Implications of monocular vision for racing drivers.

We performed two experiments to investigate how monocular vision and a monocular generalized reduction in vision (MRV) impact driving performance during racing. A total of 75 visually normal students or professional racing drivers, were recruited for the two experiments. Driving performance was evaluated under three visual conditions: normal vision, simulated monocularity and simulated monocular reduction in vision. During the driving scenario, the drivers had to detect and react to the sudden intrusion of an opponent's racing car into their trajectory when entering a turn. Generalized Linear Mixed Models (GLMMs) and ANOVA were then used to explore how monocular vision and monocular reduction in vision affect drivers' performance (crash and reaction time) while confronting them with critical situations. The results show that drivers under monocular condition are from 2.1 (95% CI 1.11-4.11, p = .024) to 6.5 (95% CI 3.91-11.13; p = .0001) times more likely to collide with target vehicles compared with their baseline (binocular) condition, depending on the driving situation. Furthermore, there was an average increase in reaction time from 64 ms (p = .029) to 126 ms (p = .015) under monocular condition, depending on the critical driving situation configuration. This study objectively demonstrates that monocularity has a significant impact on driving performance and safety during car racing, whereas performance under monocular reduction in vision conditions is less affected.

Introduction

Motor racing is a particularly dangerous sport that requires being in perfect physical condition, having fast reaction times and having good vision. During car racing, where the velocities are very high, the driver’s vision is essential to drive a powerful car safely and to detect and react in short timeframes to sudden and unexpected events. Yet, to date, there is a lack of literature for determining the visual requirements of a racing driver.

Monocular vision is an impairment that intuitively would not allow someone to perform motor racing. Although it has been reported that a one-eyed racing driver managed to complete an entire season with only one incident[1], during seven laps, the driver failed to see and react to a penalty flag that was being waved on his normally sighted left side. To date, the literature contains no evidence of an increased risk of crashing among one-eyed drivers and that literature thus is unhelpful regarding the visual requirements of a racing driver[1]. Beyond the juridical and human rights issues, it is essential to determine if monocular drivers represent a potential risk to themselves and also to other racing drivers or even spectators at a race.

To our knowledge, no study has explored the impact of monocularity on driving in a racing condition. The only paper dealing with this issue is that of Westlake [1]. However, as mentioned by the author, monocularity has been studied only in the conventional driving research field. Indeed, the literature on performance and safety of monocular drivers has largely focused on studies of commercial drivers (eg truck, delivery vehicle, taxi, bus) and have reported conflicting results. For example, some research has found that monocular drivers have more crashes and convictions [2-4] and poorer driving performance [5]. Other studies have observed that, compared with control groups or a national reference group, drivers with real or simulated monocular visual field loss have equivalent accident or conviction rates and driving performance [6-8]. Furthermore, McKnight et al. [9] assessed 40 monocular and 40 binocular commercial drivers and found no differences with respect to visual search, lane placement, clearance judgment, gap judgment, hazard detection, or information recognition. However, these authors did find that monocular drivers were less adept at reading signs at a distance during both daytime and nighttime driving than were binocular drivers. Furthermore, the definition of ‘‘monocularity” has varied widely in the literature of conventional driving, ranging from drivers who have a total absence of function in one eye, to drivers who have visual function in one eye below the minimum level for having a license. Sometimes, no definition is provided. Thus, the impact of monocularity on conventional driving has not been exhaustively addressed. Finally, it is difficult to extrapolate the studies that have been performed on conventional driving to motorsport, which is a more demanding activity during which drivers are taking risks in a competitive context.

The lack of consensus on the impact of monocularity on driving is relatively surprising given the importance of visual inputs for this activity. Indeed, monocularity causes three important deficits that can affect driving. The first deficit concerns the reduction of the visual field. A monocular individual has a peripheral field deficit of 20 to 40 degrees temporally. Thus, monocular drivers must move their head or eye to obtain information on the temporal side of their non-functional eye. The impact on driving strongly depends on the importance of this information based on the context and the ability of the driver to adapt effectively their visual comportment to retrieve that information. The second deficit is the loss of binocular summation. It has been observed that there exists a superiority of binocular over monocular visual performance with fine central acuity being further enhanced through binocular summation [10]. Although it has been demonstrated in early enucleated patients and in subjects who are functionally monocular, that the contrast sensitivity [11,12] or visual acuity of the functional eye [13,14] is better than the visual functioning of the better eye of normal binocular subjects, the performance of monocular individuals remains inferior to those with binocular vision [14,15]. Finally, monocular drivers have a poorer perception of depth than binocular individuals [9],[14]. This may not be as important as the other deficits as the racetrack is a dynamic environment. Dynamic stereopsis is weakly correlated with static stereopsis and is reduced with increasing angular velocity [16,17]. The distances and operating speeds of stereopsis suggest that, in many circumstances, stereopsis would not be useful to the race driver. In addition, the driver can use a number of other indicators to assess the perception of depth, such as the apparent size of the objects and their expansion over time, shadows, and the space between the front of the car they are driving and the car in front [1].

Although monocular drivers could be potentially seen as high-risk drivers, to date, no study has been conducted to determine if monocularity could be responsible for an objective decrease in visual performance during race driving and could impact safety. In addition, no study has evaluated the impact of a monocular generalized reduction in vision (MRV), such as that which occurs in patients with congenital amblyopia. The present study was performed to address the impact of monocularity and MRV on racing by simulating monocular visual field defects in a group of racing drivers to assess the impact of these defects on driving performance. Because it is difficult to perform these assessments in a standardized and safe way during real driving, we used a sophisticated driving simulator.

We conducted two experiments to evaluate the impact of monocularity and MRV on racing performance while driving. To our knowledge, this is the first study to do so. We predicted that under identical conditions, monocular drivers and drivers with monocularly reduced vision would be at higher risk for accidents and have longer reaction times than binocular drivers with normal visual function in both eyes.

General method

Overview

The study was conducted at the FFSA Autosport Academy (Le Mans, France) and at the Institut de la Vision, Paris, France. The study was approved by the Ethics Committee of the French Society of Ophthalmology and adhered to the tenets of the Declaration of Helsinki. All participants, or parents of minor participants, gave written informed consent to participate. No participants received compensation for participating. Seventy-five participants were recruited. All participants were professional racing drivers or students at the FFSA Autosport Academy. No exclusion criteria, such as demographic requirements, were applied when recruiting participants. The data from some participants were excluded from analysis because these participants did not complete all the tasks or because of failures in data recording.

Visual assessment

A visual assessment was performed to select only visually non-impaired participants and to be sure that no visual pathologies could affect the results of the studies. The assessment included a complete medical and ocular history, tests of binocular and monocular visual acuity (ETDRS chart in Metrovision), colour perception (Ishihara plates), contrast sensitivity (Metrovision), visual field (kinetic perimetry using a Goldmann perimeter and a III/4e stimulus), refraction (TOPCON KR-800S), and spectacle lens and frame measurements (Essilor CLE 60). We also determined the sensory eye dominance to select the eye to be experimentally impaired (Red Lens test) and to determine which Ryser filter decreased the visual acuity of the dominant eye to 0.5 log MAR.

Apparati

Simulation of visual impairments

Three visual conditions were employed in the study: baseline normal visual acuity (≤0 logMAR visual acuity) in both eyes, reduced vision (0.5 logMAR visual acuity) in the dominant eye, and monocular vision. The visual acuity restrictions were selected because they are most commonly cited as resulting in impaired driving performance [18]. MRV was achieved by placing a Ryser filter either on one lens of a pair of protective glasses (if the driver did not normally wear glasses) or on one lens of the driver’s personal glasses. Monocular vision was produced by placing an occlusive patch over the participant’s dominant eye.

Driving simulator

Streetlab’s fixed-base driving simulator (Paris, France) was used in this study. The driving simulation was handled by SCANeR™ studio Software and displayed on an immersive visual system consisting of three full HD LCD 65” screens that provided a forward field of view of 180 degrees. The simulator was equipped with a fixed-base car seat, a Fanatec force-feedback steering wheel, Fanatec braking and throttle pedals, and speakers. Both auditory feedback (i.e., engine noise) and tactile feedback (i.e., when contact was made with the curb in a corner the steering wheel jerked) were provided. Two video cameras were used to provide recordings of the driver’s sessions. The driving parameters were recorded at the frequency of 60 Hz. The virtual driving environment was the former racetrack in Barcelona-Catalunya, Spain, and the vehicle being driven was a GP2 type car.

Procedure

Participants attended two testing sessions. The first session was the selection session, during which potential participants completed the visual assessment and a Motion Sickness History Questionnaire MSHQ [19] to prevent unnecessary exposure to the driving simulation. The total duration of this session was 1h.

The second session was the driving simulator assessment. A 10–15 minutes practice drive preceded the test drive to allow the participant to become acquainted with the driving simulator and the track. The examiner issued a set of instructions to each subject before the run. Each racing driver was tested just once under each of the three visual conditions in a pseudo-random order. Participants could have a break whenever they wanted, so as to avoid Simulator Adaptation Syndrome (SAS). The total duration of the driving simulator session was around 1h.

Statistical analysis

The minimum number of participants required was determined by an a priori power analysis. Our objective was to compare the occurrence of accidents by visual condition and report the estimated ORs using a logistic regression (the analysis was modified during the revision process). To determine the number of observations required per group (n), we assumed that the accident rate expected in the control condition (Baseline) would be 5% and 20% in the monocular condition with a power of at least 90% and an α of .05. We collected 100 observations per group, or a minimum number of 17 participants, taking into account the number of tests per driving situation.

A Generalized Linear Mixed Model (GLMM) with a logit link function was applied, using an R package named lme4 [20], in R [21] to study the likelihood (odds ratio) of collision risk given the visual condition of the racing drivers. The GLMM was used to account for the possible unobserved heterogeneity caused by repeated measures from the same individuals. The binary response variable was the involvement in an accident whereby 1 = accident and 0 = no accident. We set the normal vision condition as the reference condition. We assessed overdispersion for every model (the sum of the squared Pearson residuals should be χ2 distributed)[22].

We also analyzed driver reaction time in crash avoidance using two-way repeated measures ANOVA. Data were controlled for normality, homogeneity of variances, and sphericity to insure adequate tests. In cases where the assumption of sphericity was violated, the Greenhouse-Geisser correction was applied. A threshold of p<.05 was considered significant. Dunnett’s and Newman–Keuls post-hoc multiple comparison tests were conducted to identify where differences among means existed.

We analyzed the relationships between reaction times and accident rate for each visual condition, regardless of the type of situation, using Pearson’s correlation.

Experiment 1

The purpose of Experiment 1 was to investigate how simulated recent monocularity or MRV affects driving performance and particularly the detection of hazardous situations in front of the driver, within a horizontal field of vision of 120°.

Method

Participants

31 participants initially were recruited for the experiment; however, three were withdrawn during testing due to symptoms of SAS. It also was necessary to exclude from analysis 10 participants with incomplete data (absence of data in at least 1 condition) or irrelevant data. The final sample was composed of 18 racing drivers ranging in age from 14 to 36 years (mean: 21.88 ± 7.47 years; 1 F, 17M). All participants were amateur or professional racing drivers with a mean of 6.8 ± 4.8 years of experience in AutoSport.

Driving scenarios

Three different dynamic hazardous situations were developed according to the impaired-eye side (Fig 1). All three require the drivers to detect and react to a sudden intrusion of an opponent’s racing car into the participant's trajectory in a turn. All opposing cars start from the edge of the track, within a visual field of 120 degrees maximum when the driver looks straight ahead. In the first situation, the turning direction (the direction where the drivers must look) and the location of the entering car are congruent with the visual impairment. We call this situation congruent-congruent (CC). In the second setting, the turning direction is congruent with the visual impairment, but the location of the entering car is incongruent with the visual impairment. We term this situation congruent-incongruent (CI). In the third situation, the turning direction (the direction where the drivers must look) is incongruent with the visual impairment, but the location of the entering car is congruent with the visual impairments. We call this situation incongruent-congruent (IC).

Fig 1

The three hazardous situations, respectively CC, CI and IC situations.

The triangular shape corresponds to the blind area of the drivers.

To prevent any anticipatory behavior, billboards hide the entering cars from the view of the drivers. These billboards have been placed at the inside and outside of every turn. Throughout the driving, the participant follows an opponent car with a fixed-time headway of 4s that cannot be overtaken. They also are followed by an opponent car with a fixed-time headway of 1.5s. This immersion process gives the impression to the driver of being in a race.

Six potential hazards were placed at different locations of the track for each of the three scenarios (CC, CI, IC), for 18 randomized situations. To experience all of the situations, the drivers must complete 15 laps. Six versions of the scenarios (three for each blind side) were generated to counterbalance the presentation order of the challenging events and so that each repeated condition had a different situation.

Results

Table 1 lists the visual characteristics of the 18 tested subjects with complete data. No subject had any visual pathology or visual abnormality. All had normal visual acuity overall (mean: -0.23 ± 0.06 logMAR), and their MRV reached the desired visual acuity level (mean: 0.49 ± 0.04 logMAR).

Table 1

Visual characteristics of tested subjects.

Participants	Age (yr)	Visual Acuity (logMAr)			Visual Field Diameter Goldman III/4e		Deficit within Central 10 Degrees	Contrast SensitivityAbnormality	Color Vision Abnormality	Visual Acuity with Ryser
		left	Right	Booth	Horizontal	Vertical
RD1	20	-0,30	-0,26	-0,28	176	132	no	no	no	0,46
RD2	28	-0,28	-0,22	-0,26	174	119	no	no	no	0,50
RD3	27	-0,16	-0,18	-0,26	176	120	no	no	no	0,52
RD4	31	-0,20	-0,26	-0,20	176	123	no	no	no	0,52
RD5	30	0,02	-0,10	-0,14	176	121	no	no	no	0,50
RD6	36	-0,28	-0,28	-0,30	176	112	no	no	no	0,56
RD7	21	-0,10	-0,28	-0,24	174	123	no	no	no	0,50
RD8	17	-0,26	-0,24	-0,28	173	120	no	no	no	0,50
RD9	28	-0,26	-0,28	-0,26	172	111	no	no	no	0,52
RD10	34	0,12	0,06	-0,06	174	120	no	no	no	0,50
RD11	16	-0,26	-0,28	-0,30	175	122	no	no	no	0,50
RD12	15	-0,16	-0,20	-0,24	176	124	no	no	no	0,46
RD13	15	-0,10	-0,16	-0,18	176	119	no	no	no	0,50
RD14	15	-0,24	-0,26	-0,30	171	122	no	no	no	0,40
RD15	17	-0,08	-0,06	-0,16	171	118	no	no	no	0,52
RD16	16	-0,26	-0,22	-0,30	172	118	no	no	no	0,54
RD17	16	-0,22	-0,20	-0,22	170	119	no	no	no	0,44
RD18	15	-0,2	-0,2	-0,2	171	110	no	no	no	0,44

Driving performance

We evaluated 18 participants who passed all three visual conditions. In each condition, drivers were confronted with three types of dangerous situations. Six tests were proposed for each dangerous situation. Eliminating missing data related to the failed or incorrectly performed tests, we collected 232 observations for the CC situations, 220 observations for the CI situations and 214 observations for the IC situations.

As a first step, we conducted a series of GLMMs to determine if multiple visual conditions (monocularity or MRV) were predictive of an accident and the odds of a collision happening given the visual condition. We conducted a GLMM for each potential hazard situation; i.e, CC, CI and IC (Table 2). The analysis indicated the absence of significant overdispersion for CC situations model (χ2 = 228.567; ratio = 1.002; rdf = 228; p = .477); for IC situations model (χ2 = 204.360; ratio = 0.973; rdf = 210; p = .597); and for CI situations model (χ2 = 203.934; ratio = 0.944; rdf = 216; p = .712).

Table 2

Parameter estimates from the three GLMMs for the 3 driving situations related to the occurrence of collisions for each visual condition.

Parameter	Estimate	SE	Z value	p	OR	95% CI low	95% CI high
Model for CC Situations
Intercept	-0,701	0,247	-2,839	0,001
Baseline vision	0,00	0,00
MRV	0,484	0,335	1,447	0,148	1,623	0,845	3,153
Monocular	0,751	0,333	2,258	0,024	2,119	1,112	4,111
Model for CI Situations
Intercept	-0,710	0,282	-2,514	0,012
Baseline vision	0,00	0,00
MRV	0,122	0,361	0,337	0,736	1,129	0,556	2,310
Monocular	0,286	0,352	0,812	0,417	1,331	0,668	2,679
Model for IC Situations
Intercept	-1,849	0,360	-5,136	0,0001
Baseline vision	0,00	0,00
MRV	1,001	0,432	2,314	0,021	2,720	1,188	6,587
Monocular	1,129	0,428	2,637	0,008	3,094	1,368	7,456

SE: standard error; OR: odds ratio; CI: confidence interval.

For the CC situations, monocularity was the only significant condition in the model (p = .024). Model parameters showed that racing drivers in the monocular condition have 2.1 (95% CI 1.112–4.111) greater odds of having a collision than racing drivers in the Baseline binocular condition. For CI situations, the GLMM showed that neither monocularity nor MRV was significant. Finally, for IC situations, both monocularity and MRV were significant (p = .008 and p = .021, respectively). Model parameters showed that recently monocular racing drivers had 3.094 (95% CI 1.368–7.456) greater odds of having a collision compared with the baseline condition, and racing drivers with MRV had 2.72 (95% CI 1.188, 6.587) times greater odds of having a collision compared with the baseline visual condition.

Reaction time

Mean reaction times for the three hazard situations are presented in the Fig 2. The two-way repeated measures ANOVA showed significant main effects of the visual condition, F(2, 34) = 5.587, p = .008, and a tendential effect of the hazard situation, F(2, 34) = 2.695, p = .082. The interaction effect was non-significant, F(4, 68) = .9559, p = .437.

Fig 2

The effects of visual conditions on mean reaction times for each hazard situation.

The error bars represent the 95% confidence interval.

We performed separate one-way repeated ANOVAs for each hazard situation and then conducted multiple comparisons to assess the simple main effect of each visual condition. For the CC hazard situation, the ANOVA revealed a significant main effect of visual condition, F(2, 34) = 3.630, p = .049. Dunnett's multiple comparison showed a significantly longer RT under the monocular visual condition (800 ms) than under the baseline visual condition (736 ms, p = .029). We did not observe a difference in RT between the MRV condition and the baseline visual condition (p = .837). Furthermore Newman-Keuls post-hoc test showed a significant longer RT under the monocular visual condition (800 ms) compared with the MRV condition (748 ms, p = .049). For the CI hazard situation, the effect of visual condition was not significant, F(2,34) = 1.110 p = .341. For the IC hazard situation, the analysis revealed a significant main effect of both visual conditions, F(2, 34) = 4,407, p = .019, with Dunnett’s multiple comparisons indicating that RT was significantly longer under the monocular visual condition (874 ms) than baseline visual condition (748 ms, p = .015) and tended to be significantly longer under the MRV condition (846 ms) than baseline visual condition (748 ms, p = .064). Newman-Keuls post-hoc test failed to reveal a statistical difference between the monocular visual condition and the MRV condition (p = .531).

Correlations between reaction times and accident rate

As the variables were normally distributed, we used a Pearson correlation

Regarding the baseline visual condition, the results of the Pearson correlation indicated that there was no significant association between reaction times and accident rates, (r = .28, p = .26). Regarding the MRV visual condition, the results of the Pearson correlation indicated that there was a significant positive association between reaction times and accident rates, (r = .64, p = .004). Finally, regarding the monocular visual condition, the results of the Pearson correlation indicated that there was a significant positive association between reaction times and accident rates, (r = .48, p = .044).

Experiment 2

The purpose of Experiment 2 was to investigate how simulated recent monocularity or MRV affects driving performance and particularly the detection of hazard situations entering the horizontal field of vision from the rear.

44 participants were recruited for the second experiment. Four were discarded due to early symptoms typical of SAS. It also was necessary to exclude from the analysis nine participants with incomplete data. The final sample thus was composed of 31 racing drivers, all men, ranging in age from 15 to 48 years (mean: 25.16 ± 9.91 years). All participants were amateur or professional racing drivers with a mean of 8.46 ± 7.82 years of experience in Autosport. 15 were formula racing drivers, six were kart-racing drivers, six were GT racing drivers, one was a rally-racing driver and three were driving instructors.

Two different dynamic hazardous situations were developed (Fig 3), both of which require the drivers to detect and react to the sudden intrusion of an opponent’s racing car into the participant's trajectory when entering a turn. In those situations, the driver is overtaken by an opponent vehicle as the driver’s car approaches a turn. The opponent vehicle appears suddenly, accelerates to overtake the driver, and enters the visual field of the driver at one of two visual angles: 75° or 60°, assuming that the driver is looking straight ahead in the direction their vehicle is moving. When it reaches one of the above-mentioned angles, the opponent vehicle stabilizes its speed to stay at this angle position until the driver breaks and turns the steering wheel. To prevent the presence of the opponent car in the visual field from being detected by the driver using their near visual field (less than 60°), the rear-view mirrors have been obstructed visually. This provision is consistent with the report by drivers engaged with an opponent that they sometimes neglect to look in their rearview mirror and can be surprised by the intrusion of another opponent.

Fig 3

The two hazardous situations, respectively CC and II situations for experiment 2.

The triangular shape corresponds to the blind area of the driver.

In the first situation, the turning direction (the direction where the driver must look) and the location of the entering car are congruent with the visual impairment. We call this situation congruent-congruent (CC). In the second situation, both the turning direction of the driver’s car and the location of the entering car are incongruent with the visual impairment. We call this situation incongruent-incongruent (II). Different locations were chosen for both CC and II situations according to the side of the impaired eye.

The session consisted of eight trials for the CC and four trials for the II randomized situations located at different turns of the circuit. Six versions of each situation (three for each blind side) were generated to counterbalance the presentation order of the challenging events and so that each repeated condition had a different situation.

Table 3 lists the visual characteristics of the 31 participants. No participant had any visual pathology or visual abnormality (mean visual acuity: -0.23 ± 0.05 logMAR) and their visual acuity for MRV reached the desired visual acuity level (mean: 0.49 ± 0.04 logMAR).

Table 3

Visual characteristics of tested subjects.

Participants	Age (yr)	Visual Acuity (logMAr)			Visual Field Diameter Goldman III/4e		Deficit within Central 10 Degrees	Contrast SensitivityAbnormality	Color Vision Abnormality	Visual Acuity with Ryser
		left	Right	Booth	Horizontal	Vertical
RD1	17	-0,28	-0,22	-0,30	174	123	no	no	no	0,56
RD2	29	-0,26	-0,20	-0,26	176	107	no	no	no	0,52
RD3	21	-0,08	-0,16	-0,20	171	98	no	no	no	0,46
RD4	46	-0,18	-0,20	-0,26	176	121	no	no	no	0,52
RD5	15	-0,20	-0,26	-0,26	176	124	no	no	no	0,56
RD6	16	-0,26	-0,26	-0,30	176	124	no	no	no	0,54
RD7	25	-0,28	-0,28	-0,28	173	118	no	no	no	0,50
RD8	32	0,020	0,02	-0,08	175	121	no	no	no	0,56
RD9	23	-0,02	-0,12	-0,22	168	117	no	no	no	0,52
RD10	21	-0,26	-0,22	-0,28	176	123	no	no	no	0,50
RD11	15	-0,10	-0,16	-0,18	176	119	no	no	no	0,50
RD12	35	-0,24	-0,28	-0,30	174	117	no	no	no	0,50
RD13	19	-0,28	-0,28	-0,28	171	121	no	no	no	0,50
RD14	17	-0,10	-0,16	-0,20	173	115	no	no	no	0,52
RD15	19	-0,14	-0,16	-0,20	171	117	no	no	no	0,50
RD16	36	-0,24	-0,24	-0,24	175	124	no	no	no	0,50
RD17	21	-0,18	-0,18	-0,24	171	117	no	no	no	0,54
RD18	26	-0,06	-0,12	-0,22	175	121	no	no	no	0,62
RD19	35	-0,14	-0,02	-0,14	171	122	no	no	no	0,50
RD20	16	-0,22	-0,20	-0,22	170	119	no	no	no	0,44
RD21	31	0,06	-0,20	-0,28	166	124	no	no	no	0,56
RD22	23	-0,28	-0,22	-0,30	170	117	no	no	no	0,54
RD23	15	0,02	-0,10	-0,10	172	123	no	no	no	0,48
RD24	15	-0,26	-0,20	-0,26	170	117	no	no	no	0,60
RD25	20	-0,18	-0,14	-0,22	176	123	no	no	no	0,48
RD26	22	-0,20	-0,06	-0,28	173	112	no	no	no	0,46
RD27	48	-0,20	0,06	-0,20	168	105	no	no	no	0,40
RD28	47	-0,16	-0,16	-0,22	170	90	no	no	no	0,50
RD29	24	-0,24	-0,26	-0,28	170	115	no	no	no	0,54
RD30	15	-0,16	-0,20	-0,22	170	114	no	no	no	0,52
RD31	36	-0,20	+0,12	-0,26	173	115	no	no	no	0,50

We evaluated 31 participants who passed all three visual conditions. In each condition the drivers were confronted with two types of dangerous situations. Eight tests were proposed for the CC situations and four tests for the II situations. If we eliminate the failed or incorrectly performed tests, we collected 658 observations for the CC condition and 325 observations for situation II.

We conducted a series of GLMMs to determine if monocularity or MRV were predictive of an accident and the odds of a collision occurring, given the specific visual condition. A GLMM was conducted for both the CC and II hazard situations (Table 4). The analysis indicate the absence of significant overdispersion for CC situations model (χ2 = 590.933; ratio = 0.903; rdf = 654; p = .963) and for II situations model (χ2 = 242.486; ratio = 0.973; rdf = 320; p = .999).

Table 4

Parameter estimates from the two GLMMs for the 2 driving situations related to the occurrence of collisions for each visual condition.

Parameter	Estimate	SE	Z value	p	OR	95% CI low	95% CI high
Model for CC Situations
Intercept	-2,041	0,274	-7,448	0,0001
Baseline vision	0,00	0,00
MRV	0,376	0,273	1,377	0,168	1,457	0,851	2,523
Monocular	1,87	0,263	7,102	0,0001	6,493	3,911	11,133
Model for II Situations
Intercept	-1,554	0,348	-4,463	0,0001
Baseline vision	0,00	0,00
MRV	0,083	0,351	0,239	0,811	1,087	0,538	2,211
Monocular	0,112	0,360	0,312	0,755	1,119	0,542	2,321

Monocularity in the CC situations model was the only significant condition (p<.0001). Model parameters showed that racing drivers in the monocular condition have 6.493 (95% CI 3.911–11,133) greater odds of having a collision compared with racing drivers in the baseline condition.

Discussion

This study investigated the impact of simulated recent monocularity and MRV on simulated racing driving performance. The results of the study demonstrate that driving under racing conditions with simulated recent monocularity affects driving performance. Specifically, the results of the two experiments show that, compared with the baseline condition, drivers under the recent monocular condition are 2.1 to 6.5 times more likely to collide with target vehicles, depending on the driving situation. The driving situations that pose problems in monocular conditions are those in which the target to be detected and avoided appears on the blind-eye side (CC and IC situations for experiment 1 and CC for experiment 2). The maximum risk in all situations is reached for the situation CC in experiment 2. In this situation, the target initially is outside the visual field and rapidly enters the field of view. The results therefore seem to indicate that the more the target to be detected is in the peripheral vision on the blind side, the greater is the risk of collision. This is explained by the fact that the more the target to be detected is in the periphery on the blind side, the later it will be detected, leaving less time and, therefore, less chance for the driver to avoid a collision. The increase in reaction time, observed for drivers in the monocular vision condition (the IC [126 ms] and CC [64 ms] situations in experiment 1), seems to confirm the above results. In addition, correlation analysis allows us to observe that the longer the reaction time increases in this visual condition, the greater the risk of an accident. On the other hand, the results of the CI situation of experiment 1, and the II situation of experiment 2, characterized by the fact that the targets appear on the "healthy" side of the eye, show that these situations do not pose a problem for drivers under the monocular condition.

The simulated recent MRV condition seems to have a limited impact on driving. In the IC situation, there is both an increased risk of collision, where drivers are 2.7 times more likely to be involved in a crash, and a reaction time which tended to be longer than under normal vision conditions (98ms). It is possible that the drivers in this setting optimize their trajectory by moving their head towards the tangent point or the apex. By this head movement, it is possible that the appearance of the vehicle is outside their visual field and so would consequently be detected later, making the collision more difficult to avoid. Nevertheless, correlation analysis allows us to observe that, for the monocular visual condition, the longer the reaction time increases, the greater the risk of an accident, whereas this is not the case in baseline visual condition.

The results of this study, and specifically in experiment 2, could be explained by the fact that reduced central acuity (in one eye) is not as important in racing as loss of peripheral vision and that information obtained from the peripheral field is useful or at least relevant for race car driving. Peripheral vision has characteristics that are very different from central vision. Firstly, peripheral vision makes it possible to assess the entire visual scene, especially its spatial layout, more quickly than with eye movements (and, therefore, even more quickly than with head movements). The basic understanding of a scene can be completed in less than 100ms [23,24], which would be impossible without peripheral vision [25]. It also has the capacity to extract properties for a wide variety of stimuli, such as the size of the object [26], object orientation [27], and even the average emotion of a group of faces [28,29]. It also is possible to extract the mean pedestrian heading [30] or detect movement [31] through peripheral vision that can be very useful in the context of driving. The evaluation of the average characteristic value of a group of similar objects (pedestrians, bicycles, etc.), commonly called ensemble perception, is accomplished at the expense of the ability to discriminate more finely the information specific to each object taken individually [25].

Thus, peripheral vision allows the driver to detect elements entering the visual field, acquire information on these elements and, if needed, direct the eyes to process this information more finely. In addition, it has been shown that the identification of objects on the periphery is largely facilitated via pre-saccadic attention [32]. Thus, the analysis of the environment via the peripheral vision does not require perfect central vision. A car entering in the visual field of the driver will be very quickly detected and the driver will certainly not even have to look towards the object to infer that it is a car, a truck or a motorcycle. This could explain why the simulated recent MRV is less affected than the recent monocularity condition. Our results, especially those of experiment 2, suggest that the preserved peripheral visual field in drivers with recent MRV are able to perceive the appearance of a car entering their visual field and, thus, avoid a collision. On the other hand, the absence of peripheral vision in a simulated recent monocular condition strongly penalizes the driver.

The results of this study show that acute monocularity can impact the safety of racing drivers. Even though the participants in this study were professional drivers in racing conditions, we believe that it is reasonable to extrapolate these results to conventional driving. Indeed, the configurations of the situations tested in our experiments may well apply to an on-road driving condition. For example, they may represent an overtaking situation or a sudden insertion situation into the vehicle's trajectory at a crossroads or the untimely crossing of a pedestrian outside a normal crosswalk. As indicated in the introduction that there is a lack of consensus in the literature on the impact of monocularity on safety and driving performance. This is partly due to the operational definition of monocularity that varies among studies ranging from no function at all in one eye to one eye with reduced visual function to some extent (usually visual acuity). However, we see the importance of defining the phenomenon precisely in this study as our results show that the impact is not at all the same between a recent monocular condition and a recent MRV condition.

Nevertheless, we must remain cautious in this generalization as the dynamics between sports and classic driving are not the same. Thus, it would be interesting to reproduce the design of our study in this context of open-road driving.

There are certain limitations to our study. Firstly, as the study drivers did not have a pre-existing visual impairment, the results relate only to individuals who either have suddenly become monocular or have experienced a sudden monocular generalized visual loss. Such individuals have not had time to adapt to the vision alteration by putting in place behavioral strategies to deal with the visual impairment. It would be interesting to evaluate drivers with long-standing or even congenital monocularity or MRV and who have had time to adapt to their visual disability to see if they are using strategies to deal effectively with these situations.

Another limitation of our study relates to the methodology used in experiment 2. To meet the objectives of this experiment, we forced the drivers to drive without their mirrors. Indeed, we wanted to evaluate the impact of monocularity without the possibility of behavioral adaptation. This methodology has the virtue of placing the driver in a worst-case scenario but may be perceived as a little less environmentally realistic. However, it is quite possible, for various reasons, that a mirror may no longer be accessible to the driver (eg, a mirror that is broken, fogged, or useless due to the glare of the sun). It also is possible that, in some cases, the driver may not see an element in its wing mirror. Indeed, motor racing can involve an intense struggle between drivers. They are highly engaged and allocate a very important part of their attention to interactions with the vehicles in front of them, which they seek to overtake. This can lead to a phenomenon of inattentive blindness because their attention is strongly focused on another spot, event or object. The expression “inattentional blindness” was created by Arien, Mack and Irvin Rock [33] and popularized by the invisible gorilla test, conducted by Simons and Chabris [34]. In the case of inattention blindness, drivers can miss, for example, the overtaking maneuver of another car coming from behind them.

To improve the experience, it would be interesting to use a more ecological methodology to test each driver in a single condition with a single hazardous event. Indeed, in our study, we tested each driver in the three visual conditions and with 12 or 18 hazardous situations each time. However, these events are relatively rare while racing. Thus, the appearance of events would be less predictable. Nevertheless, the implementation of this method requires the involvement of a considerable number of drivers who are very hard-to-reach people. It could also be very interesting to perform a longitudinal study with drivers who have lost vision in one eye to see if they can compensate for their deficit and evaluate their adaptation time to recover performances as good as those of binocular drivers.

The present findings suggest that a sudden monocular reduction in vision with preservation of the peripheral field does not impact either race car driving performance or safety loss but that sudden monocularity has a great impact on driving performance and safety during car racing. Accordingly, we believe that further research should be conducted on monocular drivers who have had time to adapt to their disability to determine if adaptation is possible and sufficient to lead to an improved or even normal driving performance. In addition, future research should also look at how sudden and long-standing monocularity impacts driving for every day, non-professional drivers.

This is the dataset of experiment 1 for CC condition.

(XLSX)

Click here for additional data file.

This is the dataset of experiment 1 for CI condition.

(XLSX)

Click here for additional data file.

This is the dataset of experiment 1 for IC condition.

(XLSX)

Click here for additional data file.

This is the dataset of experiment 2 for CC condition.

(XLSX)

Click here for additional data file.

This is the dataset of experiment 2 for II condition.

(XLSX)

Click here for additional data file.

This is the dataset of experiment 1 for correlation analysis.

(XLSX)

Click here for additional data file.

This is dataset of experiment 1 for reaction time analysis.

(XLSX)

Click here for additional data file.

23 Aug 2019

PONE-D-19-19649

Implications of monocular vision for racing drivers

PLOS ONE

Dear Dr. Adrian,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

The manuscript has been reviewed by two experts in transportation safety and medical biostatistics. Based on the review results, some major issues need to be fully addressed in the revised manuscript. Particularly, some of the major issues are: (1) statistical analysis method issue as brought up in the comments; and (2) sample size issue of statistical analysis. Please completely address the aforementioned major issues along with other detailed comments from both reviewers in a thoroughly revised manuscript in order for the manuscript to be recommended for acceptance.

We would appreciate receiving your revised manuscript by Oct 07 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Zhixia Li, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following in the Competing Interests section:

"I have read the journal's policy and some authors of this manuscript have the following

competing interests:

José-Alain Sahel:

Funding Sources: LabEx LIFESENSES (ANR-10-LABX-65), ERC Synergy

"HELMHOLTZ" (ERC Grant Agreement #610110), Banque publique d'Investissement

(Sightagain BPI-2014-PRSP-15), University-Hospital Institute “FOReSIGHT ((ANR-18-

IAHU-01)”, Foundation Fighting Blindness (C-CL-0912-0600-INSERM01; C-GE-0912-

0601-INSERM02).

Consultant: Pixium Vision; GenSight Biologics; SparingVision.

Personal Financial Interests: GenSight Biologics, Prophesee, Chronolife, Pixium

Vision, Tilak Healthcare, Sparing Vision.

Bahram Bodaghi is consultant for the Medical Commission of the FIA.

Neil R. Miller is study director for the QRK207 Clinical Treatment Trial for acute NAION

(Quark Pharmaceuticals) and consultant to the Regenera pharmaceutical company.

Gérard Saillant is President of the Medical Commission of the FIA.

The other authors do not have any conflicts of interest to disclose".

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests).  If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

3. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

Additional Editor Comments (if provided):

The manuscript has been reviewed by two experts in transportation safety and medical biostatistics. Based on the review results, some major issues need to be fully addressed in the revised manuscript. Particularly, some of the major issues are: (1) statistical analysis method issue as brought up in the comments; and (2) sample size issue of statistical analysis. Please completely address the aforementioned major issues along with other detailed comments from both reviewers in a thoroughly revised manuscript in order for the manuscript to be recommended for acceptance.

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Interesting paper. I only have minor comments as the follows:

* All participants are non visually impaired. It is possible that the test drivers perform worse under monocular vision case because they are not used to rely on only one eye. The drivers that adapt to the monocular vision may perform much better. The authors discussed this issue in the conclusion section. I was wondering if the authors can give some recommendations on how we can improve the experiments and analyses in the future studies.

* In line 176, why was each driver tested only once in each condition? Multiple tests in each test might generate richer samples for eliminating test errors.

* In line 188, if every situation is a separate sample, how can we perform a statistically meaningful analysis?

* In line 231, how can a human driver keep a fixed headway from the preceding vehicle?

Reviewer #2: The authors designed two experiments to investigate whether monocular vision or a monocular generalized reduction in vision (MRV) impacts driving performance during racing. I do have some comments on the statistical analysis and wish the authors could address them:

1. In the first experiment, there were 18 participants. But how many observations were obtained from each participant? Why the observations from the same participant could be viewed as independent? A mixed logistic model and the ANOVA for repeated measurements may be more appropriate.

2. Also in the first experiment, how many participants were in CC, CI, and IC, respectively? On average, there were six subjects in each scenario. The authors need to justify why the sample size is sufficient to detect meaningful difference. A similar justification may also be needed for the second experiment

3. I wonder whether the authors could explain why the reaction time was not measured and compared in the second experiment?

4. In the discussion, the authors attempted to explain how monocular vision or MRV impacts driving performance during racing through reaction time. Is it possible that the authors could do some analysis to show the connection between reaction time and collision?

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

27 Sep 2019

Response to reviewers

September 2019

In red are the response of the authors

As asked by the Editor, we include our updated Competing Interests statement in the response to reviewers’ letter.

I have read the journal's policy and some authors of this manuscript have the following

competing interests:

José-Alain Sahel:

Funding Sources: LabEx LIFESENSES (ANR-10-LABX-65), ERC Synergy

"HELMHOLTZ" (ERC Grant Agreement #610110), Banque publique d'Investissement

(Sightagain BPI-2014-PRSP-15), University-Hospital Institute “FOReSIGHT ((ANR-18-

IAHU-01)”, Foundation Fighting Blindness (C-CL-0912-0600-INSERM01; C-GE-0912-

0601-INSERM02).

Consultant: Pixium Vision; GenSight Biologics; SparingVision.

Personal Financial Interests: GenSight Biologics, Prophesee, Chronolife, Pixium

Vision, Tilak Healthcare, Sparing Vision.

Bahram Bodaghi is consultant for the Medical Commission of the FIA.

Neil R. Miller is study director for the QRK207 Clinical Treatment Trial for acute NAION

(Quark Pharmaceuticals) and consultant to the Regenera pharmaceutical company.

Gérard Saillant is President of the Medical Commission of the FIA.

The other authors do not have any conflicts of interest to disclose

This does not alter our adherence to PLOS ONE policies on sharing data and materials.

Supporting Information has been added at the end of the manuscript.

S1 Dataset. This is the dataset of experiment 1 for CC condition.

S2 Dataset. This is the dataset of experiment 1 for CI condition.

S3 Dataset. This is the dataset of experiment 1 for IC condition.

S4 Dataset. This is the dataset of experiment 2 for CC condition.

S5 Dataset. This is the dataset of experiment 2 for II condition.

S6 Dataset. This is the dataset of experiment 1 for correlation analysis.

S7 Dataset. This is dataset of experiment 1 for reaction time analysis.

Responses to Reviewer #1:

* All participants are non visually impaired. It is possible that the test drivers perform worse under monocular vision case because they are not used to relying on only one eye. The drivers that adapt to the monocular vision may perform much better. The authors discussed this issue in the conclusion section. I was wondering if the authors can give some recommendations on how we can improve the experiments and analyses in the future studies.

Answer:

We appreciate the reviewer’s question and have added in the discussion of the manuscript:

- Page 25-26 line 519-528:“To improve the experience, it would be interesting to use a more ecological methodology to test each driver in a single condition with a single hazardous event. Indeed, in our study we tested each driver in the three visual conditions and with 12 or 18 hazardous situations each time. However, these events are relatively rare while racing. Thus, the appearance of events would be less predictable. Nevertheless, the implementation of this method requires the involvement of a considerable number of drivers who are very hard-to-reach people. It could also be very interesting to perform a longitudinal study with drivers who have lost vision in one eye to see if they can compensate for their deficit and evaluate their adaptation time to recover performances as good as those of binocular drivers.”

We also made a correction page 18 line 375-376. “The session consisted of eight trials for the CC and four for the II randomized situations located at different turns of the circuit”.

* In line 176, why was each driver tested only once in each condition? Multiple tests in each test might generate richer samples for eliminating test errors.

We appreciate the reviewer’s question. We were only able to test each driver once in each visual condition because the vast majority of them are professional drivers who have a very busy schedule and therefore had only limited time to do the driving simulator test. It was therefore not possible to have more time with them or to review them after the test. Furthermore, testing multiple times in each condition might induce a learning effect, like strategies to avoid situations, that might bias the results.

* In line 188, if every situation is a separate sample, how can we perform a statistically meaningful analysis?

We appreciate the reviewer’s question. The principle is to use each tested situation independently which takes a binary form of 0 or 1 depending on whether or not there has been an accident. To evaluate these data, we use statistics that model binary variables and are analytical models commonly used in epidemiology. For example. we had initially used a logistic regression but we eventually performed a GLMM at the request of reviewer 2.

* In line 231, how can a human driver keep a fixed headway from the preceding vehicle?

We appreciate the reviewer’s question. In our simulation experiments, the driver does not maintain the time headway. It is the front car that is driven by the simulator software that provides the 4s interval. No matter what the driver does, the car in front of him adapts by accelerating or braking to obtain a constant time headway of 4 seconds.

Responses to Reviewer #2:

Reviewer #2: The authors designed two experiments to investigate whether monocular vision or a monocular generalized reduction in vision (MRV) impacts driving performance during racing. I do have some comments on the statistical analysis and wish the authors could address them:

1. In the first experiment, there were 18 participants. But how many observations were obtained from each participant? Why the observations from the same participant could be viewed as independent? A mixed logistic model and the ANOVA for repeated measurements may be more appropriate.

We appreciate the reviewer’s question and comment. In experiment 1, we tested the subject for each visual condition 18 times, with 6 tests per type of situation (CC, CI and IC). In experiment 2, we tested the subject 12 times for each visual condition with 8 tests for the CC condition and 4 tests for condition II. A number of situations failed for various reasons, such as the loss of control of the car at the time of the event or the subject's poor positioning in the curve, which reduced the number of final situations.

We initially considered the observations as independent for several reasons:

- Drivers have a very stereotypical behaviour unlike conventional driving. They are all trying to go as fast as possible by following the ideal trajectory on the circuit.

- Each test situation has been implemented in a different turn and no two turns have the same characteristics (speed, angle, dynamics, grip...).

- A driver is not homogeneous in his driving performance. Its performance is not the same according to the turn or sectors of the track.

Having said this, we have followed the recommendations of the reviewer by adding a GLMM that corresponds to a logistic regression in which we control the effects related to the participants' test-retest. We used a Generalised Linear Mixed Model (GLMM) with a logit link function to account for the possible unobserved heterogeneity caused by repeated measures from the same individuals. The results of the GLMM have been inserted in the manuscript.

On the other hand, for reaction times we had initially used an analysis of variance with repeated measurements.

Concerning the modification of the statistical analysis we added:

- Page 2 line 32. “Generalized Linear Mixed Models (GLMMs)”

- Page 2 line 36=5-37=6. “2.1 (95% CI 1.11 - 4.11, p=.024) to 6.5 (95% CI 3.91 – 11.13; p=.0001)”

- Page 9 line 191-204. “A Generalized Linear Mixed Model (GLMM) with a logit link function was applied, using an R package named lme4 [20], in R [21] to study the likelihood (odds ratio) of collision risk given the visual condition of the racing drivers. The GLMM was used to account for the possible unobserved heterogeneity caused by repeated measures from the same individuals. The binary response variable was the involvement in an accident whereby 1= accident and 0= no accident. We set the normal vision condition as the reference condition. We assessed overdispersion for every model (the sum of the squared Pearson residuals should be χ2 distributed)[22].”

- Page 13 line 272-281.” As a first step, we conducted a series of GLMMs to determine if multiple visual conditions (monocularity or MRV) were predictive of an accident and the odds of a collision happening given the visual condition. We conducted a GLMM for each potential hazard situation; i.e, CC, CI and IC (Table 2). The analysis indicated the absence of significant overdispersion for CC situations model (�2= 228.567; ratio= 1.002; rdf= 228; p= .477); for IC situations model (�2= 204.360; ratio= 0.973; rdf= 210; p= .597); and for CI situations model (�2= 203.934; ratio= 0.944; rdf= 216; p= .712).

- Page 13-14 line 283-286.”

Table 2. Parameter estimates from the three GLMMs for the 3 driving situations related to the occurrence of collisions for each visual condition.

Parameter Estimate SE Z value p OR 95% CI low 95% CI high

Model for CC Situations

Intercept -0,701 0,247 -2,839 0,001

Baseline vision 0,00 0,00

MRV 0,484 0,335 1,447 0,148 1,623 0,845 3,153

Monocular 0,751 0,333 2,258 0,024 2,119 1,112 4,111

Model for CI Situations

Intercept -0,710 0,282 -2,514 0,012

Baseline vision 0,00 0,00

MRV 0,122 0,361 0,337 0,736 1,129 0,556 2,310

Monocular 0,286 0,352 0,812 0,417 1,331 0,668 2,679

Model for IC Situations

Intercept -1,849 0,360 -5,136 0,0001

Baseline vision 0,00 0,00

MRV 1,001 0,432 2,314 0,021 2,720 1,188 6,587

Monocular 1,129 0,428 2,637 0,008 3,094 1,368 7,456

- Page 14-15 Line 287-297.” For the CC situations, monocularity was the only significant condition in the model (p=.024). Model parameters showed that racing drivers in the monocular condition have 2.1 (95% CI 1.112 - 4.111) greater odds of having a collision than racing drivers in the Baseline binocular condition. For CI situations, the GLMM showed that neither monocularity nor MRV was significant. Finally, for IC situations, both monocularity and MRV were significant (p=.008 and p=.021, respectively). Model parameters showed that recently monocular racing drivers had 3.094 (95% CI 1.368 – 7.456) greater odds of having a collision compared with the baseline condition, and racing drivers with MRV had 2.72 (95% CI 1.188, 6.587) times greater odds of having a collision compared with the baseline visual condition.”

- Page 20 line 392-399. “We conducted a series of GLMMs to determine if monocularity or MRV were predictive of an accident and the odds of a collision occurring, given the specific visual condition. A GLMM was conducted for both the CC and II hazard situations (Table 4). The analysis indicate the absence of significant overdispersion for CC situations model (�2= 590.933; ratio= 0.903; rdf= 654; p= .963) and for II situations model (�2= 242.486; ratio= 0.973; rdf= 320; p= .999).”

- Page 20 line 400-404. “Monocularity in the CC situations model was the only significant condition in the logistical regression models (p<.00015). Model parameters showed that racing drivers in the monocular condition have 6.4934.77 (95% CI 3.004911 -– 11,1337.571) greater odds of having a collision compared with racing drivers in the baseline condition.”

- Page 20-21 line 406-411.

Table 4. Parameter estimates from the two GLMMs for the 2 driving situations related to the occurrence of collisions for each visual condition

Parameter Estimate SE Z value p OR 95% CI low 95% CI high

Model for CC Situations

Intercept -2,041 0,274 -7,448 0,0001

Baseline vision 0,00 0,00

MRV 0,376 0,273 1,377 0,168 1,457 0,851 2,523

Monocular 1,87 0,263 7,102 0,0001 6,493 3,911 11,133

Model for II Situations

Intercept -1,554 0,348 -4,463 0,0001

Baseline vision 0,00 0,00

MRV 0,083 0,351 0,239 0,811 1,087 0,538 2,211

Monocular 0,112 0,360 0,312 0,755 1,119 0,542 2,321

- Page 21 line 416-419. Specifically, the results of the two experiments show that, compared with the baseline condition, drivers under the recent monocular condition are 2.1 to 6.54.7 times more likely to collide with target vehicles, depending on the driving situation.

- Page 28 line 604-608. “20. Bates D, Mächler M, Bolker BM, Walker SC. Fitting linear mixed-effects models using lme4. J Stat Softw. American Statistical Association; 2015;67. doi:10.18637/jss.v067.i01

- 21. R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna.: R Foundation for Statistical Computing; 2015.

- 22. Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH, et al. Generalized linear mixed models: a practical guide for ecology and evolution. Trends in Ecology and Evolution. 2009. pp. 127–135. doi:10.1016/j.tree.2008.10.008”

We removed:

Page 9 line 197-202. “We theorized that all situations are different from each other because the angle of the curve, the speed of approach, and the position of the hazardous event are never the same. Furthermore, a single driver could have experienced the same situation twice (test re-test) but in a completely different visual context with a major impact on driving performance. Thus, we considered every situation as a separate sample for the logistic regression.”

2. Also in the first experiment, how many participants were in CC, CI, and IC, respectively? On average, there were six subjects in each scenario. The authors need to justify why the sample size is sufficient to detect meaningful difference. A similar justification may also be needed for the second experiment

We understand the reviewer’s concern:

Each of the 18 participants passed the three conditions CC, CI and IC.

We have added the following in the manuscript:

- Page 12-13 line 266-271. “We evaluated 18 participants who passed all three visual conditions. In each condition, drivers were confronted with three types of dangerous situations. Six tests were proposed for each dangerous situation. Eliminating missing data related to the failed or incorrectly performed tests, we collected 232 observations for the CC situations, 220 observations for the CI situations and 214 observations for the IC situations.

- Page 20 line 387-391. “We evaluated 31 participants who passed all three visual conditions. In each condition the drivers were confronted with two types of dangerous situations. Eight tests were proposed for the CC situations and four tests for the II situations. If we eliminate the failed or incorrectly performed tests, we collected 658 observations for the CC condition and 325 observations for situation II.”

We added in the statistical analysis paragraph:

- Page 8-9 line182-190. “The minimum number of participants required was determined by an a priori power analysis. Our objective was to compare the occurrence of accidents by visual condition and report the estimated ORs using a logistic regression (the analysis was modified during the revision process). To determine the number of observations required per group (n), we assumed that the accident rate expected in the control condition (Baseline) would be 5% and 20% in the monocular condition with a power of at least 90% and an α of .05. We obtained a number of 100 observations per group, or a minimum number of 17 participants taking into account the number of tests per driving situation.”

3. I wonder whether the authors could explain why the reaction time was not measured and compared in the second experiment?

We appreciate the reviewer’s question. In experiment 1, drivers have to detect and react to a sudden intrusion of an opponent’s racing car into the participant's trajectory in a turn. They have no choice but to provide an action (breaking or turning) to not crash. We can thus determine a reaction time between the intrusion into the visual field and the beginning of an avoidance action. In the second experiment, the opposing cars come from behind and interfere with the drivers in their trajectory around an up coming bend. The main solution used by the drivers to avoid accidents in this experiment was to take a slightly wider turn as soon as they see the opposing car. However, as the opposing car comes from behind the driver, there is no way to determine exactly when the driver sees the opposing car and, thus, no way to determine the delay from the time the driver sees the opposing car to the time he reacts to it.

4. In the discussion, the authors attempted to explain how monocular vision or MRV impacts driving performance during racing through reaction time. Is it possible that the authors could do some analysis to show the connection between reaction time and collision?

We appreciate the reviewer’s question and have performed correlation analyses between reaction times and accident rates for each visual condition, regardless of the type of situation. As the variables were normally distributed, we used a Pearson correlation.

Regarding the baseline visual condition, the results of the Pearson correlation indicated that there was no significant association between reaction time and accident rates, (r = .28, p = .26).

Regarding the MRV visual condition, the results of the Pearson correlation indicated that there was a significant positive association between reaction time and accident rates, (r=.64, p=.004).

Regarding the monocular visual condition, the results of the Pearson correlation indicated that there was a significant positive association between reaction time and accident rates, (r=.48, p=.044).

The results of the correlational analysis show that the increase in reaction time has an impact on involvement in accidents only under conditions where vision is impaired. This can be explained by the fact that in baseline conditions, the accident relies more on the intrinsic qualities of each driver to be able to avoid the accident. On the other hand, when vision is impaired, there is a substantial increase in reaction time, which leaves only a small chance for the driver to avoid the accident, given the dynamics of the cars.

We added:

- Page 10 line 206-207. “We analyzed the relationships between reaction times and accident rate for each visual condition, regardless of the type of situation, using Pearson’s correlation."

- Page 16 line 325-334. “Correlations between reaction times and accident rate.

As the variables were normally distributed, we used a Pearson correlation.

Regarding the baseline visual condition, the results of the Pearson correlation indicated that there was no significant association between reaction times and accident rates, (r = .28, p = .26). Regarding the MRV visual condition, the results of the Pearson correlation indicated that there was a significant positive association between reaction times and accident rates, (r=.66, p=.003). Finally, regarding the monocular visual condition, the results of the Pearson correlation indicated that there was a significant positive association between reaction times and accident rates, (r=.48, p=.048).”

- Page 22 line 429-431. “. In addition, correlation analysis allows us to observe that the longer the reaction time increases in this visual condition, the greater the risk of an accident.”

- Page 22- line 442-445. “. Nevertheless, correlation analysis allows us to observe that, for the monocular visual condition, the longer the reaction time increases, the greater the risk of an accident, whereas this is not the case in baseline visual condition.”

Submitted filename: Response to reviewers.docx

Click here for additional data file.

21 Oct 2019

PONE-D-19-19649R1

Implications of monocular vision for racing drivers

PLOS ONE

Dear Dr. Adrian,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please address the Reviewer 2's comments in the minor revision.

We would appreciate receiving your revised manuscript by Dec 05 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Zhixia Li

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

Please address the Reviewer 2's comments in the minor revision.

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: All my comments have been addressed.

Thanks.

Reviewer #2: (1) It may be worth pointing out that the ANOVA is for the repeated measurements.

(2) The authors found that there is no strong correlation between the reaction time and the accident rate in the baseline condition, and argued that "this can be explained by the fact that in baseline conditions, the accident relies more

on the intrinsic qualities of each driver to be able to avoid the accident." I wonder if the authors can provide any evidence to support their claim.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Hao Liu

Reviewer #2: No

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

7 Nov 2019

Responses to Reviewer #2:

Reviewer #2: (1) It may be worth pointing out that the ANOVA is for the repeated measurements.

Answer:

We appreciate the reviewer’s question. we have modified:

Page 9 line 196-197 “We also analyzed driver reaction time in crash avoidance using two-way analysis of variances (ANOVAs) with repeated measures.” by “We also analyzed driver reaction time in crash avoidance using two-way repeated measures ANOVA.

Page 13 line 282-283 “The two-way ANOVA with repeated measures showed significant main effects of the visual condition” by “The two-way repeated measures ANOVA showed significant main effects of the visual condition”.

(2) The authors found that there is no strong correlation between the reaction time and the accident rate in the baseline condition, and argued that "this can be explained by the fact that in baseline conditions, the accident relies more on the intrinsic qualities of each driver to be able to avoid the accident." I wonder if the authors can provide any evidence to support their claim.

We appreciate the reviewer’s question. There is no specific evidence we are currently aware of and this claim is only made based on our analysis and observations. We should have used instead :“this may be explained by the fact that in baseline conditions, the accident relies more on the intrinsic qualities of each driver to be able to avoid the accident". If the reviewer is aware of any evidence we would be open to including it here.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

25 Nov 2019

Implications of monocular vision for racing drivers

PONE-D-19-19649R2

Dear Dr. Adrian,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Zhixia Li

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: The authors have addressed all comments I have in the previous reports. I do not have any further comment.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

4 Dec 2019

PONE-D-19-19649R2

Implications of monocular vision for racing drivers

Dear Dr. Adrian:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Zhixia Li

Academic Editor

PLOS ONE