WOOD, JOANNE M. PhD, FAAO; OWENS, D ALFRED PhD
Crash data from the United States indicate that nighttime fatality rates, adjusted for mileage, are three to four times higher than daytime rates.1 A number of factors probably contribute to the higher death toll at night, including increased alcohol use and driver fatigue. Less obvious and potentially more pervasive are the vision changes experienced, but not necessarily appreciated, by drivers under reduced illumination. Leibowitz and colleagues have suggested that drivers are unaware of their visual limitations at night because their visual guidance abilities are relatively unimpaired, whereas their visual recognition abilities are selectively degraded.2–5 According to the selective degradation hypothesis, drivers’ sustained ability to steer their vehicle easily at night and to see well-illuminated signs and instruments masks their diminished ability to see low-contrast objects, resulting in unjustified confidence when driving at night. This theory draws some support from traffic studies, which report that traffic speeds are as high at night as under daytime conditions.6
The changes in visual function that occur under reduced illumination are well recognized and include reductions in visual acuity in central7, 8 and peripheral locations,9 as well as reduced contrast sensitivity for all spatial frequencies.10, 11 The magnitude of these changes under nighttime driving is moderated to some extent by headlighting, street lighting, dashboard instruments, and a wide variety of retroreflective signs and markings. Recent evidence from a night driving simulator indicates, however, that vehicle guidance (steering) is unperturbed in low light, particularly for younger adults.12 Moreover, these changes in night vision are more severe in older individuals,8, 12 as a result of age-related changes in both optical and neural processes.13, 14 Interestingly, although it is well documented that many older drivers minimize or avoid driving at night,15, 16 crash data show that drivers aged 65 and older have a greater rate of involvement in fatal crashes at night than other drivers, except those younger than 25 years.17 Crash data also indicate that, although a small proportion of all nighttime collisions involve older drivers, the proportion of drivers involved in pedestrian collisions at night increases with age.18
From the standpoint of road safety, this evidence raises the question of whether visual assessment for drivers’ licensing can predict visual recognition abilities under real-world conditions, including night driving. Although most people drive under both day and night conditions, vision standards for licensing are based on photopic visual acuity in most countries. The only exception to this is Germany, which uses the Mesoptometer II test to assess low contrast acuity (using a Landolt C) under low mesopic luminance conditions (0.032 cdm2), in the presence and absence of glare (simulating low-beam headlights). The legal standards for driving at night, set by the German Ophthalmological Society, require that a driver must be able to recognize a Landolt C (6/60) at a contrast level of 1:5 to be eligible to drive a private vehicle and a contrast level of 1:2.7 to drive a commercial vehicle.19 The introduction of this standard was based on the finding that visual acuity, measured under conditions similar to those encountered at twilight, was significantly reduced in older drivers, although visual acuity measured under standard conditions was normal.20 We are not aware, however, of any on-road studies that have provided external validation for this standard.
The present investigation was designed to determine whether visual acuity or contrast sensitivity, measured under a range of luminance levels, could predict visual recognition while driving under real-world conditions. This relationship was tested for licensed drivers of three different age groups. We hypothesized that the reduction in contrast sensitivity experienced under low light conditions might be more important than the changes in resolution; hence, contrast sensitivity would be more useful for predicting recognition performance when driving at night.
There were a total of 24 participants, including eight younger drivers (mean age: 21.5 ± 2.8 years), eight middle-aged drivers (mean age: 46.6 ± 4.2 years), and eight older drivers (mean age: 71.9 ± 2.6 years) with equal numbers of women and men in each group. Participants were recruited from the general driving population. They were licensed drivers with at least 3 years of driving experience, and all reported that they drove regularly. All participants passed the minimum drivers’ licensing criterion for corrected binocular visual acuity of 6/12 (20/40). Participants wore their normal optical correction while driving.
The study was conducted in accordance with the requirements of the Queensland University of Technology Human Research Ethics Committee. All participants were given a full explanation of the experimental procedures, and written informed consent was obtained with the option to withdraw from the study at any time.
Visual acuity and contrast sensitivity were measured in the laboratory under four luminance conditions. All measures were taken binocularly. The room illumination resulted in a test chart luminance of 65 cd/m2. Participants wore the same refractive correction used habitually for driving (and worn for all of the driving assessments described here) in conjunction with the appropriate correcting lens for the working distance of each test.
After a dark-adaptation period of 30 minutes, visual acuity and contrast sensitivity were measured while the participants wore goggles that were fitted with ND filters of decreasing density: beginning with 3.0 ND (0.065 cd/m2), followed by 2.0 ND (0.65 cd/m2), 1.0 ND (6.5 cd/m2), and finally no filter (65 cd/m2). The order of luminance levels always proceeded from the dimmest to the brightest condition to minimize both the time required for dark adaptation and the potential for learning effects in successive tests.
Static high contrast visual acuity was measured using two versions of a standard logMAR chart (Australian Vision Chart No. 5) at a working distance of 3.2 m, unless visual acuity was worse than the top line of the chart, in which case shorter viewing distances were used and the results scored accordingly. Subjects were forced to guess letters, even when they were unsure, until a full line of letters was incorrectly read. Each letter seen was scored as -0.02 log units.
Pelli-Robson Letter Contrast Sensitivity
Measures of contrast sensitivity were determined using two versions of the Pelli-Robson chart at the standard working distance of 1 m. Subjects were instructed to look at a line of letters and forced to guess the letter when they were not sure until a full line of letters was incorrectly read. Each letter was scored as 0.05 log units.
Real-world visual performance was assessed while participants drove under day and night conditions on the closed road circuit at the Mt Cotton Driver Training Centre, which has been used in previous studies of driving and vision.21 The experiment was canceled if it was raining or the road surface was wet. The circuit, which is representative of a rural road, consists of a two- to three-lane bitumen (asphalt) road surface and includes hills, curves, bends, and straight sections as well as standard road signs and road markings. A 1.8 km (1.1 mile) section of the circuit was used for this study. The circuit does not include any street lighting. At night, realistic glare conditions, simulating an oncoming vehicle, were created by positioning automotive headlights mounted at the correct height and separation at two locations along the circuit. These headlights were activated when the test vehicle drove through a series of remote sensors.
The test vehicle was a 1997 Holden Commodore station wagon equipped with automatic transmission and a digital video system to measure lane position. The driver’s view of the speedometer was occluded by translucent film. High-beam headlights were active during all night tests to maintain a consistent beam pattern. In addition to the normal high beam, lower headlight intensities were obtained by mounting ND filters on the headlights, thus attenuating the luminous intensity of the beam by 0.6 (-75%), 0.9 (-87.5%), and 1.5 (-97%) log units. One should note that none of these conditions duplicates the illumination of a low-beam system because low and high beams differ in both the luminous intensity and optical distribution of the light. Low beams aim the maximum illumination downward and toward the shoulder, whereas high beams aim the maximum illumination either toward the horizon straight ahead (European designs) or toward the horizon but slightly (1°) toward the shoulder (U.S. designs).22, 23 We avoided confounding variations of intensity with those of beam pattern by using only the high-beam setting. If one compares the luminous intensity of the road and shoulder at long distances (i.e., 0° elevation, and 0° to 0.5° toward the shoulder), U.S. low beams are most closely approximated by the ND 0.9 condition, and European low beams are comparable to the ND 1.5 condition.
Participants drove around the circuit five times, once in daylight and four times at night with the headlight beams set at each of the four intensity levels. Participants were instructed to drive at a comfortable speed, to be alert for unpredictable hazards (like wild animals) as they ordinarily would on rural roads and to report relevant targets, including road signs, low-contrast road hazards, and pedestrians. Between laps, the headlight filters were changed surreptitiously in preparation for the next test run while the driver was distracted by the administration of a questionnaire. Participants also drove around the same test circuit under daytime conditions and were required to complete the same driving tasks. The order of day and night test conditions and the order of headlight intensities at night were counterbalanced across participants within all age groups.
There were 21 standard road signs located around the circuit, and participants were instructed to report all road signs and other important targets (e.g., animals or pedestrians) as they drove around the circuit. Large, low-contrast road hazards were placed at four locations along the circuit. These road hazards consisted of 15 cm × 80 cm × 220 cm (reflectance of approximately 10%) thick gray foam rubber, so that although participants could feel the hazards when hit, they had a minimal effect on vehicle control. Participants were asked to report when they saw a road hazard and to avoid it by steering around it. Performance was measured as the number of road hazards reported as seen and the number hit. Two pedestrians, who were wearing retroreflective markings of equal area but in different spatial configurations, walked along the shoulder of the opposite lane in a direction facing the oncoming test vehicle. To minimize learning effects, the pedestrians were positioned at variable locations, including both straight and curved segments of the circuit. Both pedestrians wore a black tracksuit and either a sash consisting of a single retroreflective stripe (2.5 cm wide) that extended diagonally from the right shoulder to the left hip or the same quantity of retroreflective material in narrower (0.75 cm) stripes attached to the sweatsuit at the waist, shoulders, elbows, wrists, knees, and ankles known as “biomotion.” The biomotion condition was based on research by Johansson,24 which established that luminous markings on the limb joints create a unique perceptual phenomenon called “biological motion.” Later work using video projections of the night road environment indicated that “biomotion markings” may be superior to other marking configurations,25 and this finding has been replicated in the U.S.26 and Finnish27 road environments using passengers as subjects. However, to our knowledge, it had not yet been investigated for drivers of different ages under real-world conditions. Hence, the biomotion condition was included to determine the extent to which biomotion markings could improve pedestrian visibility, relative to more conventional markings of the torso, for drivers of various ages.
Several additional dependent measures were collected, including measures of driving behavior and responses to an extensive questionnaire. The present report focuses on the clinical vision tests and their relationship to the drivers’ ability to recognize relevant road objects (road signs, road hazards, and pedestrians) while driving. This latter measure was calculated as the percentage recognition of signs, road hazards, and pedestrians correctly recognized.
The group mean data for visual acuity and contrast sensitivity are plotted as a function of luminance and age in Figure 1. This shows similar performance levels for the young and middle-aged participants, whereas the older participants had lower performance levels across all of the luminance conditions. A repeated-measures analysis of variance (ANOVA) of the visual acuity data with one within-subject factor (chart luminance) and one between-subjects factor (driver age) showed significant main effects of luminance (F[3,63] = 598.9; p < 0.001) and age (F[2,21] = 11.2; p < 0.001) and a significant interaction between luminance and age (F[6,63] = 4.6; p = 0.008). Because the data did not meet assumptions of sphericity, the Greenhouse-Geisser correction was used in computing alpha levels.
The contrast sensitivity data showed similar trends to those for visual acuity, with an ANOVA showing significant main effects for luminance (F[3,63] = 540.6; p < 0.001) and age (F[2,21] = 12.03; p < 0.001), but no significant interaction between luminance and age (F[6,63] = 0.64; p = 0.68).
Intercorrelations Between Vision Tests
The relationship between visual acuity and contrast sensitivity was also examined under the four different luminance levels; Table 1 gives the Pearson r values for the full correlation matrix. When a Bonferroni correction factor was applied, all but one of the correlations were significant at the p < 0.05 level, with the exception being the correlation between visual acuity measured under the brightest condition and contrast sensitivity measured under the lowest luminance condition.
Under standard photopic test conditions, contrast sensitivity and visual acuity were modestly correlated (r = -0.61, p < 0.05). The correlation between contrast sensitivity and visual acuity was much higher for visual acuity measured under lower luminance levels, as indicated by higher correlations (r = -0.84, p < 0.01).
Effect of Luminance and Age on Recognition While Driving
Figure 2 illustrates the change in recognition performance on the road, defined as the mean percentage recognition of all targets, as a function of luminance and the age of the driver. A repeated-measures ANOVA with one within-subject factor (light condition) and one between-subjects factor (driver age) showed significant main effects for light condition (F[4,84] = 23.1, p < 0.001) and driver age (F[2,21] = 3.48, p = 0.05), with older drivers performing worse than either middle-aged or younger subjects. The interaction between age and light condition was not significant (F[8,84] = 1.31, p = 0.25).
Relationship Between Vision Tests and Recognition While Driving
The relationship between the vision tests and drivers’ ability to recognize road objects, including signs, pedestrians, and road hazards, was examined through a series of correlational and multiple regression analyses. Because contrast sensitivity under standard photopic conditions was highly correlated with visual acuity measured under low-luminance conditions (r = 0.77-0.89) (Table 1), we first examined the correlations between standard (photopic) measures of visual acuity and contrast sensitivity and real-world recognition ability under the five different illumination conditions. Table 2 represents the portion of variance in real-world recognition that could be explained by either photopic visual acuity or contrast sensitivity (r2 values).
Table 2 clearly shows that photopic visual acuity measures did not predict variations in recognition while driving for either day or nighttime conditions. A stronger relationship was found for photopic contrast sensitivity, which showed increasing predictive power as the lighting condition was reduced, ranging from a nonsignificant 14% of the variance in daylight to 40% in the darkest condition (Fig. 3). The same pattern of correlations was obtained when analyses were restricted to data for sign recognition, excluding pedestrians, and low-contrast hazards.
Multiple regression analyses were used to determine whether prediction of real-world performance could be improved through the use of multiple vision tests. The first analysis used a stepwise regression model to determine an optimal combination of tests by entering all measures of acuity and contrast sensitivity (for all four luminance levels) as possible predictors of drivers’ recognition for each of the five light conditions. As shown in Table 3, there was no consistent pattern of “optimal” predictor variables for all five driving conditions. The addition of age as a predictor produced a small improvement in only one of the five road conditions. Two practical test combinations were suggested, however, by the fact that three vision tests appeared among the “optimal” predictors for multiple conditions, namely photopic contrast sensitivity (PR65), mesopic visual acuity (VA6.5), and photopic visual acuity (VA65). In view of the fact that photopic visual acuity is already established as the standard test, two additional “practical models” were examined to determine if adding one more test to the existing standard would offer a substantial improvement. The practical models were: Model 1: Photopic Visual Acuity and Photopic Contrast Sensitivity (i.e., VA65 and PR65) Model 2: Photopic Visual Acuity and Mesopic Visual Acuity (i.e., VA65 and VA6.5)
Table 3 presents the resulting r2 values for both of the practical models in comparison with the predictive power of the stepwise regression for all vision tests combined. These analyses showed that both “practical models” have greater merit than any single test, and both are only marginally inferior to the more complex (and impractical) combinations of all possible tests. A similar pattern of results was also found when the predictive power of photopic visual acuity and contrast sensitivity were compared with that of the two “practical models” for predicting driving recognition when all of the conditions were combined (average performance for day and the four night conditions) and that of the difference in performance between day and the darkest night condition (Table 4), although practical model 2 did offer improved predictive power when considering all conditions combined.
Our findings confirm that reduced luminance and increasing age have a detrimental affect on drivers’ recognition ability measured while traveling on a closed road circuit, as well as on clinical measures of visual acuity and contrast sensitivity. Importantly, our results demonstrate that, contrary to commonly accepted licensing standards, visual acuity measured under standard photopic testing conditions did not predict drivers’ recognition ability under either day or nighttime road conditions. Rather, photopic contrast sensitivity provided a better prediction of recognition while driving, especially under the dimmest nighttime condition. Multiple regression analyses, which evaluated the combined predictive power of both visual acuity and contrast sensitivity measured under a wide range of luminance levels, did not reveal a consistent optimal combination. Further evaluation of two practical models indicated, however, that adding either photopic contrast sensitivity or mesopic visual acuity to the standard acuity test can provide a much more useful alternative to current drivers’ licensing vision standards.
The results for the clinical tests of visual function are in accordance with previous studies that have also demonstrated that under reduced luminance levels, both visual acuity7 and contrast sensitivity measured with the Pelli-Robson chart28, 29 are decreased, and these effects are exacerbated for older participants. The present study extends these findings in a potentially useful way to the prediction of performance in real-world conditions. The correlation between visual acuity and contrast sensitivity measured under standard luminance conditions was relatively modest, which is in accord with previous findings.30 Contrast sensitivity measured under standard photopic test conditions was significantly correlated with visual acuity measured at lower luminance levels, however, and appears to tap into similar mechanisms, assessing sensitivity at intermediate spatial frequencies.
The finding that the visual function of older participants is degraded to a greater extent under low luminance compared with standard photopic conditions has prompted many to suggest that older drivers should pass a low luminance visual acuity examination to be eligible to drive at night.8, 20, 31 Similar conclusions were drawn by Anderson and Holliday,32 who showed that photopic measures of visual acuity did not predict visual acuity measured under nighttime driving conditions in a roadside vehicle, which is consistent with the results of the study reported here. The present findings provide new evidence based on target recognition in the road environment that 1) photopic visual acuity is of limited value in predicting performance of the current population of drivers, and 2) prediction of real-world performance can be enhanced by use of a photopic test of contrast sensitivity. Still better predictions of drivers’ performance were obtained by the use of two vision tests, which supplement the current standard with either photopic contrast sensitivity or mesopic visual acuity. Selection of the best practical combination will require further research to establish appropriate criteria for combining the two scores and to examine logistic aspects of the second test. Photopic contrast sensitivity has the advantage of a standardized procedure under normal lighting conditions, whereas mesopic visual acuity has the advantage of seeming familiar to nonprofessional examiners and those to be tested. On the other hand, measurements of acuity under mesopic conditions can vary widely as a result of small uncontrolled variations in luminance,31, 33 which may be problematic in light of the fact that 1) it may be impractical to control luminance precisely, and 2) there are no standardized procedures for assessing mesopic acuity.
Our data collected while participants were driving under real-world conditions, demonstrate that recognition ability under nighttime conditions is reduced compared with daytime ability, and that these effects are more severe for older drivers. These data are in accordance with previous studies, which have shown that older drivers have poorer sign recognition at nighttime compared with younger drivers34 and have greater difficulty in recognizing roadside pedestrians.26, 35, 36
We should note that the correlations between contrast sensitivity and real-world recognition performance reported here may be an underestimate of the strength of the relationship because of individual differences in the speed traveled on the closed road circuit. Our results, reported elsewhere, showed that the older participants drove more slowly than the younger participants.37 It is likely that these differences in speed tended to reduce the recognition scores of the younger drivers and to enhance those of older drivers. If so, the correlations reported in Tables 2 through 4 would have been higher if all participants had traveled at the same speed. As it is, however, the present data provide a more valid estimate of the predictive strength of clinical vision tests for “normal” (unconstrained) driving behavior.
The potential use of the Pelli-Robson chart in driver licensing is also supported by other researchers who have found that contrast sensitivity is a significant factor contributing to the prediction of crash rates in older drivers38 and closed road driving performance under daytime conditions.39 Interestingly, the Pelli-Robson chart has also been cited in the debate in the medical literature regarding the use of night vision testing for licensing. Jory,40 in a letter to the editor, proposed the Pelli-Robson chart for night vision testing, whereas Leung41 argued that there are a number of other equally important factors contributing to night driving ability, including dark adaptation rate, glare sensitivity, and scotopic retinal sensitivity. These letters typify much of the debate on vision and driving in being opinion-based rather than evidence-based. The study reported here is the first to provide evidence-based data to support the proposal that specific combinations of two tests of visual function can predict visibility during nighttime driving.
The authors thank Queensland Transport for allowing the use of the facilities at the Mt Cotton Driver Training Centre and the staff of the Mt. Cotton Centre for their generous cooperation and support. The authors are grateful for the assistance of Justin Owens, Daniel Whittam, and Mark Woolf during data collection, and Prof. Richard Tyrrell and Dr. Michael Sivak for helpful advice during preparation of the manuscript. This study was supported by grants from the Australian Research Council, Queensland University of Technology, and Franklin & Marshall College.
Joanne M. Wood, PhD, FAAO
Centre for Health Research–School of Optometry
Queensland University of Technology
Victoria Park Rd
Kelvin Grove Q4059, Australia
2. Leibowitz HW, Owens DA. Nighttime driving accidents and selective visual degradation. Science 1977;197:422–3.
3. Leibowitz HW, Owens DA. Behavioral implications of civil twilight. In: Light and Color in the Open Air. Technical Digest Series, vol 12. Washington, DC: Optical Society of America; 1990:105–8.
4. Leibowitz HW, Owens DA, Post RB. Nighttime driving and visual degradation. SAE Technical Paper Series, #820414. Warrendale, PA: Society of Automotive Engineers; 1982.
5. Owens DA. Twilight vision and road safety: seeing more than we notice but less than we think. In: Andre, J, Owens DA, Harvey LO Jr, eds. Visual Perception: The Influence of H W Leibowitz. Washington, DC: American Psychological Association; 2003:157–80.
6. Herd DR, Agent KR, Rizenbergs RL. Traffic accidents: day versus night. In: Transportation Research Board, Commission on Sociotechnical Systems, National Research Council. Traffic Accident Analysis and Application of Systems Safety. Washington, DC: National Academy of Sciences; 1980:25–30.
7. Johnson CA, Casson EJ. Effects of luminance, contrast, and blur on visual acuity. Optom Vis Sci 1995;72:864–9.
8. Sturr JF, Kline GE, Taub HA. Performance of young and older drivers on a static acuity test under photopic and mesopic luminance conditions. Hum Factors 1990;32:1–8.
9. Bedell HE. Eccentric regard, task, and optical blur as factors influencing visual acuity at low luminances. In: Johnson CA, Leibowitz HW, eds. Night Vision: Current Research and Future Directions. US6RD8087N38. Washington, DC: National Academy Press; 1987:146–61. Available at: http://www.nap.edu/books/POD259/html/146.html
. Accessed April 20, 2005.
10. De Valois RL, Morgan HC. Psychophysical studies of monkey vision. II. Squirrel monkey wavelength and saturation discrimination. Vision Res 1974;14:69–73.
11. Woodhouse JM, Barlow HB. Spatial and temporal resolution and analysis. In: Barlow HB, Mollon JD, eds. The Senses. Cambridge: Cambridge University Press; 1982:133–64.
12. Owens DA, Tyrrell RA. Effects of luminance, blur, and age on nighttime visual guidance: a test of the selective degradation hypothesis. J Exp Psychol Appl 1999;5:115–28.
13. Kline DW, Schieber F. Vision and aging. In: Birren JE, Schaie KW, eds. Handbook of the Psychology of Aging, 2nd ed. New York: Van Rostrand Reinhold; 1985:296–331.
14. Sloane ME, Owsley C, Alvarez SL. Aging, senile miosis and spatial contrast sensitivity at low luminance. Vision Res 1988;28:1235–46.
15. Kosnik WD, Sekuler R, Kline DW. Self-reported visual problems of older drivers. Hum Factors 1990;32:597–608.
16. Waller PF. The older driver. Hum Factors 1991;33:499–505.
17. Mortimer RG, Fell JC. Older drivers: their night fatal crash involvement and risk. Accid Anal Prev 1989;21:273–82.
18. Owens DA, Brooks JC. Drivers’ Vision, Age, and Gender as Factors in Twilight Road Fatalities (Report No UMTRI-95-44). Ann Arbor, MI: The University of Michigan Transportation Research Institute; 1995.
19. Harms H, Nolte W. Anleitung fur die augenarztliche Untersuchung und Beurteilung der Eignung zum Fuhren von Kraftfahrzeugen der DOG. In: Conrads H, Gramberg-Danielsen B, eds. Richtlinien und Untersuchungsanleitungen, Berufsverband der Augenarzte Deutschlands. Heidelberg: Kaden; 1984:43.
20. Aulhorn E, Harms H. [The examination on fitness for driving at darkness with the mesoptometer.] Klin Monatsbl Augenheilkd 1970; 157:843–73.
21. Wood JM, Troutbeck R. Effect of visual impairment on driving. Hum Factors 1994;36:476–87.
22. Andre J, Owens DA. The twilight envelope: a user-centered approach to describing roadway illumination at night. Hum Factors 2001;43: 620–30.
23. Schoettle B, Sivak M, Flannagan MJ. High-beam and low-beam headlighting patterns in the US and Europe at the turn of the millennium. SAE Technical Paper Series, #2002-01-0262. Warrendale, PA: Society of Automotive Engineers; 2002.
24. Johansson G. Visual perception of biological motion and a model for its analysis. Percept Psychophys 1973;14:201–11.
25. Owens DA, Antonoff RJ, Francis EL. Biological motion and nighttime pedestrian conspicuity. Hum Factors 1994;36:718–32.
26. Luoma J, Schumann J, Traube EC. Effects of retroreflector positioning on nighttime recognition of pedestrians. Accid Anal Prev 1996; 28:377–83.
27. Luoma J, Penttinen M. Effects of experience with retroreflectors on recognition of nighttime pedestrians: comparison of driver performance in Finland and Michigan. Transportation Res Part F: Traffic Psychol Behav 1998;1F:47–58.
28. Taub HA, Sturr JF. The effects of age and ocular health on letter contrast sensitivity and high and medium contrast acuity as a function of luminance. Clin Vision Sci 1991;6:181–9.
29. Kogure S, Membrey WL, Fitzke FW, Tsukahara S. Effect of decreased retinal illumination on frequency doubling technology. Jpn J Ophthalmol 2000;44:489–93.
30. Rubin GS, West SK, Munoz B, Bandeen-Roche K, Zeger S, Schein O, Fried LP, SEE Project Team. A comprehensive assessment of visual impairment in a population of older Americans. The SEE Study. Salisbury Eye Evaluation Project. Invest Ophthalmol Vis Sci 1997;38:557–68.
31. Sturgis SP, Osgood DJ. Effects of glare and background luminance on visual acuity and contrast sensitivity: implications for driver night vision testing. Hum Factors 1982;24:347–60.
32. Anderson SJ, Holliday IE. Night driving: effects of glare from vehicle headlights on motion perception. Ophthal Physiol Opt 1995;15: 545–51.
33. Johnson CA. Effects of luminance and stimulus distance on accommodation and visual resolution. J Opt Soc Am 1976;66:138–42.
34. Sivak M, Olson PL, Pastalan LA. Effect of driver’s age on nighttime legibility of highway signs. Hum Factors 1981;23:59–64.
35. Chrysler ST, Danielson SM, Kirby VM. Age differences in visual abilities in nighttime driving field conditions. Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting; Santa Monica, CA; Human Factors and Ergonomics Society; 1996:923–7.
36. Wood, JM, Tyrrell RA, Carberry TP. Limitations in drivers’ ability to recognise pedestrians at night. Hum Factors 2005; in press.
37. Wood JM, Owens DA, Woolf MI, Owens J. Predicting night-time visibility while driving. J Vision 2002;2:331a. Available at: http://journalofvision.org/2/7/331
. Accessed April 20, 2005.
38. Decina LE, Staplin L. Retrospective evaluation of alternative vision screening criteria for older and younger drivers. Accid Anal Prev 1993;25:267–75.
39. Wood JM. Age and visual impairment decrease driving performance as measured on a closed-road circuit. Hum Factors 2002;44:482–94.