. Many of older drivers would not do such novel in-car tasks in their normal driving, just as they do not use a cell phone while driving.35 However, our study, even if not representing usual driving behavior of all participants, presented a setting that tested time sharing performance, especially switching of visual attention from roadway to target and back, keeping the task relatively constant and simple.

In conclusion, these results show that older drivers have increasing difficulties in controlling time sharing in highway driving already at the age of 65 to 70 years. The results suggest that older drivers should compensate for their declined cognitive performance at the strategic level by avoiding self-imposed time sharing while driving. This decline should also be taken into account by the car industry when planning new in-vehicle user interfaces, and by highway engineers when they design road and traffic information systems.

Back to Top | Article Outline


The project funded by the Road Administration, “Head on crashes—causation and trends,” made this study possible. The authors thank Mr. Timo Valkas and Jaakko Haapasalo, BSc, for building and programming the special keyboard, and Jukka Harjula, BA, Juha Karola, BA, and Lasse Warjus, MA, for on-road data collection and Mr. Aki Laukkanen for help in vehicle data analysis. This study was presented at the 3rd International Congress of Traffic and Transport Psychology, ICTTP 2004, Nottingham, United Kingdom, September 5 to 9, 2004.

Heikki Summala, PhD

Traffic Research Unit

Department of Psychology

P.O. Box 9

00014 U

Helsinki, Finland


Back to Top | Article Outline


1. Hills BL. Vision,visibility, and perception in driving. Perception 1980;9:183–216.
2. Shinar D, Schieber F. Visual requirements for safety and mobility of older drivers. Hum Factors 1991;33:507–19.
3. Owsley C. Vision and driving in the elderly. Optom Vis Sci 1994;71:727–35.
4. Ball K, Owsley C. The useful field of view test: a new technique for evaluating age-related declines in visual function. J Am Optom Assoc 1993;64:71–9.
5. Owsley C, McGwin G Jr. Vision impairment and driving. Surv Ophthalmol 1999;43:535–50.
6. Zwahlen HT, Adams CCJ, DeBald DP. Safety aspects of CRT touch panel controls in automobiles. In: Gale AG, Freeman MH, Haslegrave CM, Smith P, Taylor SP, eds. Vision in Vehicles II: Proceedings of the Second International Conference on Vision in Vehicles, Nottingham, UK, September 14 to 17, 1987. Amsterdam: North Holland; 1988:335–44.
7. Wierwille WW. Visual and manual demands of in-car controls and displays. In: Peacock B, Karwowski W, eds. Automotive Ergonomics. London: Taylor & Francis; 1993:299–320.
8. Wikman AS, Summala H. Time sharing by drivers with neurological problems. Paper presented at the 2nd International Conference on Traffic and Transport Psychology, ICTTP 2000, September 4 to 7, 2000, Bern, Switzerland.
9. Wikman AS, Haikonen S, Summala H, Kalska H, Hietanen M, Vilkki J. Time-sharing strategies in driving after various cerebral lesions. Brain Inj 2004;18:419–32.
10. Cerella J. Aging and information-processing rate. In: Birren, JE, Schaie KW, eds. Handbook of the Psychology of Aging, 3rd ed. San Diego: Academic Press; 1990:201–21.
11. Van Zomeren AH, Brouwer WH. Clinical Neuropsychology of Attention. Oxford: Oxford University Press; 1994.
12. Brouwer WH, Waterink W, Vanwolffelaar PC, Rothengatter T. Divided attention in experienced young and older drivers—lane tracking and visual analysis in a dynamic driving simulator. Hum Factors 1991;33:573–82.
13. Korteling JE. Effects of skill integration and perceptual competition on age-related differences in dual-task performance. Hum Factors 1991;33:35–44.
14. Ponds RW, Brouwer WH, van Wolffelaar PC. Age differences in divided attention in a simulated driving task. J Gerontol 1988;43:P151–6.
15. Withaar FK, Brouwer WH, van Zomeren AH. Fitness to drive in older drivers with cognitive impairment. J Int Neuropsychol Soc 2000;6:480–90.
16. Salthouse TA, Miles JD. Aging and time-sharing aspects of executive control. Mem Cognit 2002;30:572–82.
17. Salthouse TA, Atkinson TM, Berish DE. Executive functioning as a potential mediator of age-related cognitive decline in normal adults. J Exp Psychol Gen 2003;132:566–94.
18. Meiran N, Gotler A. Modelling cognitive control in task switching and ageing. Eur J Cogn Psychol 2001;13:165–86.
19. Mäntylä T. Remembering to remember: adult age differences in prospective memory. J Gerontol 1994;49:P276–82.
20. West RL. An application of prefrontal cortex function theory to cognitive aging. Psychol Bull 1996;120:272–92.
21. Summala H, Nieminen T, Punto M. Maintaining lane position with peripheral vision during in-vehicle tasks. Hum Factors 1996;38:442–51.
22. Summala H, Lamble D, Laakso M. Driving experience and perception of the lead car’s braking when looking at in-car targets. Accid Anal Prev 1998;30:401–7.
23. Horrey WJ, Wickens CD. Focal and ambient visual contributions and driver visual scanning in lane keeping and hazard detection. Paper presented at the 48th Annual Meeting of the Human Factors and Ergonomics Society, New Orleans, Louisiana, September 20 to 24, 2004. Available at: Accessed May 3, 2005.
24. Folstein MF, Folstein SE, McHugh PR. ‘Mini-mental state.’ A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189–98.
25. Reitan RM. Validity of the Trail Making Test as an indicator of organic brain damage. Percept Motor Skills 1958;8:271–6.
26. Wechsler D. Wechsler Adult Intelligence Scale: Revised. New York: Psychological Corp; 1981.
27. Lundberg C, Hakamies-Blomqvist L, Almkvist O, Johansson K. Impairments of some cognitive functions are common in crash-involved older drivers. Accid Anal Prev 1998;30:371–7.
28. Schultheis MT, Garay E, Millis SR, Deluca J. Motor vehicle crashes and violations among drivers with multiple sclerosis. Arch Phys Med Rehabil 2002;83:1175–8.
29. Rizzo M, McGehee DV, Dawson JD, Anderson SN. Simulated car crashes at intersections in drivers with Alzheimer disease. Alzheimer Dis Assoc Disord 2001;15:10–20.
30. Stutts JC, Stewart JR, Martell C. Cognitive test performance and crash risk in an older driver population. Accid Anal Prev 1998;30:337–46.
31. Hyvärinen L. The LH Symbol Tests. Long Island City, NY: The Lighthouse Inc; 1992.
32. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986;51:1173–82.
33. Wikman AS, Nieminen T, Summala H. Driving experience and time-sharing during in-car tasks on roads of different width. Ergonomics 1998;41:358–72.
34. Lezak MD. Neuropsychological Assessment, 3rd ed. Oxford: Oxford University Press; 1995.
35. Pöysti L, Rajalin S, Summala H. Factors influencing the use of cellular (mobile) phone during driving and hazards while using it. Accid Anal Prev 2005;37:47–51.

older drivers; visual attention; time-sharing; divided attention; driving performance; driving ability; open road

© 2005 American Academy of Optometry