The relation between driving errors and executive functioning in intellectually able young novice drivers with autism
Introduction
Driving is an important step towards gaining autonomy as it allows the development and maintenance of work-related and social contacts (Cox, Reeve, Cox, & Cox, 2012). Yet, the ability to drive safely is hard to acquire, especially for persons on the autism spectrum. The official diagnosis of an autism spectrum disorder (ASD) refers to “a neurodevelopmental disorder characterized by impairments in social interaction and communication, as well as repetitive behaviors and restricted interests” (American Psychological Association, 2013). A recent study from the US compared adolescents with and without ASD from the Children’s Hospital of Philadelphia healthcare network patient database. Although 83.5% of the neurotypical adolescents acquired a driving license by the age of 21, only 1 in 3 of the ASD adolescents did so (Curry, Yerys, Huang, & Metzger, 2018).
Driving comprises several subtasks running in parallel, between which one must be able to switch in a smooth manner (e.g., shifting gears, steering, changing lanes, and keeping traffic rules into account). Sudden changes in the traffic environment (e.g., traffic density, weather conditions) are additional difficulties. Hence, driving is a complex goal-directed task that places high demands on perceptual, cognitive, and motor processes (Monahan et al., 2013, Ross et al., 2014, Ross, Jongen, Brijs, et al., 2015). Therefore, it is not surprising that driver errors contribute to 70–75% of driver collisions, indicating that driver errors are directly related to traffic safety (Allahyari et al., 2008, Stanton and Salmon, 2009).
Executive functions (EFs), such as set shifting, working memory, and response inhibition, refer to a cluster of higher-order cognitive processes mediated by the prefrontal cortex, which enable an individual to perform goal-directed actions and problem solving (Rapport, Orban, Kofler, & Friedman, 2013). The relation between EFs and driving ability has already been investigated in neurotypical young novice drivers as they constitute a risk group for crashes.
For one part, this increased accident risk has been explained by insufficient driving experience (McCartt et al., 2003, Sagberg and Bjørnskau, 2006). Another explanation involves the fact that the adolescent brain has not fully matured yet (Gogtay and Thompson, 2010, Gogtay et al., 2004, Steinberg, 2005). Neuroscientific evidence shows that the brain areas providing behavioral ‘drive’, the limbic system, mature early. Meanwhile, the areas responsible for control over behavioral drive, the fronto-striatal connections, mature into young adulthood (Casey et al., 2011, Stevens et al., 2007). Different effects of this maturation process can be postulated. First, this developmental imbalance can create an excessive amount of 'drive', which in turn may result in risky behavior (Steinberg, 2005). The effect of this imbalance is especially prominent in male drivers, who weigh the benefits of risk taking more heavily than the costs compared to female drivers (Gardner & Steinberg, 2005). Second, many aspects of driving (e.g., vehicle control: Gugerty, 2011) only become automated over time with increasing driving experience. Since non-automated tasks require a larger investment of cognitive resources, novice drivers need to devote more of their already sparse resources to the driving task (Ross et al., 2014). One important driving ability is hazard avoidance, “the process of avoiding a collision with a hazard from initial searching for hazards through to the successful selection of an appropriate response” (Crundall & Pradhan, 2016). It is possible to identify several sub processes in hazard avoidance, for instance, hazard searching, fixation, mitigation, reaction, and response. A full description of all processes goes beyond the scope of this article. We refer the reader to Crundall and Pradhan (2016) for a detailed definition and delineation of all the different sub processes. In the current study, we included hazards since responses towards them depend both on driving experience and cognitive resources, aspects that are both relevant for young novice drivers.
ASD is often accompanied by EF difficulties such as problem-solving, cognitive flexibility, WM, self-monitoring, and generating novel solutions when adjusting to unexpected changes (Chen et al., 2016, Hill, 2004, Hughes et al., 1994, Pellicano, 2012, Turner, 1999). Thus, adolescents with ASD simultaneously fall into two potential risk categories: they belong to the novice driver population and show EF difficulties. Nevertheless, research on driving performance of novice drivers with ASD is still too scarce (e.g., Huang et al., 2012, Ross, Jongen, Brijs, et al., 2015), and often does not relate driving performance to EF, or does not include hazards. A summary of the research in ASD that related driving performance to EF or included hazards is summarized below.
Cox et al. (2016), studied a sample ranging from 15 to 23 years of age, and showed a different response to increased WM load in ASD compared to neurotypical controls. Increased WM demands resulted in decreased steering and braking in the ASD group, whereas it increased steering and braking in the control group, during a simulated drive. Classen, Monahan, and Hernandez (2013) linked increased driving errors (e.g., speed regulation, lane maintenance) to selective and divided attention in both pre-licensed and licensed adolescents with ASD. Furthermore, Daly, Nicholls, Patrick, Brinckman, and Schultheis (2014) speculated that driving errors might relate to EF difficulties. Via self-report, they found that licensed adults with ASD considered themselves as ‘poor drivers’, and also reported to commit more driving errors compared to non-ASD participants. Chee, Lee, Patomella, and Falkmer (2017) used a driving simulator, the Driving Behaviour Questionnaire (DBQ), and measures of cognitive and visual-motor ability. They found a worse performance in ASD participants, compared to a typically developing control group, with respect to some measures. Specifically, they reported more lapses (i.e., inability to focus and effectively allocate and sustain attention) during driving, more driving mistakes, and slower reactions in complex situations, during simulated driving. However, ASD participants did not show as much tailgating as the control group. Finally, some errors could be related to insufficient attentional capacity in the ASD group. Chee et al. (2017) investigated several driving measures (i.e., speed exceedances, collisions, pedestrians hits, centerline crossings, red light tickets, and tailgating). As for hazardous situations, Chee et al. (2017) only measured the ultimate outcome (i.e., collisions). However, measurement of reaction times to approaching hazards provides additional relevant information as slower reaction times lead to an increased collision risk (Bishop, Biasini, & Stavrinos, 2017).
Some studies investigated hazard avoidance in particular and compared responses to social with responses to non-social hazards. Hazards can be defined as social in case of a clearly visible person, compared to non-social in case of an object such as a car. In non-social conditions, the hazard can involve a human element such as a driver in a car, but the human element should not be visible (Bishop et al., 2017, Sheppard et al., 2010, Sheppard et al., 2017). For instance, Sheppard and colleagues used video clips and found atypical processing of road hazards. The latter study further specified that this was probably caused by slower attention orienting. Although in the first study ASD participants were found to respond more slowly to social hazards (Sheppard et al., 2010), the second study (Sheppard et al., 2017) did not find such a difference. One limitation of both studies was that they used videos instead of actual driving. A third study from Bishop et al. (2017) used driving simulation and found differences in hazard avoidance performance between yound adults with and without ASD that related to the social nature of hazards. Specifically, participants without ASD responded quicker to social hazards, whereas participants with ASD responded just as quick to social and non-social hazards (Bishop et al., 2017). Although these studies distinguished social and non-social hazards, none of them used the distinction previously proposed by Crundall and colleagues (Crundall et al., 2012, Crundall et al., 2010), i.e., behavioral prediction (BP) hazards (e.g., a parked car pulls out in front of the driver after a passenger has left the vehicle), environmental prediction (EP) hazards (e.g., two pedestrians are hidden by a bus shuttle and start to cross when the driver passes by), and dividing and focusing attention (DF) hazards (e.g., in a small curvy road, an approaching lorry comes from a small blind bend, pulls out to avoid crashing into a pedestrian, and occupies the driver's lane). One could hypothesize different performance of ASD drivers based on the respective category. For instance, due to difficulties with multitasking and mental flexibility (e.g., Rajendran et al., 2011, Van Eylen et al., 2011), one could expect more difficulties with DF hazards as these contain multiple potential hazards between which one must alternately switch attention.
In sum, there are indications that EFs could play a role in the driving performance of young adults with ASD. Given the relation between driving errors and traffic safety, it is important to better understand driving errors in ASD, and their relation with underlying EF mechanisms. Moreover, special attention is given to road hazards.
This study aimed to replicate and extend previous research to answer the following three questions: (1) do ASD participants exhibit a divergent pattern of performance on EF tests related to driving, compared to their neurotypical peers? (2) do ASD participants exhibit a divergent pattern of driving performance compared to their neurotypical peers? and (3), are differences in driving performance related to performances on EF tasks?
Section snippets
Methods
The current paper expands on the proceeding paper that was presented at Road Safety and Simulation (RSS) 2017 (Ross et al., 2017). While the proceeding paper focused mainly on the response to hazards and working memory, the current analyses expanded the topic to driving errors in general, and to multiple EF measures (see below). This study was approved by the ethical committees of Hasselt University and the Catholic University of Leuven (reference number ML10787).
Results
The descriptive statistics of the two groups are shown in Table 2. Table 3 displays the correlations between the EF measures per group.
Participants with ASD appeared to perform worse on the SSRT than the control group, explaining 4.6% of variance in SSRT. Nevertheless, group did not significantly predict SSRT performance, F(1,32) = 1.55, β = −0.22, p = .22. Meanwhile, group did significantly predict performance on the WM and UFOV tasks. Group explained 15.2% of the variance in the WM task, F
Discussion
The aim of this study was to investigate the relation between driving errors and EFs in young novice drivers with ASD. To this aim, we examined whether adolescents with ASD would show a divergent driving performance and response to hazards compared to neurotypical adolescents, and whether these differences would relate to differences in EFs.
Implications
The current results imply that high functioning young adults with ASD can be considered as quite capable drivers once they learn how to drive. However, future studies could focus on dividing and focusing hazards, and possibly include them in ASD driver training programs. One important way to perform training is a driving simulator. The scant research that exists indeed suggests merits of using driving simulators to train driving in people with ASD (e.g., Cox et al., 2017, Wade et al., 2017). To
Limitations
First, the sample size limited the statistical power and generalizability of this study, although the sample size of the current study is fairly equal to sample sizes in other comparable studies (e.g., Cox et al., 2016, Classen et al., 2013, Reimer et al., 2013). A larger sample size would have made it easier to obtain significant results, and would have given a more reliable estimation of effect sizes. Furthermore, a larger sample size would have allowed a full test of the model, instead of
Conclusion
In conclusion, people with ASD showed worse EF performance than the neurotypical control group in the attention and WM domains, whereas their level of inhibition was comparable to that of the control group. The young novice drivers with ASD in our study were pretty skilled drivers. Dependent on the driving measure, the driving performance of young novice drivers with ASD was considered worse, equal, or even better, compared to the control group. Importantly, relations between EF and driving
Funding
This research was funded by the Belgian Marguerite-Marie Delacroix support fund (GV/B-226).
Acknowledgments
We would like to thank Weixin Wang for programming the scenario, and Dirk Roox and Marc Geraerts for the technical support. Moreover, we would like to thank Peter Vermeulen for his support and advice regarding the interpretation of results related to context blindness.
Ethical approval
This study was approved by the ethical committees of Hasselt University and the Catholic University of Leuven (reference number ML10787).
References (86)
- et al.
Personality traits and executive functions related to on-road driving performance among older drivers
Accident Analysis & Prevention
(2011) - et al.
Executive function deficits and neural discordance in children with autism spectrum disorders
Clinical Neurophysiology
(2009) - et al.
Commentary training improves responsiveness to hazards in a driving simulator
Accident Analysis & Prevention
(2010) - et al.
Some hazards are more attractive than others: Drivers of varying experience respond differently to different types of hazard
Accident Analysis & Prevention
(2012) - et al.
The relations between specific measures of simulated driving ability and functional ability: New insights for assessment and training programs of older drivers
Transportation Research Part F: Traffic Psychology and Behaviour
(2016) - et al.
Mapping gray matter development: Implications for typical development and vulnerability to psychopathology
Brain and Cognition
(2010) - et al.
An examination of the relationship between measures of impulsivity and risky simulated driving amongst young drivers
Accident Analysis and Prevention
(2017) Evaluating the theory of executive dysfunction in autism
Developmental Review
(2004)- et al.
Task switching in autism: An EEG study on intentions and actions
Neuropsychologia
(2018) - et al.
Evidence for executive dysfunction in autism
Neuropsychologia
(1994)