The Effects of Visual Field Loss from Stroke on Performance in a Driving Simulator : Optometry and Vision Science

Secondary Logo

Journal Logo

FEATURE ARTICLE – PUBLIC ACCESS

The Effects of Visual Field Loss from Stroke on Performance in a Driving Simulator

Bro, Tomas MD, PhD1∗; Andersson, Jan PhD2

Author Information
doi: 10.1097/OPX.0000000000001928
  • Free

Abstract

FU1

Visual field loss is a common consequence of stroke and precludes driving in many countries. However, legal visual requirements for drivers' licenses are largely without scientific basis.1 Some studies have shown that individuals with visual field loss are more unsafe drivers than normally sighted individuals, but other studies have failed to show any connections.2 In the absence of strong scientific evidence, the visual requirements for driving differ both between and within countries.3

Safety and performance are two major outcome measures used in research on driving. Safety is defined by adverse events that might be studied in accident statistics, surveys, or simulators. Performance is driver behavior when maneuvering the vehicle that may be tested either on-road or in a simulator. Assessing drivers on-road may provide conditions more similar to real-world driving but have limitations in systematically testing of hazardous situations. Simulator tests, on the other hand, make it possible to test the driver's behavior under repeatable and standardized conditions, yet simulations are not a perfect imitation of reality and can create motion sickness.2

STROKE AND VISUAL FIELD LOSS

Stroke is the second leading cause of death and a major cause of disability worldwide.4 Visual field loss after stroke is very common. Hemianopia and quadrantanopia are the most frequent consequences.5 A community-based study detected homonymous defects in 8.3% of 194 post-stroke patients.6 Asymptomatic but perimetric visual field loss may be found in 33% of patients after transient ischemic attack or minor stroke.7 In an assessment of 61 patients 9 months after stroke, homonymous visual field loss confirmed with Humphrey perimetry was found in 16%.8 Partial recovery usually occurs within the first 3 months. Improvement beyond 6 months is rare.9

Previous Studies of Visual Field Loss from Stroke and Simulated Driving

Several on-road studies have assessed driving with visual field loss from stroke. Drivers with hemianopia had reduced rating for maneuvers on a group level but may drive without obvious errors in the majority of cases, which highlights the necessity for individual evaluations.10 Safe driving seems to be more achievable with quadrantanopia than hemianopia.11,12

Studies have also been performed using simulated driving performance. Patients with homonymous visual field loss are shown to have reduced ability to detect objects and have more collisions on the affected side of the visual field.13,14 Central visual field loss seems to have a stronger negative effect on safe driving than does peripheral loss.15 Lane position may be displaced toward the nonaffected side of the visual field.16 Paracentral visual field loss causes slower reactions to pedestrians in blind areas and more missed responses.17 Other studies have shown that drivers might compensate for the impaired detection, possibly through increased explorative head and eye movements.18,19

Limited research has been completed regarding how different types of stroke-related visual field loss could affect driving safety and performance. Many studies have also grouped macular-sparing hemianopia together with complete hemianopia without distinction.12,14,16,19

The study by Papageorgio et al.15 provided an exception to the above and included detailed information about the area spared from damage in the affected hemifield within a central radius of 30°. Furthermore, no previous study of simulated driving with stroke has included more than 30 cases. The purpose of this study was to compare driving simulator performance of a large cohort of stroke participants with a normally sighted control group. Furthermore, we investigated the effects of different extents and types of visual field loss on driving ability.

METHODS

Participants

From 2014 to 2015, healthy normally sighted individuals were recruited to perform a driving simulator test at the Swedish Road and Traffic Research Institute in Linköping. The inclusion criteria involved healthy individuals between 55 and 75 years of age who drove approximately 15,000 km/y. These 83 individuals were paid 100 euros for participation. Their driving behavior was then used to create reference values for critical behavior used to define a passed test.

Thereafter, between 2016 and 2018, drivers with withdrawn licenses due to visual field loss caused by several reasons could apply for the test. It was initially performed only for research purposes but was later accepted by the Swedish Transport Agency as a cause for dispensation (i.e., return of the driver's license) if no other medical complications were present, which made the participants highly motivated. The cost for the individual was approximately 2000 euros. Over and above this fee, the participants also spent time and money traveling to the simulator from all parts of Sweden.

Because the loss of a driving license may have dire consequences for the individual's welfare,20 the interest in the test was very high. More than 300 individuals with visual field loss completed it. This study analyzes the results from the driving simulator test for 153 participants with visual field loss due to stroke. Other reasons for visual field loss were glaucoma, laser treatment for diabetic retinopathy, or optic nerve drusen.21–23 Ethical approval was given by Linköping University Committee (Dnr 2014/124-31).

Background Information

Before study participation, the individuals submitted medical records and visual field charts to the study investigators. Diagnosis and visual field examinations were therefore always done in advance, in a clinical setting. If medical information was missing, this was requested from the Swedish Transport Agency.

Field of Vision

All stroke participants had lost their driver's license because of severe visual field loss according to the Swedish legislation. This legislation requires that all corresponding test points (the maximum sensitivity for an overlapping location from two monocular perimetric examinations) within 10° radius must be at least 20 dB, and only one corresponding test point between 10 and 20° is allowed to be less than 10 dB. The examination should be performed with Humphrey perimetry target size III or equivalent static threshold perimetry,24 and the license is revoked only if the visual field loss is permanent. All study participants had lesions present for 6 months or more. Inclusion criteria for the visual field analysis were an examination with Humphrey perimetry 24-2/30-2 (Carl Zeiss Meditec, Oberkochen, Germany), Henson Zata 24-2 (Topcon, Tokyo, Japan), or Octopus G standard (Haag-Streit Diagnostics, Köniz, Switzerland). Visual field defects were classified as follows: homonymous sectoranopia, homonymous hemianopia sparing the macula, quadrantanopia, partial or incongruous homonymous hemianopia, complete homonymous hemianopia, or homonymous scotomatous defects according to previously defined criteria (Fig. 1).5

F1
FIGURE 1:
Classification of visual field defects (the right eye is on the right, and the left eye is on the left; all examinations with Humphrey perimetry 24-2). (A) Homonymous sectoranopia. (B) Homonymous hemianopia sparing the macula. (C) Quadrantanopia. (D) Partial or incongruous homonymous hemianopia. (E) Complete homonymous hemianopia. (F) Homonymous scotomatous defect.

In addition, the binocular integrated visual field was calculated by merging monocular test results from each eye, using the point with the higher sensitivity from each test. Mean sensitivity was evaluated in the following clusters: central (0 to 10°) and peripheral (10 to 20°), right and left, and superior and inferior (Fig. 2).

F2
FIGURE 2:
Test point clusters in an integrated visual field.

Simulator

The driving simulator Sim III (Swedish Road and Traffic Research Insitute, Linköping, Sweden) consisted of a real truncated car body.25 Moving road patterns and landscape were shown on one large screen in front of the car. Six projectors created one seamless image with a width of approximately 140°. Three LCD displays were used to simulate rearview and side mirrors. A vibration table provided high-frequency reproduction of road unevenness and moved the compartment/cab relative to the projected image (Fig. 3).

F3
FIGURE 3:
Picture of Sim III at VTI with the motion platform and the interior used in the test.

Procedure

The procedure was the same for the normally sighted control group and the stroke participants, except that the control group also was screened for visual field defects with Humphrey perimetry 24-2. None in this group had visual field loss. All participants signed an informed consent form before the test drive. All participants could abort at any time, without giving any explanation. However, all drivers completed the test.

Simulator Scenario

The driving scenario contained three types of roads with different speed limits: city driving (30 to 50 km/h), rural road (70 km/h), and motorway (110 km/h). The drive took approximately 50 minutes to complete, depending on the participant's chosen speed. The actions made by buses, cars, and pedestrians were initiated by the speed and distance from the car driven by the participant. During the development of the scenario, 90° turns were included. However, the pilot testing revealed that more than half of the drivers experienced simulator sickness. Therefore, the final version did not include any sharp turns. It should also be noted that collisions were not followed by a “crash experience.” In this way, all participants could complete the scenario without negative psychological effects from previous critical driving behavior. Further details can be found in our previous publications.21–23

Assessment of Passed versus Failed

All test results were digitally stored in protocols with a video recording of the full driving scenario. The data from the stroke participants were retrospectively analyzed by two independent traffic safety experts: one traffic inspector and one traffic safety researcher. In total, seven traffic inspectors and five traffic safety researchers were involved in the assessment. The instruction to the traffic inspector was to use their understanding of a normal on-road driving license test session according to the Swedish guidelines.26 The traffic safety researchers instead used a rating scale, based on 95% confidence intervals from the control group.27 Still, the classification into pass or failed was finally always based on subjective assessments. If the experts disagreed in their opinion, an additional traffic safety researcher performed a third assessment. However, this was rare because the interrater agreement was high with 93% overlap. None of the experts had any information about the side and degree of the driver's visual field loss when making pass/fail decisions.

Design

This study had a cross-sectional design with a comparison component. The dependent measures used in this study were several driving measures (safety measures: collisions and failure to give way; performance measures: time headway and reaction time). The independent variables were normally sighted versus stroke, passed versus failed, and classification, severity, and side of visual field loss.

Collisions

The test included 33 possible collisions with both other cars and pedestrians. Seven collisions were objects in front of the car, 6 in the back, 14 to the right, 2 to the left, and 4 both to the right and to the left.

Failure to Give Way

Failure to give way was measured 11 times and included only pedestrians. Seven were to the right, and four were to the left. We used the minimum distance to the pedestrian and considered distances between 0 and 1 m as hazardous and 1 and 2 m as risky. Failure to properly give way was a specific situation occurring when the test driver came too close to pedestrians either by stopping or by passing too close. Distances less than 1 m were deemed as hazardous. This cutoff level was constructed from subjective evaluation of data from the control participants.

Time Headway

Time headway is the distance to different moving objects divided by the experimental vehicle speed.28 Time headway was measured 29 times during the test. Values less than 1 second were considered as critical, motivated by previous research on traffic safety.29

Reaction Time

Reaction time was measured 17 times during the test with objects that suddenly appeared and required braking. This measure could therefore be evaluated only if braking occurred, and the event did not end in lane changing, overtaking, or even collision. Reaction times for 2 seconds were therefore not relevant, and the behavior in the specific event could instead be captured by other measures.

Speed

Mean speed was measured on three different stretches with different recommended speeds (motorway [110 km/h], rural [70 km/h], and urban [30 to 50 km/h]).

Lane Position

Lateral position of the stroke participant was compared with that of the control participants; that is, a positive value indicated (Tables 1, 2) that the stroke participants drove more to the left and a negative value more to the right.

TABLE 1 - Driving simulator data for controls and post-stroke patients with 95% confidence intervals
Control Stroke total P Stroke passed Stroke failed P
n 83 153 99 54
Average age (y) 65 62 .04 60 (58 to 62) 66 (63 to 69) <.001
Female sex (%) 36 11 <.001 13 (13 to13) 7 (7 to 7) .33
Time since stroke onset (y) 5.5 (4.6 to 6.4) 5.6 (4.4 to 6.8) 5.4 (3.8 to 7) .83
Average speed city driving (km/h) 41 (40 to 42) 39 (38 to 40) .08 38 (37 to 39) 40 (37 to 43) .14
Average speed rural road (km/h) 80 (78 to 82) 74 (73 to 75) <.001 75 (74 to 76) 73 (71 to 75) .04
Average speed motorway (km/h) 106 (104 to 108) 103 (102 to 104) .05 104 (103 to 105) 100 (97 to 103) <.001
Drivers with collisions (%) 18 (18 to 18) 22 (21 to 23) .47 20 (19 to 21) 26 (24 to 28) .39
Drivers with hazardous (0–1 m) FGW (%) 2 (2 to 2) 3 (3 to 3) .65 1 (1 to 1) 6 (5 to 7) <.001
Driver with risky (1–2 m) FGW (%) 18 (18 to 18) 17 (17 to 17) .85 9 (8 to 10) 31 (29 to 33) <.001
Average critical (<1 s) THW events 3.9 (3 to 4) 4.5 (4.2 to 4.8) .06 3.8 (3 to 4) 5.6 (5 to 6) <.001
Average RT (s) .66 (1 to 1) 0.67 (0.65 to 0.69) .57 0.65 (1 to 1) 0.71 (1 to1) <.001
Average LP city driving (m) 0 (−0.04 to −0.04) −0.11 (−0.14 to −0.08) <.001 −0.11 (−0.15 to −0.07) −0.11 (−0.18 to −0.05) .95
Average LP rural road (m) 0 (−0.06 to −0.06) −0.04 (−0.07 to −0.01) .28 −0.04 (−0.08 to −0.01) −0.03 (−0.08 to −0.02) .62
Average LP motorway (m) 0 (−0.11 to −0.11) 0.29 (0.22 to 0.37) <.001 0.31 (0.22–0.4) 0.27 (0.11 to 0.42) .11
FGW = failure to give way; LP = lateral position compared with control (+ to the left, − to the right); RT = reaction time; THW = time headway.

TABLE 2 - Comparison of drivers with hemianopia or quadrantanopia to the left or right with 95% confidence intervals
Complete/partial/macular-sparing hemianopia (B + D + E) P Quadrantanopia
(C)
P
Left Right Left Right
n 21 20 25 42
Passed (%) 57 (56–58) 65 (64 to 66) .61 68 (67 to 69) 62 (62 to 62) .63
Drivers with collisions (%) 38 (37 to 39) 15 (15 to 15) .11 16 (16 to 16) 24 (24 to 24) .44
Drivers with hazardous (0–1 m) FGW (%) 5 (5–5) 0 (—) .32 0 (—) 0 (—)
Driver with risky (1–2 m) FGW (%) 19 (19 to 19) 20 (20 to 20) .94 12 (12 to 12) 17 (16 to 17) .59
Average critical (<1 s) THW events 5.2 (3.9 to 6.6) 4.8 (4 to 5.5) .54 3.5 (2.8 to 4.2) 4.5 (3.9 to 5.1) .04
Average RT (s) 0.69 (0.63 to 0.75) 0.68 (0.64–0.73) .76 0.65 (0.61–0.69) 0.67 (0.64 to 0.71) .50
Average LP city driving (m) −0.2 (−0.3 to −0.1) 0 (−0.1 to −0.1) .02 −0.2 (−0.3 to −0.1) −0.1 (−0.1 to 0) .07
Average LP rural road (m) −0.1 (−0.2 to −0.1) 0 (0 to −0.1) <.001 −0.1 (−0.1 to 0) 0 (0 to 0.1) .04
Average LP motorway (m) 0.1 (0 to −0.2) 0.5 (0.2 to −0.8) .01 0.4 (0.1 to 0.6) 0.3 (0.2 to 0.4) .33
B–E = visual field clusters according to Fig. 1; FGW = failure to give way; LP = lateral position compared with control (+ to the left, − to the right); RT = reaction time; THW = time headway.

Statistical Considerations

The statistical tests used were independent t tests and z test. Logistic regression was performed with success rate as the criteria variable (passed/failed): age, sex, time since onset of stroke, and classification of visual field loss (Fig. 1) and average sensitivity in different clusters of the integrated visual field (Fig. 2) were used as explanation factors. The α level of P < .05 was always used.

Car Accidents among Dispensation Cases

One part of the evaluation of the simulator-based driving test was to conduct a follow-up of all individuals who had regained their driving licenses after a successful performance in the simulator. We used the Swedish Traffic Accident Data Acquisition database that includes all road accidents with personal injuries in Sweden reported by the police or emergency hospitals.

RESULTS

Participants

The stroke participants matched the control group by age but consisted mainly of men. Even if some patients experienced simulator sickness, all completed the scenario. Sixty-five percent of the stroke participants passed the simulator test (95% confidence interval, 57 to 72%, with Wilson procedure with a correction for continuity; Table 1). Younger patients were more successful than the older ones (Fig. 4). Fifty-one percent of individuals 70 years or older passed the test, compared with 70% of individuals younger than 70 years (P < .05 with z test of proportions).

F4
FIGURE 4:
Percentage of passed patients with stroke, in different age groups, with total number of individuals in each group.

Stroke versus Normal-sighted Participants

Statistical analysis of reaction time, time headway (with independent t tests), collisions, and failure to give way (with Fisher exact test) did not show significant differences between stroke participants and controls. Stroke participants drove more slowly than the normally sighted on motorway and rural roads (with independent t tests; Table 1). Reaction time could be measured only if the driver did not have his/her foot on the brake pedal before the object of interest was presented, and only reaction times between 0 and 2 seconds were included in the analysis. Reaction time was therefore measured in only 71% of possible events in the stroke group and 54% in the control group.

Performance and Safety Measures for Stroke Participants

Fisher exact test revealed that failed stroke participants had both more hazardous (0 to 1 m) and risky failure to give way (1 to 2 m). An independent two-tailed t test for critical time headway was also significant, meaning that failed drivers had a reduced ability to lower their velocity when they were close to the moving objects in the scenario. Using the same statistical method, reaction time for passed participants was significantly faster than for failed participants. Failed drivers drove a little more slowly on motorway and rural stretches but not in city driving (Table 1).

Taken together, the failed group was less safe and performed less well on all measures except collisions even if they drove a little more slowly on some stretches. This finding revealed that both safety measures and performance measures explained the success rate for the passed drivers.

Visual Field Loss as a Predictor for Driver Safety

The inclusion criteria for the visual field analysis were met in 122 individuals, 80% of all stroke participants. Individuals with homonymous hemianopia with macular sparing had the highest pass rate, and scotomatous defects had the lowest (Fig. 5). No significant differences between right- and left-sided visual field loss could be detected on any of the measures, except lane position, patients with hemianopia had their average lateral position significantly displaced toward the nonaffected side (Table 2). Logistic regressions were performed on success rate with age, sex, time since onset of stroke, and one of five different measures of visual field loss (1: classification according to Fig. 1, 2 to 5: mean sensitivity in an integrated visual field in different clusters according to Fig. 2). None of these analyses revealed any significant factors other than age (P < .01) (Table 3).

F5
FIGURE 5:
Percentage of passed patients after stroke, classified on different visual field defects, with total number of individuals in each group. Median age for each group is presented within brackets.
TABLE 3 - Multivariate of possible factors influencing pass/fail rate
Independent variable OR (95% CI) P
Analysis 1
 Average age (y) 0.94 (0.89–0.98) <.01
 Male sex 0.63 (0.08–3.43) .62
 Time since onset of stroke (y) 0.95 (0.84–1.07) .38
 VFL A <0.01 (NA) .99
 VFL B <0.01 (NA) .99
 VFL C <0.01 (NA) .99
 VFL D <0.01 (NA) .99
 VFL E <0.01 (NA) .99
 VFL F <0.01 (NA) .99
Analysis 2
 Age (y) 0.94 (0.89–0.98) <.01
 Male sex 0.6 (0.08–2.86) .55
 Time since onset of stroke (y) 0.95 (0.83–1.08) .39
 IVF0–10 0.98 (0.87–1.11) .80
 IVF10–20 1.01 (0.86–1.18) .93
 Right sided 0.77 (0.33–1.8) .55
Analysis 3
 Age (y) 0.93 (0.89–0.98) <.01
 Male sex 0.57 (0.08–2.76) .51
 Time since onset of stroke (y) 0.95 (0.83–1.08) .39
 IVFRight 0–10 1.03 (0.94–1.12) .52
 IVFLeft 0–10 0.94 (0.82–1.06) .29
 IVFRight 10–20 0.96 (0.87–1.05) .38
 IVFLeft 10–20 1.06 (0.95–1.19) .28
 Right sided 0.29 (0.04–2.05) .22
Analysis 4
 Age (y) 0.94 (0.89–0.98) <.01
 Male sex 0.58 (0.08–2.77) .52
 Time since onset of stroke (y) 0.95 (0.83–1.08) .41
 IVFSuperior 0–10 0.98 (0.86–1.1) .70
 IVFInferior 0–10 1.03 (0.9–1.17) .70
 IVFSuperior 10–20 1.00 (0.89–1.11) .95
 IVFInferior 10–20 0.99 (0.89–1.1) .84
 Right sided 0.79 (0.33–1.84) .58
Analysis 5
 Age (y) 0.94 (0.89–0.98) <.01
 Male sex 0.61 (0.08–2.92) .56
 Time since onset of stroke (y) 0.95 (0.84–1.08) .43
 IVFaffected side 0–10 0.99 (0.93–1.06) .83
 IVFaffected side 10–20 1.00 (0.92–1.09) .96
CI = confidence interval; IVF = mean sensitivity in an integrated visual field in different clusters according to Fig. 2; OR = odds ratio; VFL = visual field loss according to Fig. 1.

Car Accidents On-road among Dispensation Cases

The majority (96%) of all individuals with a passed test applied for a renewed driver's license. Among these, two individuals were rejected for a renewed license because of other medical circumstances. None of the 93 stroke participants with a regained license were involved in a motor vehicle accident according to the Swedish Traffic Accident Data Acquisition database in October 2020, 2.8 to 6.1 years (median, 3.8 years) after the simulator test.

DISCUSSION

This study of simulator-driving performance of individuals with different types of homonymous visual field loss showed that a classification of homonymous visual field loss from stroke might predict driver safety on a group level. Younger patients were considerably more successful than older. Failed participants with visual field loss did not have any more collisions than those who passed. Instead, they had more “near” collisions, as evidenced by the greater numbers of events with hazardous failures to give way and more critical time headways.

To the best of our knowledge, this is the largest study so far of simulated driving behavior for individuals with visual field loss after stroke. Other strengths of the study are the specified information about the visual field, the detailed simulator scenario, and that the patients performed the test with the aim of regaining their licenses, which guaranteed good participation.

However, the study also had important weaknesses. The control group was not sex matched with the study group. This may have affected the result because men and women may have different comfort with speed and willingness to take risks.30 In addition, the study group was slightly but significantly younger than the controls. Sharp turns could not be used (to avoid simulator sickness), which made the simulator less like reality. Looking at visual field classification, a possible trend could be detected with higher pass rate for sectoranopia and hemianopia compared with the remaining classifications. However, when age was included as a covariate in a logistic regression model, visual field classification was not a significant predictor. The most important deficiency of the study is that the normally sighted participants were not assessed in terms of passed and failed. Therefore, we do not know if all controls passed the test. At the same time, this group created the reference values to pass. It should also be noted that the high cost of the testing (2000 euros) could affect the selection of participants. Individuals with less affected driving and higher financial resources were probably more disposed to apply for the test. This could explain the high success rate of 65%.

Even individuals with severe visual field loss might be safe drivers, which conforms to previous research.10 All participants drove at almost the same speed and did not use speed as an adaptation for increased safety,31 which means that the participants with visual field loss who passed did not compensate by reducing speed. Our study showed a success rate of 64% for drivers with quadrantanopia compared with 56% for drivers with hemianopia. This finding is consistent with previous research of on-road driving performance, which has shown the distribution of 88 versus 73%.11 However, individuals with homonymous hemianopia with macular sparing were even safer drivers than individuals with quadrantanopia. We could also verify that lane position in hemianopia was displaced toward the nonaffected side,16 which supports the thesis that drivers try to increase the safety margin on their blind side. Most possible collisions came from the right for two reasons. First, this is the natural situation in right-hand traffic (parked cars, buses, bicyclists, etc.). Second, the time to react on object that interferes with your trajectory is shorter. Despite this fact, drivers with right-sided defects were not less successful than drivers with left-sided defects, which is difficult to explain. Because head and ocular movements were not recorded, this possible compensatory strategy could not be evaluated.

The individuals with renewed driver's licenses had no higher risk of on-road collisions. However, the significance of this result is unknown because of the low number of license renewals (93). Besides, road accidents are fortunately rare events, even during a period of years. Hence, it is not possible to conclude that the simulator assessment can yet discriminate between safe and unsafe drivers. However, the data support that the method is valid but needs further evaluation. This conforms to previous studies that have shown that simulator testing is a well-standardized method appropriate for assessment of driving performance in individuals with binocular visual field loss.32

CONCLUSIONS

In this large cohort, driver safety could not be predicted from the type of homonymous visual field loss. Even individuals with severe visual field loss might be safe drivers. Therefore, it seems reasonable to provide an opportunity for individualized assessments of practical fitness to drive in circumstances of licensing issues. An on-road test by a certified driving examiner is currently considered the clinical criterion standard. Driving simulators may provide a useful complement to evaluate responses to critical situations under safe, repeatable, and standardized conditions. However, the optimal cutoff level for such a test remains to be discovered.

REFERENCES

1. Bowers AR. Driving with Homonymous Visual Field Loss: A Review of the Literature. Clin Exp Optom 2016;99:402–18.
2. Owsley C. The Vision and Driving Challenge. J Neuroophthalmol 2010;30:115–6.
3. Bro T, Lindblom B. Strain out a Gnat and Swallow a Camel?—Vision and Driving in the Nordic Countries. Acta Ophthalmol 2018;96:623–30.
4. Katan M, Luft A. Global Burden of Stroke. Semin Neurol 2018;38:208–11.
5. Zhang X, Kedar S, Lynn MJ, et al. Homonymous Hemianopias: Clinical-anatomic Correlations in 904 Cases. Neurology 2006;66:906–10.
6. Gilhotra JS, Mitchell P, Healey PR, et al. Homonymous Visual Field Defects and Stroke in an Older Population. Stroke 2002;33:2417–20.
7. Falke P, Abela BM Jr., Krakau CE, et al. High Frequency of Asymptomatic Visual Field Defects in Subjects with Transient Ischaemic Attacks or Minor Strokes. J Intern Med 1991;229:521–5.
8. Townend BS, Sturm JW, Petsoglou C, et al. Perimetric Homonymous Visual Field Loss Post-stroke. J Clin Neurosci 2007;14:754–6.
9. Zhang X, Kedar S, Lynn MJ, et al. Natural History of Homonymous Hemianopia. Neurology 2006;66:901–5.
10. Elgin J, McGwin G, Wood JM, et al. Evaluation of On-road Driving in People with Hemianopia and Quadrantanopia. Am J Occup Ther 2010;64:268–78.
11. Wood JM, McGwin G Jr., Elgin J, et al. On-road Driving Performance by Persons with Hemianopia and Quadrantanopia. Invest Ophthalmol Vis Sci 2009;50:577–85.
12. Racette L, Casson EJ. The Impact of Visual Field Loss on Driving Performance: Evidence from On-road Driving Assessments. Optom Vis Sci 2005;82:668–74.
13. Bowers AR, Mandel AJ, Goldstein RB, et al. Driving with Hemianopia, I: Detection Performance in a Driving Simulator. Invest Ophthalmol Vis Sci 2009;50:5137–47.
14. Bowers AR, Ananyev E, Mandel AJ, et al. Driving with Hemianopia: IV. Head Scanning and Detection at Intersections in a Simulator. Invest Ophthalmol Vis Sci 2014;55:1540–8.
15. Papageorgiou E, Hardiess G, Ackermann H, et al. Collision Avoidance in Persons with Homonymous Visual Field Defects under Virtual Reality Conditions. Vision Res 2012;52:20–30.
16. Bowers AR, Mandel AJ, Goldstein RB, et al. Driving with Hemianopia, II: Lane Position and Steering in a Driving Simulator. Invest Ophthalmol Vis Sci 2010;51:6605–13.
17. Matthew BP, Bowers AR, Albu A, et al. Hazard Detection by Drivers with Paracentral Homonymous Field Loss: A Small Case Series. J Clin Exp Ophthalmol 2011;2011:1.
18. Smith M, Mole CD, Kountouriotis GK, et al. Driving with Homonymous Visual Field Loss: Does Visual Search Performance Predict Hazard Detection?Br J Occup Ther 2015;78:85–95.
19. Bahnemann M, Hamel J, De Beukelaer S, et al. Compensatory Eye and Head Movements of Patients with Homonymous Hemianopia in the Naturalistic Setting of a Driving Simulation. J Neurol 2015;262:316–25.
20. Nyberg J. Welfare Consequences for Individuals with License Withdrawal Due to Visual Field Loss. Linköping, Sweden: The Norwegian Road and Transport Research Institute; 2019. Available at: http://www.diva-portal.org/smash/record.jsf?pid=diva2%3A1362594&dswid=-1163. Accessed July 21, 2022.
21. Bro T, Andersson J. The Effects of Visual Field Loss from Glaucoma on Performance in a Driving Simulator. Acta Ophthalmol 2022;100:218–24.
22. Bro T, Andersson J. The Effects of Visual-field Loss from Panretinal Photocoagulation of Proliferative Diabetic Retinopathy on Performance in a Driving Simulator [published online January 13, 2022]. Eye (Lond). doi:10.1038/s41433-021-01832-3.
23. Bro T, Andersson J. The Effects of Visual Field Loss from Optic Disc Drusen on Performance in a Driving Simulator [published online March 21, 2022]. Neuro-ophthalmology. doi:10.1080/01658107.2022.2038640.
24. Transportstyrelsens Författningssanling (TSFS). The Swedish Transport Agency's Regulations and General Advice on Medical Requirements for Possession of a Driving License, etc. Consolidated Electronic ed. 2010. Available at: https://www.transportstyrelsen.se/TSFS/TSFS%202010_125k.pdf. Accessed July 21, 2022.
25. Nordmark S, Jansson H, Palmkvist G, et al. The New VTI Driving Simulator—Multi Purpose Moving Base with High Performance Linear Motion. In: Proceedings of the Driving Simulation Conference (DSC) 2004. Paris, France; 2004:45–55.
26. The Swedish Driving Test from a Cognitive Perspective. Criteria for Examination Tasks—Standardized Driving Test. Borlänge, Sweden: Vägverket (Swedish Road Authority); 2000.
27. The Swedish National Road and Transport Research Institute (VTI). Protocol for Evaluation of Driving Ability, 2015. Linköping, Sweden: VTI; 2015.
28. Mamdoohi AR, Zavareh MF, Hyden C, et al. Comparative Analysis of Safety Performance Indicators Based on Inductive Loop Detector Data. Promet-Zagreb 2014;26:139–49.
29. Östlund J, Peters B, Thorslund B, et alThe Adaptive Integrated Driver-Vehicle Interface Deliverable (AIDE). Driving Performance Assessment Methods and Metrics. Information Society Technologies Programme; AIDE D2.2.5. Available at: http://www.aide-eu.org/pdf/sp2_deliv_new/aide_d2_2_5.pdf. Accessed September 2, 2021.
30. Ferrante C, Varladi V, De Blasiis MR. Gender Differences Measured on Driving Performances in an Urban Simulated Environment. In: Cassenti D, ed. Advances in Human Factors and Simulation; AHFE 2019. Advances in Intelligent Systems and Computing. Vol. 958. Cham, Switzerland: Springer International Publishing AG; 2020:144–56.
31. Brémond R, Dommes A, Engel L. Driving at Night with a Cataract: Risk Homeostasis?Transp Res F Traffic Psychol Behav 2018;53:61–73.
32. Ungewiss J, Kubler T, Sippel K, et al. Agreement of Driving Simulator and On-road Driving Performance in Patients with Binocular Visual Field Loss. Graefes Arch Clin Exp Ophthalmol 2018;256:2429–35.
Copyright © 2022 American Academy of Optometry