A scoping review investigating motorcyclists’ visual gaze using eye-tracking methodology

A scoping review investigating motorcyclists’ visual gaze using eye-tracking methodology.

Introduction. Motorcyclists are one of the most vulnerable road user groups worldwide world (World Health Organization, 2015). The UK Department for Transport (2023) reports the most common crash contributory factor is ‘the driver or motorcyclist failed to look properly’. To better understand motorcyclists’ visual gaze, it is necessary to review the prior research which uses eye-tracking methodologies.

Methods. A scoping review was deemed appropriate due to limited prior research; following Arskey and O’Malley (2005) five stage guidelines. 18 articles were selected for review according to search and inclusion criteria.

Results. Key outcomes include: motorcyclists gaze at the road surface more than car drivers, experienced motorcyclists have visual gaze strategies which target potential hazard zones, and it is possible to train novices’ gaze strategies.

Conclusions. Despite the heterogeneity of research designs and eye-tracking technology; findings demonstrate commonalities in motorcyclists’ visual gaze, furthermore, gaze training can enhance hazard perception with long-term effect.

Keywords

Eye tracking, Motorcycle safety, Collision prevention, Visual attention

1. Introduction

Motorcyclists are one of the most vulnerable road user groups in the U.K. with an average of 6 deaths and 104 serious injuries per week (Department for Transport, 2023). Motorcycles account for 1% of the miles driven on the UK roads but comprise 18% of those Killed or Seriously Injured (KSI) (RoSPA, 2017). Furthermore, motorcycling is a vulnerable form of transport throughout the world (World Health Organization, 2015).

Motorcycle crashes tend to result in motorcyclist KSI due to minimal safety protection, unlike that of cars which has a metal cage, seatbelts, and airbags. Instead, the onus of protection is based on the motorcyclist wearing protective clothing and riding in a manner which anticipates and negotiates hazards appropriately (Araujo et al., 2017). However, wearing protective clothing is often a choice. For example, some countries do not have compulsory laws for wearing of protective clothing and several countries have loopholes that motorcyclists exploit to avoid wearing a helmet, despite knowing that helmets can save lives (Araujo et al., 2017).

Notwithstanding legalities of protective equipment, motorcyclists should aim to ride in a manner to prevent a crash occurring. For example, the Motorcycle Accidents in Depth Study (MAIDS) found that 37.4% of crashes were due to rider error (ACEM, 2004). Moreover, the handling and control characteristics are different to that of other motorised forms of transport because motorcycles can lean in corners and stability can be compromised due to road and weather conditions (Cho & Wu, 2004). Therefore, motorcycling requires multiple physical and cognitive demands including visual attention.

Visual attention can be explained as the ability to select and attend to objects in a field of view (Ungerleider & Kastner, 2000). However, a visual scene can contain multiple objects which may overload visual and cognitive processing (Lamme, 2003). Therefore, it is important to know which objects to select and attend to. For example, amongst several attention strategies, attention may be directed to items of salience via a top-down process which is primed to look for specific objects or a bottom-up process which is reactionary due to unexpected yet salient objects (Connor et al., 2004).

Motorcyclists use visual attention for a variety of reasons, such as for direction control, hazard perception, or to correctly identify road signage. Consequently, optimal visual attention may be one of the most important factors in prevention of crashes. Yet, U.K. Government statistics which outline contributory factors to crashes state that 44% of crashes regardless of vehicle type are due to ‘failed to look properly’ and 22% ‘failed to judge the other person’s path or speed’ (Department for Transport, 2023). Indicating there may be deficits in visual attention of road users.

Therefore, to investigate reduction of motorcycle crashes, motorcyclists’ visual attention should be considered. Eye-tracking equipment has been used to assess visual gaze for a variety of situations and populations; from war veterans’ attention towards traumatic stimuli to young women’s attention towards fashion items (Ju & Johnson, 2010; Kimble et al., 2010). Thus, using eye-tracking to investigate motorcyclists’ visual gaze is an appropriate methodology. A brief literature search indicated a limited yet varied approach to motorcyclists’ eye-tracking research. Therefore, a scoping review was considered appropriate to identify current knowledge and research gaps (Munn et al., 2018). The research question was, ‘What is the existing research regarding motorcyclists’ visual gaze using eye-tracking equipment?

The research aims and objectives included:

• Investigate motorcyclists visual gaze: where do motorcyclists look?
• Record different technologies used to track motorcyclists’ visual gaze.
• To investigate the methodology and outcomes of existing research regarding motorcyclists’ visual gaze.
• Collate, compare, and contrast research outcomes.
• To explore the potential application of research outcomes.
• Identify limitations and gaps in research.
• Suggest directions for future research.

2. Methodology

Arskey and O’Malley (2005) provide a comprehensive description for scoping reviews, recommending that scoping studies aim to investigate key concepts quickly and comprehensively in a research area. Additionally, they suggest a scoping review does not need to critically assess the research literature using a specific analysis tool (Arksey & O’Malley, 2005: Munn et al., 2018). A literature search regarding motorcyclists’ visual gaze demonstrated no existing scoping or systematic review. Thus, it was decided that conducting a scoping review to explore the state of the prior research was appropriate due to the broader scoping methodology.

Arksey and O’Malley (2005) recommend a five-stage framework to ensure rigour and replicability for scoping reviews. The stages are: (1) identify the research question, (2) identify relevant research, (3) selection of research, (4) chart the research data, (5) collate, summarise, and report the results.

In addition to following the Arksey and O’Malley (2005) framework, the scoping review followed PRISMA ScR guidelines and Joanna Briggs Institute (JBI) guidance (Tricco, et al., 2018; Peters et al., 2020).

2.1 Eligibility criteria

The eligibility criteria for research to include in the scoping review are listed below and shown in detail in table 1.

1) English language only.

2) Peer reviewed research published in journals and conference articles or published Government research.

3) Use of eye-tracking equipment. Which may be specific to its time, for example early eye-tracking equipment was crude compared with current technology.

4) Participants who are solely motorcyclists or instructed to behave as a motorcyclist: or motorcyclists who hold both a motorcycle licence and a car driving licence yet are not instructed to, ‘behave as you would when driving a car’.

There were no restrictions made regarding the publication date because eye-tracking is a recent methodology used for transport research. Additionally, methodological factors, such as sample size or research setting (laboratory versus field-based data collection) were not restricted. However, restrictions were placed regarding the publication type of the research, consequently editorials and letters were excluded.

 

Table 1. Eligibility criteria for the research included in the scoping review.

Criterion	Inclusion	Exclusion
Date Range.	All dates.
Language.	English.
Research Status.	Published in a peer reviewed journal or published Government research.	Editorials, Commentaries, Letters.
Methodology – Equipment.	Eye-tracking used with a motorcycle/motorcycle simulator.	Eye-tracking for other road users: car, lorry, bus, bicycle.
Methodology – Participants.	Motorcyclists or participants instructed to behave as motorcyclists (such as on a simulator).	Solely non-motorcyclists: car, lorry, bus, cyclist research. Motorcyclists who are instructed to behave as a car driver or other motorist type.
Methodology – Scenario.	Either motorcycles in the field or laboratory based with either motorcycle simulators and/or projected road-based scenarios.
Methodology – desired outcome	Eye tracking as a measurement for visual attention for road traffic behaviours and motorcycle control	Solely using eye-tracking as a method for assessing mental workload.

2.2 Information sources

The search was carried out between 01.03.2023 – 01.10.23. The search strategy was two-fold. The initial search was for published peer reviewed research from journal articles, published conference proceedings, and published Government research. Secondly, a cited references search was used when reading the literature and subsequently the use of publication lists on author websites/university profile pages of authors who regularly publish in this research area.

Sources of information included electronic databases: CINAHL, PsychArticles, PsychInfo, ScienceDirect, Scopus, and Web of Science. The databases contained journals from varied disciplines including Psychology, Accident Prevention, and to a lesser extent Engineering journals; due to subscription restrictions at the organisation where the researcher is based. Google Scholar was used for a wider search to include published Government research.

Boolean operators were used for the database and journal search. Search terms mostly used truncated words to explore wider possibilities of terms and spellings. For example, motorcycles are known by many names including Motorbike and Powered Two-wheeler. Consequently, search terms for motorcycle included; motorcycle (motorc*), motorbike (motorb*), bike*, powered two-wheel*.

It was decided that using the NOT search parameter could produce unwanted exclusions. For example, due to the potential dual usage of the word for bike meaning both motorcycle or bicycle. Consequently, research which included bicycles was excluded as part of the title and abstract screening.

Search terms for visual attention included: vision (vis*), visual perception (visual percep*), visual attention (visual atten*). Search terms for and eye-tracking included: eye-tracking (eye track*), Eye-movement (eye move*), and eye-fixation (eye fix*). An example of a search in PsychInfo database is: Motorc* AND Vis* AND Eye track*.

Arksey and O’Malley (2005) suggest that researchers may not have the same comprehensive search skills as a specialist librarian. Consequently, once initial search was completed by the researcher, a specialist university librarian was consulted to assess the search.

2.3 Selection of sources of evidence

Alongside using the eligibility and exclusion criteria, a Population, Intervention, Comparison and Outcome (PICO) was used to frame the research question and was incorporated for assessing the literature: Participants = Motorcyclists, Intervention = Eye-tracking, Comparison = Lab-Based and Field-Based research, Outcome = Visual Gaze and Visual Fixations.

The exclusion criteria were specific to methodology used. For example, eye-tracking data can be used to judge mental workload, yet the research may not report where the motorcyclist is directing their visual attention. Additionally, research was excluded which used participants who held both a motorcycle and car licence but instructed them to think or behave as if driving a car.

Publication selection started by removing duplicates from database search. Then the Title and Abstract were screened for eligibility criteria. Full-Text screening was used when the Title and Abstract did not state clearly whether eye-tracking equipment was used, or which methodology was employed such as instructions to participants, figure 1 includes the screening procedure.

Articles were added to Mendeley bibliographic software for storage and retrieval.

 Figure 1. Prisma flowchart of the scoping review research selection process

 

2.4 Quality Appraisal Process

The sources of literature were assessed independently by the author. However, once literature was screened and the final research selection was collated, the author sought secondary validation from a wider team comprised of professionals from psychology and health disciplines.

Article data retrieved included: Author(s), Date, Journal, Right or Left-hand drive country, Participants, Methodology including equipment and experimental processes, Results and Outcomes, and Researcher Conclusions.

It is important to note worldwide nuances in language and driving conditions. For example, some countries use the word ‘Pavement’ to mean ‘road surface’, whereas other countries use the word ‘Pavement’ for the strip of paved land beside a road that pedestrians use (Collins, 2023). Furthermore, 35% of the world’s countries drive on the left-hand side, therefore the researcher made note of this distinction to identify whether there were differences or comparisons in motorcyclists’ visual attention and field of view (Worldstandards, 2023).

To aid analysis of the selected eye-tracking research articles, the researcher referred to the JBI Critical Appraisal Checklist for Quasi-Experimental Studies (non-randomized experimental studies) (Tufanaru, et al., 2020).

2.5 Data Charting

Data from the selected research articles was charted according to the research question, eligibility criteria, aims and objectives for this scoping review, see Table 2.

2.5 Data Charting

 Table 2. Data Chart of selected research, key information, and findings.

No	Research Citation	Participants	Methodology	Results/Outcome	Conclusion/Compare/Contrast
1	Mortimer, R.G. and Jorgeson, C.M., 1975. Comparison of eye fixations of operators of motorcycles and automobiles. SAE Transactions, pp.930-935.	Two males aged 21 and 28. Right-hand drive. USA	Hemet mounted camera – solely measuring eye fixations were wired to recording equipment which was mounted to the motorcycle tank. A passenger carried a camera to record the road route. Participants ride a motorcycle then drive a car through pre-arranged route. This allowed comparison of eye fixations. Each vehicle speed approx. 45mph.	1. Mean duration for eye fixation was longer when riding a motorcycle. 2. Motorcycling eye fixations are more to the right side of the road (the verge) when no oncoming traffic. Right side fixation was more frequent even through left and right-hand bends. 3. Motorcycling eye fixations were 14.1% nearer (30-76m) to the motorcycle than when driving.* 4. Moderately more fixation time at oncoming vehicles for motorcycle than car.	Results suggest motorcyclists look more at road surface than car drivers. Especially at road surface danger points such as the verge where more debris/road delineation may be present. This could have an impact on how riders perceive other objects outside of their field of view, for example road signage could be missed. *Note – at 45mph, 30m is closer than safe stopping distance. Safe stopping distance at 40mph is 36m.
2	Nagayama, Y., Morita, T., Miura, T., Watanabem, J. and Murakami, N., 1979. Motorcyclists’ visual scanning pattern in comparison with automobile drivers’. SAE Transactions, pp.934-945.	Experiment 1 Three males aged 24 – 36 Left-hand drive. Japan Experiment 2 Three males aged 26 – 36	Helmet mounted NAC Eye-Mark Recorder Model 4, with fibre optic cables to a fuji colour film. The motorcyclist carried a video recorder in a rucksack to record the road and when in the car an assistant held the video recorder. 2k section of arterial road in Osaka – at 50kph. Note – this is using ‘eye marks’ an early form of tracking eye movement and fixations. Extension on experiment 1 with additional I.V.s 3x types of vehicles – 50cc bike, 400cc bike, car. 3x speed – Low 30kmh, Med 45kmh, High 60kmh 1km stretch of road	1. The proportion of road surface in the visual field of a rider is much larger than that of car driver. 2. Eye fixations on the road surface are more frequent in motorcyclists. 3. There are more vertical eye movements in motorcyclists, looking from near to far road surface. 4. 30% of rider fixations are road surface. Whereas driver fixations rarely look at road surface instead drivers look further ahead and more at objects such as road signs. 1. Regardless of engine size/speed motorcyclists look at the near ground more than driver 2. Vertical gaze shifts increase with speed on a motorcycle – especially that larger motorcycle 3. The effect of speed has an different affect for motorcyclist vs driver. At 30kph speed the motorcyclist gaze is high, then drops low at 45 kph, then raises at 60kph. At low speed the car driver gaze is high. 4. On the 400cc there was an increase in convergent fixations at speed compared with the 50cc and the car. 5. mean fixation duration is longer in car drivers than 400cc motorcycle. 6. 400cc – the higher the speed the shorter the fixation duration.	1, 2, 3. Comparison with Mortimer and Jorgeson (1975), motorcyclists fixate more on the near-ground area of the road. 4. Results suggest that drivers looking further ahead and higher allows them to have a wider field of vision which can help pick up more detail in their peripheral vision too. Experiment 2. With increased speed, motorcyclists eye movements increase – this may be a function of vertical eye movements from looking both at the foreground and then into the distance.
3	Di Stasi, L.L., Álvarez-Valbuena, V., Cañas, J.J., Maldonado, A., Catena, A., Antolí, A. and Candido, A., 2009. Risk behaviour and mental workload: Multimodal assessment techniques applied to motorbike riding simulation. Transportation research part F: traffic psychology and behaviour, 12(5), pp.361-370.	60 Undergraduates at Granada University Spain. Aged 18 – 31. 31 Females, 29 Males. None had riding experience or a driving licence. Eight participants excluded due to equipment failure. Right-hand drive. Spain.	A Honda Riding Training (HRT) Simulator was used with a moped style arrangement. A road scenario was played on a 19 inch screen and an Eyelink II was used to record eye movements. Three road scenarios were used, which included hazards such as cars and unexpected objects. Each participant had two test rides and then the experimental ride – 5 min on an urban circuit, with dual carriageway, mixed traffic flow, intersections and eight potential COC situations. Data recorded included: 1, Eyelink II – eye saccades and fixations. 2, HRT – riding errors, COCs, and road code violations (e.g. speeding). 3, Ergonomics Evaluation Questionnaire (EEQ) 4, Mental Workload Test (MWT)	1. High-risk participants had fewer saccades of shorter duration and higher peak of saccadic velocity than low-risk participants. 2. Data indicated that subjective and eye motion measures of mental workload were better correlated in high-risk than low-risk participants.	1. It may be that fewer, faster saccades may indicate shallow processing of the task. For example, visual investigation of the riding environment involves assessing objects to identify if they are hazardous or estimating speed of traffic. This could require deep processing skills and longer visual fixations. Low risk individuals may be taking longer to assess objects in the visual field and thus obtaining more accurate information. 2. Peak of saccadic velocity was a better indicator of subjective mental workload than other eye movement measures, because it correlated with 61% of the subjective scales in the high-risk group.
4	Papakostopoulos, V., Nathanael, D. and Marmaras, N., 2010, August. An explorative study of visual scanning strategies of motorcyclists in urban environment. In Proceedings of the 28th Annual European Conference on Cognitive Ergonomics (pp. 157-160).	Three males aged 25 – 37. Each with 5 – 7 years motorcycling experience. Right-hand drive. Greece.	Helmet mounted eye tracking device. 30 minute set route; including four-lane motorway and urban roads. On completion of the ride participants were interviewed and required to comment on their ride while watching the video playback.	1. On the motorway and straight segments of road, the motorcyclists spent most of their time fixating far ahead. 2. On bends, the motorcyclists fixated on the tangent point. 3. On urban roads approaching intersections, motorcyclists used a different scanning strategy to that of the motorway and straight sections of road with few hazards. 4. At intersections motorcyclists search for potential hazards such as abnormalities in road surface or behind parked vehicles. This uses an increased speed visual scanning strategy, both horizontally and vertically. 5. Approaching intersections motorcyclists visual scanning strategy starts by tracking with successive fixations all segments of the traffic scene but with recurrent refocusing on the crossing road to verify that the apparent intentions of crossing vehicles stay unchanged. 6. Additionally the motorcyclist tracks using successive fixations the edge line of the vehicle ahead as well as its left or right wheel to monitor any imminent manoeuvring. 7. When other motorcycles appear in the foreground; fixation patterns change dramatically. Typically, the motorcyclists fixate to the helmet of a leading motorcyclist, seemingly paying less attention to adjacent automobiles. (This phenomenon possibly indicates an attempt of the motorcyclists to minimize the cognitive effort of scanning the road/traffic environment ahead by taking advantage of the paths opened by the motorcyclist ahead. This may be a strategy fostering cognitive economy.)	1. straight roads and motorways offer a wider field of view to fixate ahead and provide ability to utilise peripheral vision. 2. The tangent point is also known as the ‘limit point’. Taught widely in UK Post Compulsory training and to a lesser extent in compulsory training. It is unknown in this study the extent of training the riders have had. 3. Motorcyclists are actively searching for hazards in an environment that has a higher expectation of hazards. 4. Motorcyclists are trying to anticipate hazards expected at an intersection. Link road surface with Mortimer and Jorgeson (1975) and Nagayama et al. (1979). 5. Motorcyclists are anticipating other road users’ actions and movements. 6. As above – but more specific indicators of intended action such as wheels turning. A similar visual scan technique is taught in UK post-compulsory advanced motorcycle training. 7. Visual interest towards other motorcyclists could be a function of the social aspect of motorcycling and salience bias.
5	Hosking, S.G., Liu, C.C. and Bayly, M., 2010. The visual search patterns and hazard responses of experienced and inexperienced motorcycle riders. COC Analysis & Prevention, 42(1), pp.196-202.	37 Participants, with 3 levels of experience in motorcycling/driving. 1, EM – ED = Experienced with motorcycle and car driving. Age 28 – 69. 2, IM – ED = Inexperienced motorcycle and experienced car. Age 21 – 48. 3, IM – ID = Inexperienced motorcycle and car. Age 18 – 25. Data from six participants were excluded due to technical issues or simulator sickness. Left-hand drive. Australia.	Honda NSR 150 Riding simulator. Road scenarios projecting onto a large screen. FaceLab eye-tracking system. 2x riding scenarios: rural scenario was a 10-min ride on a one-lane dual carriageway in an 80–100km speed zone, and the urban/town scenario was a 5-min ride on a dual-lane dual carriageway in a 60km zone. Each participant had a 5-minute practise on a different scenario to that presented in the experiment. Hazard identification speed was recorded between when the hazard was apparent and when the participant pressed the button. Visual identification of hazard speed was recorded between when the hazard was initially present and eye fixation time. Experimental scenarios contained hazards that were agreed upon by a focus group of experienced motorcyclists (not those participating in this study). Hazards include, potholes, cars turning across the motorcyclist’s path, lorries straddling lanes. Participants pressed a button (fitted to the handlebar) when they detected a hazard. Immediately after a scenario had been completed, the hazards that were identified by the participant were replayed and paused at the time corresponding to each button press. Participants were then asked to name the hazard that they had responded to.	1. Irrespective of riding or driving experience, it took longer for riders to fixate on hazards in the rural road scenario. 2. Mean first fixation time data found that there was no significant main effect of riding and driving experience. However, there was a significant effect in response time for EM-ED and to a lesser extent ID-EM. (this is button press – not eye movement) 3. There was no significant interaction between riding and driving experience and scenario type for first fixation. 4. There were significant effects between scenario type and viewing plane. 5. The was a significant two-way interaction between hazard presence and viewing plane and scenario type and viewing plane. 6. There was a significant three-way interaction between viewing plane, hazard presence, and riding and driving experience. 7. the spread of fixations were significantly different for the IM – ED group in the no hazard condition compared with the IM – ID group in the hazard condition. 8. On the horizontal plane, the variance in visual fixation angles was significantly larger for the CBD/residential scenario than that of the rural scenario in both the hazard and no hazard conditions. The variance in visual fixation angles was also reduced when a hazard was present in the CBD/residential scenario but not in the rural scenario. 9. When a hazard is present, both EM-ED and IM-ED change their visual scanning strategies. Scanning was reduced.	1. Possibly because hazards are ‘expected less’ on rural roads. 2. Despite all groups see the hazard at the same time those with highest experience respond to the hazard quicker via button press. Thus, cognitive appraisal and prior experience of hazards is more accurate and quicker for experienced groups. Note: Viewing plane and fixations change in hazardous situations for EM-ED and to a lesser extent IM-ED; consequently, experience is a factor in visual scanning strategies. However, those with motorcycle experience had significantly different visual scanning strategies, whereby the field of view was narrower for potential hazard zones. Indicating that experience on a motorcycle is necessary to have best hazard perception skills.
6	Underwood, G., Humphrey, K. and Van Loon, E., 2011. Decisions about objects in real-world scenes are influenced by visual saliency before and during their inspection. Vision research, 51(18), pp.2031-2038.	77 Participants each with at least two years driving/riding experience. 50 car drivers: mean experience – 13.3 years. 27 motorcycle riders – 22 had a car licence too. Mean experience 14.3 years. Left-hand drive. U.K.	SR Eyelink II tracker 120 photographs of a road scene. (Taken at a junction on a dual carriageway where two lanes of traffic travel in the same direction. Pictures used had traffic coming from left or right). The primary condition was to judge if it was safe to enter the carriageway by pressing a button ‘safe’ or ‘unsafe’. Eye movement data were analysed to investigate attention towards the vehicles in the pictures – logging both number of fixations and duration of fixations. 10 pictures had motorcycles approaching at mid-distance and 10 had cars approaching at mid-distance. (65m from the junction) 50 pictures had relatively little traffic and were designed to elicit positive ‘‘safe-to-pull-out’’ responses. 50 pictures had near or heavy traffic that was expected to encourage cautious response. Only the pictures showing vehicles in the mid-distance were of interest to the analysis, as these were expected to result in the most difficult decision making, with some road users declaring the roadway safe-to-enter and others regarding these distances as unsafe.	1. Riders had longer fixation durations on the vehicles than drivers.	The motorcyclists’ longer fixation may indicate deeper cognitive processing such as to gauge hazard potential of the vehicle and assess speed of the vehicle.
7	Di Stasi, L.L., Contreras, D., Cándido, A., Cañas, J.J. and Catena, A., 2011. Behavioral and eye-movement measures to track improvements in driving skills of vulnerable road users: First-time motorcycle riders. Transportation research part F: traffic psychology and behaviour, 14(1), pp.26-35.	33 Participants. 17 inexperienced rider undergraduates aged 19 – 23 at Granada university. 16 expert riders aged 25 – 50 with more than 10 years of motorcycling experience (belong to local Vespa club) Right-hand drive. Spain.	Eye-link II head mounted eye-tracker. Honda Riding Trainer (HRT) Simulator. Road scenario played on 19” screen. Mental Workload Test used to estimate attention. All participants familiarised with the HRT with test rides. Experiment 1 Ride an urban circuit for 5 minutes. 8 potential COC situations. First circuit had no traffic, second circuit had traffic. Training. First time riders had a training session. Experienced riders did not have a training session and completed experiment 2 immediately after experiment 1. Experiment 2. Highway scenario – dual carriageway, straight road, no intersections. 8 potential COC situations. As above with no traffic followed by traffic.	1. Experiment 1 – expert riders gazed in relevant locations – such as rear-view mirrors and road margins and less on irrelevant spaces. No significant differences in experiment 2.	This indicates it is possible to train inexperienced riders to look at specific areas such as where to look for hazards. Comparison with Mortimer and Jorgeson (1975), Nagayama et al.  (1979), and Papakostopoulos et al. (2010).
8	Cheng, A.S., Ng, T.C. and Lee, H.C., 2011. A comparison of the hazard perception ability of COC-involved and COC-free motorcycle riders. COC Analysis & Prevention, 43(4), pp.1464-1471.	109 motorcyclists. 46 involved in an COC in the three years prior. 16 years mean experience. Mean annual mileage 11,944 km. Right-hand drive. China.	iViewX head mounted eye tracker system. Motorcycle simulator. Road scenarios projected onto a curved screen. 1 practise riding scenario and 1 experimental riding scenario. The hazard was boxes falling from a lorry. Other measures: 1, Chinese motorcycle rider driving violation questionnaire. 2, Symbol digit modalities test. 3, Colour trails test. 4, Digit vigilance test.	1. Those who have been involved in COCs are 0.3s later in glancing at a hazard.	Those who detect hazards this much later are 25 times more likely to be involved in a COC than those who detect it earlier. Thus, indicating that quick and accurate identification of hazards is required to respond appropriately to prevent COC occurring.
9	Muttart, J.W., Peck, L.R., Guderian, S., Bartlett, W., Ton, L.P., Kauderer, C., Fisher, D.L. and Manning, J.E., 2011. Glancing and stopping behaviour of motorcyclists and car drivers at intersections. Transportation research record, 2265(1), pp.81-88.	32 participants. Yet due to technical errors eye-tracking data was used from 23. Motorcyclists who also drive cars. Participants had motorcycle licence for a minimum of 5 years and covered at least 6000 miles per year. Four participants were riding instructors with 100k miles per year. Right-hand drive. USA.	650cc Kawasaki motorcycle Dodge Charger car Both fitted with equipment to measure riding/driving such as braking and acceleration. Mercury Marquis car used as a lead car, driven by testers. Toyota Prius driven passed participants to reinforce ecological validity. Inflatable pedestrians placed around the course – also one of the research team walked parts of the route for realism. Ride or drive set route for 25 minutes with each vehicle – with various intersections. (A housing estate which had not been completed) Mobile eye-tracking system – glasses fitted and calibrated to each participant. 5x threat conditions. 1, Left turn 2, Right turn 3, Left turn across opposing traffic 4, Stop 5, Not looking at pavement (road surface)	1. 19 participants had fewer glances toward the area of most significant hazard potential before turning when riding a motorcycle compared with driving. 2. Motorcyclists look at pavement (road surface) considerably more than when driving. 3. Motorcyclists made 30% more pavement glances at right turns and when approaching sand across the road surface. 4. Motorcyclists’ search area was nearly 10 degrees larger both vertically and horizontally. 5. Experienced motorcyclists maintained an average glance location 2 degrees to the left of straight ahead, whereas those with less than 5 years of riding maintained an average fixation location nearly 9 degrees to the right. 6. Motorcyclists were also less likely to make a glance to the right before arriving at the stop line	1. One possible cause of this relatively poor performance is that many of these motorcyclists reported that they were taught to follow your nose when they were making a left turn. The rider’s concern for motorcycle control may be at the expense of intersection safety. 2. The motorcyclists gaze at road surface, comparison with Mortimer and Jorgeson (1975), Nagayama et al.  (1979), Di Stasi et al. (2011), and Papakostopoulos et al. (2010). 5. On U.S. roads, this result means that the experienced riders were attending to traffic ahead and oncoming traffic more effectively and that inexperienced riders spent more time glancing at objects off the road to the right. 6. Is this a function of choosing machine control over vision? Or look at road surface whilst stopping?
10	Megías, A., Maldonado, A., Catena, A., Di Stasi, L.L., Serrano, J. and Candido, A., 2011. Modulation of attention and urgent decisions by affect-laden roadside advertisement in risky driving scenarios. Safety Science, 49(10), pp.1388-1393.	22 participants aged 18 – 25 Undergraduates from Granada university Right-hand drive. Spain	Eye-Link II eye-tracker Honda riding trainer Road scene projected onto a screen 12 road situations: 6 with hazards, 6 without. Hazards were typical to normal riding conditions – pedestrians in the road and oncoming road users. Avoidance of the hazard required using the front brake. Emotional cues were static roadside advertisements of varying emotional salience. 24 pictures in total. 8 – Unpleasant 8 – Neutral 8 – Pleasant To estimate the effects of affect-laden roadside advertisement on attention and braking behaviour we analysed eye fixation time while the cues were on, as well as accuracy and reaction time of braking to the subsequent risk situation.	1. Stronger gaze distraction, both in the size of the time window and in the magnitude of fixation times, is elicited by emotional pictures more than by neutral ones. The deepest and longest distraction is elicited by negative emotional advertisements. 2. Negative advertisements got more total fixation time (1907 ms) than the positive advertisements (1760 ms), and both more than the neutral ones (1601 ms). 3. Negative pictures got more fixations (5.77 fixations) than positive ones (5.35). 4. Negative pictures had later gaze disengagement.	Positive and negative emotional pictures can act as a distraction from allocating attention to the road. Yet, viewing negative pictures can also speed up braking responses.
11	Nathanael, D., Portouli, E., Gkikas, K. and Papakostopoulos, V., 2012. What does a motorcyclist look at while driving at urban arterials?. Work, 41(Supplement 1), pp.4900-4906.	Three males riders: mean age 30. Riding experience: mean 10 years Right-hand drive. Greece.	SMI iView XHED2 Mid-weight motorcycles. Rode on a circuit with motorway and urban roads. Speed of traffic was a variable 1. Free flow 2. Restricted flow 3. Congested	1. Longer fixations on Urban roads than motorway 2. Duration of fixations was higher in restricted flow than in free flow or congested flow on motorways and higher than free flow on an urban road. 3. On urban roads riders looked towards the centre of the field of view. Whereas on motorways riders looked closer to the front of the motorcycle. 4. On urban roads, participants looked more centrally in restricted, more to the right in free flow and even more to the right in congested flow. 5. On urban roads, in free flow and restricted flow riders looked closer to the front wheel. 6. For both road types, moving vehicles are the most common object to fixate on. 7. In free flow, fixations shift to other objects – such as pedestrians and road signage. 8. In case of congested flow on motorways riders fixated equally often on motorcycles and passenger cars. This may be because the motorcycle riders shift their focus from the jammed vehicles, who no longer pose a risk, to the moving motorcycles ahead of them, which they consider as “pathfinders” inside the dense car grid.	1. This may be due to motorcyclists trying to calculate potential hazards and their relative speed. 2. Restricted flow is likely more demanding of attention as other road users are in close proximity but travelling quickly – unlike congested flow where speeds are much lower. 3. Central fixations allow for wider scanning for hazards such as pedestrians. On a motorway there may be less likelihood of peripheral hazards and more likelihood of road surface being a hazard for the motorcyclists. Comparison with Mortimer and Jorgeson (1975), Nagayama et al. (1979), Papakostopoulos et al. (2010), Di Stasi et al. (2011), and Muttart et al. (2011). 4. Possibly restricted flow providing quick and attention demanding conditions which require a wider field of view for attention. In congested or free flow conditions there may be more hazards coming from the right-hand side (pedestrians stepping off the pavement and cars pulling out of side roads) 5. Where external demands allow, shifts in visual attention are towards road surface. 6. Moving vehicles seen as being the most hazardous to personal safety. 7. Freed external demands on attention allow wider visual search. 8. Comparison with Papakostopoulos (2010) – visually seeking motorcycles. But – is this also a function of social aspects of motorcycling.
12	Smith, T., Garets, S. and Cicchino, J., 2013. The effect of sight distance training on the visual scanning of motorcycle riders: a preliminary look (No. DOT HS 811 689). United States. National Highway Traffic Safety Administration. Office of Behavorial Safety Research. (Published Gov report)	31 (end data) Participants Split into three groups 1. 7 beginner-trained (by Team Oregon training school) (mean age 40) 2. 12 beginner-untrained (mean age 36). 3. 12 experienced riders (mean age 39). Right-hand drive. USA.	Giga Bit-E eye-tracker system MotionNode system mounted on the motorcycle for riding stats Closed course for manoeuvres such as figure of eight, and 9.4 mile open road route. Repeated measures design: 3 time points – baseline, 6 months, and 12 months.	1. Sight distance to safe stopping distance dropped below safe stopping distance significantly for Beginner-Untrained riders. 2. Mean gaze 95% confidence ellipse for the beginner-untrained riders was significantly larger than that for the experienced riders. 3. Beginner-Trained riders demonstrated significant positive effects from their gaze training at the 12-month test.	1. Link with Mortimer (1975) with visual field too close to the bike for safe stopping distance. Note; this sight to stopping distance was particularly reduced on curved areas of the course which may have hazard identification and time to respond to hazards greatly reduced. 2. Likely that gaze area of experienced and trained riders is specific to where hazards are expected. 3. There is an indication of long-term effects of the positive influence from explicit gaze training. Road surface comparison with Mortimer and Jorgeson (1975), Nagayama et al. (1979), Papakostopoulos et al. (2010), Di Stasi et al. (2011), Muttart et al. (2011), and Nathaneal et al. (2012).
13	Lobjois, R., Siegler, I.A. and Mars, F., 2016. Effects of visual roll on steering control and gaze behaviour in a motorcycle simulator. Transportation research part F: traffic psychology and behaviour, 38, pp.55-66.	12 motorcyclists who ride at least 5000km per year and have held their licence for at least 1 year. Right-hand drive. France.	125cc frame motorcycle simulator. Road scenario on screen. Pertech monocular, head-mounted eye-tracker. Record the eye-gaze when changing the tilt and roll of the road scenario.	1. Motorcyclists look at road surface often. 2. Gaze point distribution deviated beyond the tangent point by a visual angle of 2° on the horizontal and vertical axes. 3. More than 80% of gaze points directed beyond the tangent point in the horizontal direction – beyond the exit point of the bend.	1. Motorcyclists looking at the road surface is comparable with Mortimer and Jorgeson (1975), Nagayama et al. (1979), Papakostopoulos et al. (2010), Di Stasi et al. (2011), Muttart et al. (2011), Muttart et al. (2011), Nathaneal et al. (2012), and Smith et al. (2013) 2. This type of behaviour suggests that motorcyclists did not track the tangent point properly, but rather a steering point on the road ahead. 3. The motorcyclist looking to anticipate the exit of the bend rather than the tangent of the bend itself, this may be a contributary factor for COCs on bends due to overlooking key information regarding shape of bend or other hazards.
14	Muttart, J.W., Bartlett, W., Peck, L.R., Zafian, T.M., Yamani, Y., Kauderer, C., Fisher, D., Samuel, S. and Dinakar, S., 2017 January. Motorcyclists’ Glance Behaviors at Unsignalized Intersections: Left Turn across Path (No. 17-06688). Conference: Transportation Research Board. Washington, DC	20 Motorcycle-Driver participants (3 female). 10 Driver participants (2 female). Right-hand drive. USA.	Each motorcyclist used their own motorcycle to ride the fixed route. There was a Honda car provided for all participants for the driving condition. The motorcycles and car were fitted with a video V-Box system to provide a 120 degree angle field of view forward and sideways. Furthermore, speed and positioning were recorded via GPS. Participants wore MobileEye eye tracking glasses. All participants followed the same on-road route. The Motorcycle-Driver participants both rode and drove the route: this was counterbalanced. One of the left-turn intersections was traffic controlled with lights, the other two did not have lights. Glances were measured: ahead (towards oncoming traffic), before the left-turn intersection, and after taking the left-turn.	1. Motorcyclist – Driver participants glanced significantly more ahead before taking the left-turn, especially at intersections that do not have lights.	1. This additional glancing behaviour is likely to anticipate oncoming traffic and help judge oncoming speed, with the intention to avoid oncoming traffic while taking the left-turn.
15	Muttart, J., Bartlett, W., Bakhtiari, S., Zhang, T., Samuel, S., Zafian, T., Peck, L.R. and Kauderer, C., 2017, June. Comparison of Glancing Behaviors of Riders and Drivers at Unsignalized Intersections Involving Right Turns. In Driving Assessment Conference (Vol. 9, No. 2017). University of Iowa.	20 Motorcycle-Driver participants (3 female). 10 Driver participants (2 female). Right hand drive. USA.	Each motorcyclist used their own motorcycle to ride the fixed route. There was a Honda car provided for all participants for the driving condition. The motorcycles and car was fitted with a video V-Box system to provide a 120 degree angle field of view forward and sideways. Participants wore  MobileEye eye tracking glasses. All participants followed the same on-road route. The Motorcycle-Driver participants both rode and drove the route: this was counterbalanced. Two comparisons regarding number of glances before and after the right turn at an intersection.	1. The Motorcycle-Driver made more glances to the left to the left when riding compared to when driving after the intersection than before. Furthermore, they made more glances to the right after entry to the intersection than before the intersection.	1. Like Muttart et al. (2017) above, the glancing behaviour indicates the motorcyclists aimed to identify traffic which may conflict with the rider at the intersection.
16	Papakostopoulos, V., Nathanael, D. and Psarakis, L., 2019. Semantic content of motorcycle riders’ eye fixations during lane-splitting. Cognition, Technology & Work, pp.1-13.	Six male motorcycle riders, aged 24–39 years. Used their Own motorcycle. Right-hand drive. Greece.	SMI iView X™ HED2 eye-tracker system Rode a 15-minute route on an urban arterial motorway. Analysis was for ‘lane splitting’ known as ‘Filtering’ in the UK. Once finished a ‘think aloud’ dissemination technique was used – where the motorcyclist watched the footage from their and commented on it. Riding sessions were between 17.00 – 19.00 in good daylight.	1. Fixations were attributed to five object types, namely, imminent Road surface ahead (7%), 1st car (8%), 2nd car (17%), 3rd car (21%), and traffic ahead (46%). Eye-fixation results on traffic objects show that all six riders tended to fixate less on the proximal cars, allocating more fixations in monitoring the distant cars and even more on the traffic ahead. 2. When fixating on the cars there were eight most common car body parts the motorcyclists fixated on: (1) RWT = rear wing/tyre, (2) RL = rear light, (3) RQP = rear quarter pillar, (4) RW = rear window, (5) RE = roof edge, (6) FSW = front side window, (7) WM = wing mirror, and (8) FWT = front wing/tyre Regarding front tyres, all six riders mentioned that they intentionally monitor them to spot as early as possible an imminent lateral displacement of the car being overtaken.	1. Motorcyclists looking at the road surface is comparable with Mortimer and Jorgeson (1975), Nagayama et al. (1979), Papakostopoulos et al. (2010), Di Stasi et al. (2011), Muttart et al. (2011), Nathaneal et al. (2012), Smith et al. (2013), and Lobjois et al. (2016). 2. Motorcyclists were trying to anticipate other road users’ intentions and movements. Also note these participants were experienced motorcyclists, consequently they may have a learnt optimal visual strategy. It would be interesting to use novice riders as a comparison group.
17	Lobjois, R. and Mars, F., 2020. Effects of motorcycle simulator configurations on steering control and gaze behavior in bends. Journal of experimental psychology: applied.	18 Participants. Mean age: 32. Mean motorcycle licence: 7 years. Mean distance: 11,916 km per year. Right-hand drive. France.	Eye tracking as a perception of realism in a simulator. Investigating 3 different simulator configurations of roll and steering. 1, fixed-base with direct steering 2, motion-based with direct steering 3, motion-based with reverse steering Yamaha 125cc simulator, with various feedback options: including handlebar movement, tilt and roll of the simulator. Road scenario projected onto a screen. Pertech head-mounted monocular eye-tracker. Gaze behaviour was assessed in the present study by determining the horizontal deviation of gaze relative to the Tangent Point (TP) on the bend, which the point where the direction of the inside edge line seems to reverse from the rider’s viewpoint: also known as Limit Point in advanced riding literature. Participants completed the Simulator Sickness Questionnaire.	1. Differences in gaze were found between steering conditions: motion based reverse steering and both the direct steering conditions. 2. 73% of gaze points went beyond the TP of the bend. 3. The distribution of gaze points was also found to change as a function of bend section. A, Approaching the bend entry, more fixations were made in the direction of the lane centre. B, The first half of the bend, riders scanned the road well beyond the TP.	1. A difference in gaze when applying different steering conditions may be due to reverse (counter) steering is more realistic to real-life riding. 2. Higher percentage compared with similar prior car driver studies (Mars & Navarro, 2012). This may be due to increased speed of riders and a seating position which offers greater viewing angles. Or, perhaps a result of the ‘look where you want to steer’ instruction commonly given in motorcycle training. 3. More gaze towards the centre of the lane is possibly to identify where to position the motorcycle going into the bend. This is comparable to Lobjois et al. (2016) with motorcyclists looking for the exit point. Note that Papakostopoulos et al. (2010) motorcyclists looked at the TP whereas in this research and Lobjois et al (2016) motorcyclists looked beyond the TP. This may be a result of the difference in simulator vs field-based research.
18	Wong G and Wong Y. D. 2022. Young male motorcycle rider perception response times to abrupt and gradual onset hazards.	55 Participants. 28 with motorcycle licence. Mean age 23.5. 27 without motorcycle licence. Mean age 24.4. Left-hand drive. Singapore.	Eye tracking used to measure gaze towards abrupt onset and gradual onset hazards. Sensors built into the motorcycle to measure response times to hazards. The sensors included the accelerator: increase or release and braking. Accelerator was the key measurement. SensoMotoric Instruments (SMI) wearable eye-tracking glasses. The measurements were, Time to First Fixation from onset of the hazard and the Fixation to release of accelerator. Honda Wave motorcycle simulator with screen used to display the 3D computer generated scenarios. Two scenario types were used; junction hazards and linked hazards. Each with sub-categories of early onset and gradual onset hazards.	1. Gradual onset hazards have longer response times. Both the Time to First Fixation and Accelerator Release time was longer. 2. No significant difference in accelerator response time according to experience/license; although qualitatively P value was close to significance. 3. Significant difference in Time to First Fixation for experienced riders in the linked hazards condition. No significant different in the junction hazards condition.	3. It may be possible that Experienced Riders are cognitively primed to look for the hazards in the linked hazards condition, hence the quicker Time to First Fixation.

3. Results

3.1 Data Synthesis

A narrative synthesis approach was used due to the methodological and statistical heterogeneity of the research. Narrative synthesis uses a textual approach to summarise research which is difficult to compare using statistical methods such as meta-analysis (Popay et al., 2006). The synthesis included classification and concept mapping to identify comparisons (Mulrow et al., 1998).

This scoping review contained eighteen research articles from nine countries. Five articles originated from USA. Three articles from Spain and Greece. Two articles from France. One article each from Japan, Australia, China, Singapore and U.K. 30% of research originated from left-hand drive countries.

3.2 Motorcyclists’ Visual Gaze and Visual Fixations

3.2.1 Motorcyclists’ visual gaze can be directed towards the road surface.

Nine out of the eighteen (50%) articles reported that motorcyclists look at the road surface. Furthermore, Mortimer and Jorgeson (1975) identified motorcyclists eye fixations were nearer (within 30-76 metres) to the motorcycle than when driving a car 14.1% of the time. Similarly, Nagayama et al. (1979) found motorcyclists fixate on road surface 30% of the time compared with car drivers who barely fixated on the road surface. Moreover, Papakostopoulos, et al. (2019) report that motorcyclists look at the near road surface 7% of the time.

3.2.2 Motorcyclists’ visual gaze at the road surface can be closer than safe stopping distance.

Smith et al. (2013) identified the sight distance to safe stopping distance dropped below safe stopping distance significantly for Beginner-Untrained riders. Even though it was not explicitly reported by Mortimer and Jorgeson (1975), calculations from the sight distance to stopping distance for their experienced rider participants was below safe stopping distance; because the motorcyclists were travelling an average of 45 miles per hour with a safe stopping distance of 44 metres yet periodically their visual gaze was as close as 30m. However, Smith et al. (2013) found that it is possible to train beginner riders to look further ahead beyond the safe stopping distance. Furthermore, the effect of sight training was present 12 months later, indicating that eye-gaze training can have a lasting impact on vision.

3.2.3 Motorcyclists demonstrate different gaze according to situations.

Motorcyclists demonstrate a targeted field of view and adjust their gaze strategies depending on situation.

Papakostopoulos et al. (2010) discovered that motorcyclists change their visual gaze depending on road situation. For example, on a motorway, the motorcyclists fixated far ahead with visual gaze moving on a vertical plane, whereas at intersections the motorcyclists actively looked for hazards such as road surface and used a vertical and horizontal plane. Likewise, Hosking et al. (2010) identified that motorcyclists used different visual strategies according to road scenario and increased horizontal eye movements in the urban scenario. Moreover, duration of eye-fixation changed according to situation. For example, Nathanael et al. (2012) indicated durations were longer in urban and restricted flow traffic compared with free-flowing motorway situations.

Motorcyclists also adjust their gaze towards other road users who may be hazardous to the motorcyclist. For example, Papakostopoulos et al. (2019) found when filtering in between two lanes of traffic the motorcyclists’ fixated on the 2^nd and 3^rd car in front of them as these cars presented highest potential for danger. The motorcyclists self-reported they were trying to anticipate the movements of the cars, such as by looking at the front wheel to judge if the car was going to change direction (Papakostopoulos et al., 2019).

3.2.4 Experience may be a factor in visual strategies for hazard perception.

Hoskin et al. (2010) reported those with motorcycle experience had significantly different visual scanning strategies, whereby the field of view was narrower for potential hazard zones. Likewise, Di Stasi et al. (2011) found experienced motorcyclists targeted their gaze towards relevant locations where there may be hazards, such as rear-view mirrors, and their visual gaze was less in irrelevant locations which presented little risk. Similarly, Muttart et al. (2011) identified experienced motorcyclists targeted their gaze more often towards oncoming traffic compared with inexperienced motorcyclists whose gaze was in irrelevant zones for hazards. Furthermore, experienced motorcyclists exhibit quicker time to first fixation towards hazards when presented with linked hazard situations (Wong & Wong, 2022).

Importantly, it may be possible to train inexperienced motorcyclists’ gaze towards hazardous scenarios. For example, Di Stasi et al. (2011) findings demonstrated that once inexperienced participants received training to look for hazardous areas there was no significant difference in visual scanning strategy between experienced and inexperienced participants.

3.2.5 Motorcyclists’ visual gaze when negotiating bends.

Papakostopoulos et al. (2010) reported that motorcyclists fixate on the Tangent Point (TP) of a bend. The TP is where the arc of the bend ceases or changes. However, Lobjois et al. (2016 & 2020) found motorcyclists try to look past the TP towards the exit point. Note, the Lobjois et al. (2016 & 2020) research is laboratory based on a riding simulator whereas the Papakostopoulos et al. (2010) research is field-based on a motorcycle. Consequently, there could be fundamental differences in gaze due to other variables related to research methodology.

3.2.6 Motorcyclists’ visual gaze negotiating intersections/junctions.

Muttart et al. (2011) reported 19 of 23 participants had fewer glances toward the area of most significant hazard potential before turning when riding a motorcycle compared with when driving, and less likely to make a glance to the right before arriving at the stop line. Muttart et al. (2011) suggests this may be a diversion of attention due to a function of intentional control of motorcycle handling to keep the motorcycle upright. Conversely, Muttart et al. (2017a, 2017b) found that motorcyclists glanced in the direction of potential hazards on the approach to an intersection and once they had taken the turn. Furthermore, Underwood et al. (2011) discovered motorcyclists had longer fixations on oncoming traffic compared with car drivers when deciding whether it was safe to pull out of a junction into a road.

3.2.7 What draws motorcyclists’ visual gaze?

Research suggests that motorcyclists’ visual gaze is directed to areas of potential hazards such as road surface and oncoming vehicles. However, there are additional factors that draw motorcyclists’ visual gaze such as roadside advertising and other motorcyclists. Papakostopoulos et al. (2010) and Nathanael et al. (2012) found the participant motorcyclists looked towards other nearby motorcyclists more than nearby vehicles. In both research it was suggested that this may be a function of using other motorcyclists as pathfinders through traffic. Additionally, Papakostopoulos et al. (2010) suggested that looking at other motorcyclists is potentially lightening cognitive load for visual scanning.

Megias et al. (2011) discovered that advertisements containing positive and negative scenes which provoke an emotional response draw gaze far more than neutral scenes. Additionally, negative scenes had most fixations and longest fixation durations than any other scene type.

3.3 Different technologies used to track motorcyclists’ visual gaze.

Nine studies were laboratory-based using either motorcycle simulators or photographs of road scenes. Nine studies were field-based with a motorcyclist wearing eye-tracking equipment and riding a pre-arranged route. Consequently, it is prudent to identify equipment used for field-based and laboratory-based research separately.

Early eye-tracking equipment was crude and cumbersome. For their field-based research, Mortimer and Jorgeson (1975) and Nagayama, et al. (1979) used an eye-marker which was head mounted with a modified helmet worn by the rider. The recording equipment was either attached to the motorcycle or worn by the motorcyclist. To record the road route; Mortimer and Jorgeson (1975) had an assistant holding a video recorder as a passenger on the motorcycle and in the car. Nagayama et al. (1979) used a video recorder worn by the motorcyclist.

Papakostopoulos et al. (2010) used an unspecified portable eye-tracker mounted to the rider’s helmet. The motorcyclists’ raw eye-tracking data were turned into scan-paths and the scan-paths were later superimposed to the original video.

Muttart et al. (2011, 2017a, 2017b) used a MobileEye eye-tracking system which consisted of safety glasses with a forward-facing camera and a camera aimed toward the corneal reflections of the eye. The two images from each camera were overlaid during calibration.

Nathanael et al. (2012) used an SMI iView XHED2, which records the traffic scene from the rider’s point of view and identifies the rider’s fixations points. The scene was played-back off-line with the identified fixation points overlaid to the traffic scene.

Smith et al. (2013) used a Giga Bit-E eye-tracker system which was synchronized with a Speedbox transducer that monitored the motorcycle speed as well as the instantaneous GPS location. The data included the forward view of the motorcycle rider as well as a visual overlay of the x and y axis eye location within the field of view of the rider. Papakostopoulos et al. (2019) used an SMI iView X™ HED2 eye-tracker system, like that of Nathanael et al. (2012).

The laboratory-based research used an eye-tracker in conjunction with a motorcycle simulator and a road scenario played on a screen in front of the participant.

The Eye-Link ii tracker was used by Di Stasi et al. (2009 & 2011), Megias et al. (2011), and Underwood et al. (2011). A Pertech monocular eye-tracker was used by Lobjois et al. (2016 & 2020). Cheng et al. (2011) used an iViewX head mounted system. Finally, Wong and Wong (2022) used SMI glasses.

3.4 Record of participants.

Most of the field-based research using motorcyclists included a low number of participants with an average of 13. For example, Mortimer and Jorgeson (1975) used only 2 male participants. Conversely, Smith et al. (2013) had 31 participants, which was the largest number in the field-based research. Additionally, most research used only male participants, or there was a low number of female participants.

The laboratory-based research had larger numbers of participants, average 47, with better gender representation. Lobjois et al. (2016) used the smallest number of participants 12, and Cheng et al. (2011) had the largest number of participants 109. Yet, Di Stasi et al. (2009) and Megias et al. (2011) used mostly undergraduate university students who were not motorcyclists.

4.Discussion

This scoping review demonstrates the low number of research articles regarding motorcyclists’ visual gaze, consequently research findings may not be representative of the whole motorcycling population. Yet, despite the low number of articles, there are key recurring outcomes, for instance motorcyclists’ visual gaze towards road surface is reported in 50% of the research. Moreover, the research demonstrates that motorcyclists typically look towards areas of potential hazards such as oncoming cars and road margins. Additionally, experience may be a key factor for developing targeted search strategies and motorcyclist’s visual gaze can change according to road scenario with differences in gaze and fixation duration in urban and motorway environments with different traffic density.

The comparisons with car drivers visual search are quite similar in some instances. For example, Di Stasi et al. (2011) found experienced motorcyclists’ visual gaze targeted to relevant locations such as rear-view mirrors, likewise, Mourant and Rockwell (1972) reported that inexperienced drivers looked less in their rear-view mirrors compared with experienced drivers.

Additionally, motorcyclists’ visual search strategy can change according to road situation (Hosking et al., 2010; Nathanael et al., 2012; Papakostopoulos et al., 2010 & 2019). These results are like Crundall et al. (2012) as car drivers with differing experience had different search strategies according to different road situations. Furthermore, experienced motorcycle riders have different visual scanning paths compared with novice riders, this too has findings reflected in car driver studies. For example, Chapman and Underwood (1998) report that experienced drivers had shorter fixation durations and less vertical variance in fixation locations than novices. Moreover, Crundall and Underwood (1998) found that novice drivers did not vary their search strategy across three road types, whereas the experienced drivers significantly adjusted their search according to road type.

Regardless of road type, findings from this scoping review suggest motorcyclists frequently look too close to the road surface in front of them, which in turn could affect the safe stopping distance for the motorcycle (Mortimer & Jorgeson 1975, Smith et al., 2012). However, motorcyclists should monitor road surface due to the hazard that poor road surfaces can present for safe handling of the motorcycle (Department for Transport, 2014). For example, the sand used on an intersection in the Muttart et al. (2011) research could cause motorcyclists to lose traction. Consequently, this poses a conundrum for safe handling of the motorcycle and adequate forward observations.

Yet, if experience can influence optimal visual scanning strategies, then it may be possible to incorporate both visual scanning for road surface and for looking beyond the safe stopping distance. Findings indicate it is possible to train novice motorcyclists to adjust their visual gaze to match that of experienced motorcyclists. For example, Di Stasi et al. (2011) found improvements in novices’ visual gaze strategy after giving them training and likewise Smith et al. (2012) found that novices could be trained to look beyond the safe stopping distance. Therefore, future research could investigate optimal visual gaze training for novice riders as part of motorcycle training courses.

There are a several limitations to the visual attention research in this scoping review, much is directly related to the type of methodology used. For example, the field-based research used very low participant numbers and predominantly males. Yet, according to Holmqvist et al. (2011) eye-tracking research may not require large numbers of participants for statistical significance. However, the field-based research consistently used low numbers of participants and crucial data may not be recorded from only two or three participants.

Nonetheless, the field-based research offers a strength of ecological validity which laboratory-based research cannot fulfil. Conversely, the laboratory research provided increased experimental control. Whereas it is likely that each field-based motorcyclist encountered many different variables even when travelling the same route and the same time of day. Consequently, each research article, regardless of its methodology has its strengths and limitations. Some findings are recurring regardless of methodology, such as motorcyclists targeted visual gaze strategies towards hazards or looking towards the road surface.

From this scoping review it is clear there is a limited amount of research regarding motorcyclists’ visual gaze, which is concerning considering that motorcyclists worldwide are a vulnerable road user group. More research should be targeted towards how to improve motorcyclists’ visual gaze towards hazards. Experienced motorcyclists may provide a basis for examples of optimal visual strategy. For example, Wong and Wong (2022) identified quicker time to first fixation for a linked hazard condition, potentially indicating a cognitive priming element for those with experience. Furthermore, investigating crash prone versus crash free motorcyclists, such as with Cheng et al. (2011), may provide useful insights into visual gaze strategies and response to hazards to avoid collisions.

5. Scoping review limitations

This scoping review for motorcyclists’ visual gaze is the first of its type, which has provided an overview of the prior research and methodology used. Importantly the results demonstrate some recurring themes which might have real world application for improvement in optimal visual attention training for motorcyclists and suggests future topics of research. However, there are limitations to this review which should be considered.

This scoping review was conducted by a sole researcher, yet JBI suggests that scoping studies and systematic reviews should be conducted by at least two researchers to provide balance and prevent bias (Peters et al., 2020). The author was aware of this guideline and had taken steps to prevent bias and improve rigour by having regular meetings with a wider team of academics. The draft and accompanying research studies were reviewed by four academics to seek objective feedback. Additionally, the researcher followed the Arksey and O’Malley framework (2005) alongside Prisma ScR and JBI guidance (Tricco et al., 2009: Peters, et al., 2020), as adherence to such guidance provides a basis for academic rigour.

Moreover, the author had a specialist librarian check the search terms and search strategy in accordance with Arksey and O’Malley (2005) guidance. Yet, because the institution where the researcher is based does not subscribe to engineering journals, there are search limitations which may lead to omission of relevant journals which may contain motorcyclists’ visual gaze for transport development research.

6. Conclusions

The scoping review question was, ‘What is the existing research regarding motorcyclists’ visual gaze using eye-tracking equipment’. As such the results of this review demonstrate there is limited research regarding motorcyclists’ visual gaze using eye-tracking equipment. Additionally, the research is split between laboratory and field-based methodology and the equipment used ranges from complex eye-tracking technology to basic eye-monitoring units and cumbersome video recorders.

This scoping review has identified commonalities in research findings regarding such as motorcyclists gazing at road surface more than car drivers. However, differences in gaze strategy have been identified between experienced and novice motorcyclists and importantly that novices may be taught to change their gaze strategy to better identify hazards, with long-term effect. Consequently, the results of this scoping review might provide a foundation on which to apply research findings to real-world situations, such as when training novice motorcyclists. Moreover, this scoping review can assist with planning future research to further investigate motorcyclists’ visual gaze. Finally, findings from this scoping review may help inform public policy. For example, there is evidence to support the maintenance of road surfaces to prevent motorcyclists from fixating on poor road surfaces to the detriment of identifying other hazards, which in turn will benefit all road users.

Funding statement

The funders are Bournemouth University and DocBike U.K., a motorcycle collision and injury prevention charity.

Acknowledgements

Acknowledgements to Professor Ann Hemingway, Professor Peter Hills, Assistant Professor Ursula Rolfe, and Dr Ian Mew for reading through this scoping review and providing feedback. Furthermore, I offer gratitude to the specialist librarian at Bournemouth University, Caspian Dugdale, for assistance with the literature search.

Disclosure statement

The author declares there is no conflict of interest.

References

ACEM (2004). MAIDS In-depth investigation of accidents involving powered two wheelers. Association of European Motorcycle Manufacturers (ACEM), Brussels. Available from: https://www.maids-study.eu/index.php?error=hastolog [Accessed October 2023]

Araujo, M., Illanes, E., Chapman, E. and Rodrigues, E., 2017. Effectiveness of interventions to prevent motorcycle injuries: systematic review of the literature. International journal of injury control and safety promotion, 24(3), pp.406-422.

Arksey, H. and O’Malley, L., 2005. Scoping studies: towards a methodological framework. International journal of social research methodology, 8(1), pp.19-32.

Chapman, P.R. and Underwood, G., 1998. Visual search of driving situations: Danger and experience. Perception, 27(8), pp.951-964.

Cheng, A.S., Ng, T.C. and Lee, H.C., 2011. A comparison of the hazard perception ability of accident-involved and accident-free motorcycle riders. Accident; Analysis & Prevention, 43(4), pp.1464-1471.

Cho, H.J. and Wu, Y.T., 2004, October. Modeling and simulation of motorcycle traffic flow. In 2004 IEEE International Conference on Systems, Man and Cybernetics (IEEE Cat. No. 04CH37583) (Vol. 7, pp. 6262-6267). IEEE.

Crundall, D., Chapman, P., Trawley, S., Collins, L., Van Loon, E., Andrews, B. and Underwood, G., 2012. Some hazards are more attractive than others: Drivers of varying experience respond differently to different types of hazard. COC Analysis & Prevention, 45, pp.600-609.

Crundall, D.E. and Underwood, G., 1998. Effects of experience and processing demands on visual information acquisition in drivers. Ergonomics, 41(4), pp.448-458.

Collins, 2020. Definition of Pavement. Available from: https://www.collinsdictionary.com/dictionary/english/pavement. [Accessed October 2023].

Connor, C.E., Egeth, H.E. and Yantis, S., 2004. Visual attention: bottom-up versus top-down. Current biology, 14(19), pp.R850-R852.

Department for Transport, 2014. Contributory factors to reported road accidents 2014. Available from https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/463043/rrcgb2014-02.pdf [Accessed October 2023]

Department for Transport. 2023. Reported Road casualties Great Britain: motorcyclist factsheet 2022. Available from: https://www.gov.uk/government/statistics/reported-road-casualties-great-britain-motorcyclist-factsheet-2022/reported-road-casualties-great-britain-motorcyclist-factsheet-2022 [Accessed November 2023]

Di Stasi, L.L., Álvarez-Valbuena, V., Cañas, J.J., Maldonado, A., Catena, A., Antolí, A. and Candido, A., 2009. Risk behaviour and mental workload: Multimodal assessment techniques applied to motorbike riding simulation. Transportation research part F: traffic psychology and behaviour, 12(5), pp.361-370.

Di Stasi, L.L., Contreras, D., Cándido, A., Cañas, J.J. and Catena, A., 2011. Behavioral and eye-movement measures to track improvements in driving skills of vulnerable road users: First-time motorcycle riders. Transportation research part F: traffic psychology and behaviour, 14(1), pp.26-35.

Hosking, S.G., Liu, C.C. and Bayly, M., 2010. The visual search patterns and hazard responses of experienced and inexperienced motorcycle riders. COC Analysis & Prevention, 42(1), pp.196-202.

Holmqvist, K., Nyström, M., Andersson, R., Dewhurst, R., Jarodzka, H. and Van de Weijer, J., 2011. Eye tracking: A comprehensive guide to methods and measures. OUP Oxford.

Ju, H.W. and Johnson, K.K., 2010. Fashion advertisements and young women: Determining visual attention using eye tracking. Clothing and Textiles Research Journal, 28(3), pp.159-173.

Kimble, M.O., Fleming, K., Bandy, C., Kim, J. and Zambetti, A., 2010. Eye tracking and visual attention to threating stimuli in veterans of the Iraq war. Journal of anxiety disorders, 24(3), pp.293-299.

Lamme, V.A., 2003. Why visual attention and awareness are different. Trends in cognitive sciences, 7(1), pp.12-18.

Lobjois, R. and Mars, F., 2020. Effects of motorcycle simulator configurations on steering control and gaze behavior in bends. Journal of experimental psychology: applied, 26(1), p.108.

Lobjois, R., Siegler, I.A. and Mars, F., 2016. Effects of visual roll on steering control and gaze behavior in a motorcycle simulator. Transportation research part F: traffic psychology and behaviour, 38, pp.55-66.

Megías, A., Maldonado, A., Catena, A., Di Stasi, L.L., Serrano, J. and Candido, A., 2011. Modulation of attention and urgent decisions by affect-laden roadside advertisement in risky driving scenarios. Safety Science, 49(10), pp.1388-1393.

Mortimer, R.G. and Jorgeson, C.M., 1975. Comparison of eye fixations of operators of motorcycles and automobiles. SAE Transactions, pp.930-935.

Mourant, R.R. and Rockwell, T.H., 1972. Strategies of visual search by novice and experienced drivers. Human factors, 14(4), pp.325-335.

Mulrow, C., Langhorne, P. and Grimshaw, J., 1997. Integrating heterogeneous pieces of evidence in systematic reviews. Annals of internal medicine, 127(11), pp.989-995.

Munn, Z., Peters, M.D., Stern, C., Tufanaru, C., McArthur, A. and Aromataris, E., 2018. Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC medical research methodology, 18(1), p.143

Muttart, J., Barttlett, W., Bakhtiari, S., Zhang, T., Samuel, S., Zafian, T., Peck, L.R. and Kauderer, C., 2017(a), June. Comparison of Glancing Behaviors of Riders and Drivers at Unsignalized Intersections Involving Right Turns. In Driving Assessment Conference (Vol. 9, No. 2017). University of Iowa.

Muttart, J.W., Bartlett, W., Peck, L.R., Zafian, T.M., Yamani, Y., Kauderer, C., Fisher, D., Samuel, S. and Dinakar, S., 2017(b). Motorcyclists’ Glance Behaviors at Unsignalized Intersections: Left Turn across Path. Conference: Transportation Research Board. (No. 17-06688). Washington, DC

Muttart, J.W., Peck, L.R., Guderian, S., Bartlett, W., Ton, L.P., Kauderer, C., Fisher, D.L. and Manning, J.E., 2011. Glancing and stopping behavior of motorcyclists and car drivers at intersections. Transportation research record, 2265(1), pp.81-88.

Nagayama, Y., Morita, T., Miura, T., Watanabem, J. and Murakami, N., 1979. Motorcyclists’ visual scanning pattern in comparison with automobile drivers’. SAE Transactions, pp.934-945.

Nathanael, D., Portouli, E., Gkikas, K. and Papakostopoulos, V., 2012. What does a motorcyclist look at while driving at urban arterials?. Work, 41(Supplement 1), pp.4900-4906.

Papakostopoulos, V., Nathanael, D. and Marmaras, N. 2010. An explorative study of visual scanning strategies of motorcyclists in urban environment. In Proceedings of the 28th Annual European Conference on Cognitive Ergonomics. pp. 157-160.

Papakostopoulos, V., Nathanael, D. and Psarakis, L., 2019. Semantic content of motorcycle riders’ eye fixations during lane-splitting. Cognition, Technology & Work, pp.1-13.

Peters MDJ, Godfrey C, McInerney P, Munn Z, Tricco AC, Khalil, H., 2020. Chapter 11: Scoping Reviews. In: Aromataris E, Munn Z (Editors). JBI Manual for Evidence Synthesis, JBI, 2020. Available from https://synthesismanual.jbi.global.  https://doi.org/10.46658/JBIMES-20-12

Popay, J., Roberts, H., Sowden, A., Petticrew, M., Arai, L., Rodgers, M., Britten, N., Roen, K. and Duffy, S., 2006. Guidance on the conduct of narrative synthesis in systematic reviews. A product from the ESRC methods programme Version, 1, p.b92.

RoSPA. 2017. Road Safety Factsheet. Available from: https://www.rospa.com/rospaweb/docs/advice-services/road-safety/road-crashes-overview.pdf [Accessed November 2023]

Smith, T., Garets, S. and Cicchino, J., 2013. The effect of sight distance training on the visual scanning of motorcycle riders: a preliminary look (No. DOT HS 811 689). United States. National Highway Traffic Safety Administration. Office of Behavioral Safety Research.

Tufanaru C, Munn Z, Aromataris E, Campbell J, Hopp L. 2020. Chapter 3: Systematic reviews of effectiveness. In: Aromataris E, Munn Z (Editors). JBI Manual for Evidence Synthesis [Online]. JBI, 2020. Available from: https://synthesismanual.jbi.global [accessed September 2023]

Tricco AC, Lillie E, Zarin W, O’Brien KK, Colquhoun H, Levac D, et al. 2018. PRISMA Extension for Scoping Reviews (PRISMAScR): Checklist and Explanation. Ann Intern Med. 2018;169:467–473. doi: 10.7326/M18-0850.

Underwood, G., Humphrey, K. and Van Loon, E., 2011. Decisions about objects in real-world scenes are influenced by visual saliency before and during their inspection. Vision research, 51(18), pp.2031-2038.

Ungerleider, G and Kastner, S., 2000. Mechanisms of visual attention in the human cortex. Annual review of neuroscience, 23(1), pp.315-341.

Wong, G. and Wong, Y.D., 2022. Young male motorcycle rider perception response times to abrupt-and gradual-onset hazards. Accident Analysis & Prevention, 165, p.106519.

World Health Organization, 2015. Global status report on road safety 2015. World Health Organization. Accessed from: https://www.who.int/publications/i/item/9789241565066. [Accessed September 2023].

Worldstandards, 2023. List of all left & right-driving countries around the world. Available from: https://www.worldstandards.eu/cars/list-of-left-driving-countries/. [Accessed October 2023].