Accident Analysis and Prevention 73 (2014) 236–241

Hostility, driving anger, and dangerous driving: The emerging role ofhemispheric preference

Yori Gidron a,*, Esma Gaygısız b, Timo Lajunen c

a Faculty of medicine & pharmacy, the Free University of Brussels—VUB, BelgiumbDept. of psychology, Middle Eastern Technical University, TurkeycDept. of psychology, Norwegian University of Science and Technology, Trondheim, Norway

A R T I C L E I N F O

Article history:Received 2 May 2011Received in revised form 8 August 2014Accepted 11 September 2014Available online xxx

Keywords:HostilityDriving angerDangerous drivingHemispheric preferenceCognitive exercises

A B S T R A C T

Background: Various studies have implicated psychosocial variables (e.g., hostility) in risk of dangerousdriving and traffic accidents. However, whether these variables are related to more basic neurobiologicalfactors, and whether such associations have implications for the modification of psychosocial risk factorsin the context of driving, have not been examined in depth. This study examined the relationshipbetween hemispheric preference (HP), hostility and self-reported dangerous driving, and the ability toaffect driving anger via hemisphere activating cognitive exercises (HACE).Methods: In Study 1, 254 Turkish students completed questionnaires of hostility, HP and driving behavior.In Study 2, we conducted a “proof of concept” experimental study, and tested effects of left, right andneutral HACE on driving anger, by exposing N = 650 Turkish students to written scenarios including eitherlogical (left hemisphere), visuo-spatial (right hemisphere) or “mild doses” of both types of contents(control).Results: In Study 1, left-HP was associated with higher hostility and with more dangerous driving, andhostility mediated the relationship between L-HP and reported driving behavior. In Study 2, only right-HACE led to immediate significant reductions in self-reported driving anger.Conclusions: Left-HP is related to hostility and to dangerous driving, and it may be possible to partlyreduce driving anger by right-HACE. Future studies must replicate these findings with objectivemeasures, more enduring interventions and longer follow-ups.

ã 2014 Published by Elsevier Ltd.

Contents lists available at ScienceDirect

Accident Analysis and Prevention

journa l homepage: www.e l sev ier .com/ locate /aap

1. Introduction

Traffic accidents (TA) are a complex multi-factorial publichealth problem, resulting from environmental, vehicle and humanfactors. Of the many human factors previously associated with TA,one, namely hostility, has received relatively a lot of attention.Hostility is the relative stable tendency to behave antagonistically,think cynically and feel anger across contexts (Barefoot, 1992).Hostility was found to correlate with self-reported TA anddangerous driving (e.g., Gidron et al., 2003). Furthermore, inprospective studies, hostility predicted TA and dangerous driving(e.g., Norris et al., 2000). These findings have importantimplications for TA, because of two reasons. First, they may serveto help identify in advance people at risk of dangerous driving andof TA, for closer monitoring and prevention. Second, though mainlybased on correlation designs, these findings propose to conduct

* Corresponding author.E-mail address: Yori.Gidron@vub.ac.be (Y. Gidron).

http://dx.doi.org/10.1016/j.aap.2014.09.0110001-4575/ã 2014 Published by Elsevier Ltd.

intervention trial studies aimed at reducing hostility to possiblyreduce TA. Such interventions have been developed for othercontexts (Cardiac patients; Gidron et al., 1999). Of greaterimportance, anger-management interventions have been devel-oped and shown to be effective specifically in high-anger drivers(Deffenbacher et al., 2002).

However, though statistically significant, the effect sizes ofcognitive-behavioral interventions on reducing anger or hostilityare often limited (e.g., Gidron et al., 1999). It is possible that morebasic neurobiological factors underlie or “drive” the relationshipbetween hostility and dangerous driving. Such factors could notonly help to explain the origin of such hostility and dangerousdriving in part, but may also guide additional, and possibly moreeffective neuro-scientifically-based forms of intervention toprevent dangerous driving. Past studies have associated hostilitywith excessive right hemisphere activity (e.g., Demaree andHarrison, 1997). Behaviorally, hostility, and particularly angerexpression can be divided into anger-out, the tendency towardsexplosive anger expression, and anger-in, the tendency to inhibitanger expression (Spielberger et al., 1985). In contrast to Demaree

Y. Gidron et al. / Accident Analysis and Prevention 73 (2014) 236–241 237

and Harrison (1997), one study found that left hemisphere activitywas associated with anger-out (Hewig et al., 2004). Anger-out maybe more relevant to dangerous driving since it can be manifestedexternally by observable behaviors including rude gestures,honking with the horn, and even aggressive acts towards drivers,all closely related to and possibly leading to dangerous driving.Such externally manifested hostility may also provoke otherdrivers, resulting in a possible vicious circle of hostile driving.However, it remains unknown whether left or right hemispherefunctions are related to dangerous driving, and whether hostility,the broader personality construct which includes anger andaggression, plays a role in this relationship. Hemispheric prefer-ence (HP) or cerebral asymmetry reflects the relatively stabletendency to activate or use more certain functions of onehemisphere versus the other (e.g., Davidson, 2004). Recently,left-HP was indeed discussed as related to anger/hostility (Hof-man, 2008). The present research examined the relationshipbetween HP, hostility and driving anger, and dangerous driving. InStudy 1, we examined the relationship between HP, hostility andself-reported dangerous driving behavior, beyond the effects of ageand gender. Following the finding in Study 1 that left-HP wasassociated with both hostility and dangerous driving, Study2 reflected a “proof of concept”, and examined the effects ofcognitive exercises aimed to minimally activate the righthemisphere in contrast to the left hemisphere, on driving anger.

2. Study 1

2.1. Method

2.1.1. Participants and procedureThe data were collected from 254 university student volunteers

in Ankara, Turkey (158 males and 96 females). Two-hundred andfour participants had a driving license. Only the driving licenseholders were included in the final analyses about hemisphericpreference, hostility and aberrant driving, whereas the wholesample was used for assessing the reliability of the instruments.The participants were assured for their anonymity and confidenti-ality. The participants filled out questionnaires assessing aberrantdriver behavior (Driver questionnaire by Reason et al., 1990), ascale assessing hemispheric preference (HP), the brief New–Busshostility scale (Gidron et al., 2001), and demographic variables, asdescribed below.

Among the drivers with a driving license, 68% were men, 32%were women. Participants had a mean age of 21.8 years (range 18–28 years, SD = 1.85 years). The mean lifetime mileage was9541.75 km (range 0–300,000 km, SD = 30,334.5 km) and meandriving experience time was 2.6 years (range 0–7 years,SD = 1.5 years). The average annual mileage was 3094.6 km (range0–100,000 km, SD = 9168.0 km).

2.1.2. Measures

2.1.2.1. Demographic measures. Respondents answered questionsabout their age, gender, student status, the number of years a fulldriving license had been held, lifetime mileage driven, and annualmileage.

2.1.2.2. Driver behavior questionnaire (DBQ). The Driver behaviorquestionnaire (DBQ) with extended violations was used tomeasure self-reported violations and errors (Reason et al., 1990).Violations were defined as “deliberate deviations from thosepractices believed necessary to maintain the safe operation of apotentially hazardous system.” Errors can be defined as a “failure ofplanned actions to achieve their intended consequences that caninvolve the unwitting deviation of action from intention (slips and

lapses) or departure of planned actions from some satisfactorypath toward desired goals (mistakes).” Hence, the violations arecommitted deliberately but errors are done involuntarily.Moreover, violations and mistakes are potentially risky whereasslips and lapses are harmless.

The extended version of the DBQ (Lawton et al., 1997) used inthis study included aggressive violations (3 items; e.g., to soundyour horn to indicate your annoyance), ordinary violations (8items; e.g., pull out of junction so far that the driver with right ofway has to stop and let you pass), mistakes (8 items; e.g.,misjudging the speed of another vehicle when overtaking), andlapses (8 items; e.g., forget where you left your car in a car park).Participants were asked to indicate how often they committedeach of the 28 behaviors in the previous year on a six-point scale(0 = never, 5 = very often). The Turkish translation and the factorstructure of the DBQ had been validated in studies conductedamong both professional and non-professional drivers (see Özkanet al., 2006). In the present sample, the alpha reliabilities of the“slips and lapses”, “mistakes”, “ordinary violations” and “aggres-sive violations” were 0.86, 0.91, 0.89, and 0.75, respectively.

2.1.2.3. Hemispheric preference (HP) scale. The hemisphericpreference (HP) scale (Wegner and Wells, 1985) was used toassess HP. This is a valid 12-item test, for which each item has fourresponse options: two reflect mostly left-HP functions(verbal, logical), and two reflect mostly right-HP functions(visual, creative). The left-HP index is computed by subtractingthe total number of right-HP responses from the total left-HP ones,dividing this fraction by their total, and multiplying it by 100. Thus,higher scores on the HP index reflect left-HP.

2.1.2.4. The brief New–Buss hostility scale. The brief New–Busshostility scale (Gidron et al., 2001) was translated to Turkish andused as a measure of trait hostility. The New–Buss hostility scaleincludes 8 items to which participants reply by using a 5-pointscale (1 = extremely uncharacteristic of me, 5 = extremelycharacteristic of me). The original aggression questionnaire byBuss and Perry (1992), from which the New–Buss was derived, hasbeen translated to Turkish and used in Turkey before (see Güleçet al., 2008). Moreover, Gidron et al. (2001) reported that the New–

Buss hostility scale was cross-culturally valid against multiplecriteria including deviant driving, other measures of anger/hostility and atherosclerosis in men. In the present study, theinternal alpha reliability coefficient of the scale was 0.75.

2.1.2.5. Statistical analysis. First, we tested correlations betweenall background measures and the main study variables. Second, wetested whether HP and hostility correlated with each other andwith dangerous driving using Pearson correlations. Finally, wetested a mediation model, by examining whether after statisticallycontrolling for hostility, HP still correlated with driving behavior,using analyses of covariance since HP was dichotomized at themedian. The latter was done, to increase interpretability of theresults, due to hypothesized differences between left and right-HPparticipants (Davidson, 2004).

3. Results

Groups of HP were created by splitting the HP scores at themedian. No age differences were found between participants withright-HP (M = 25.02, SD = 9.01 years) and those with left-HP (24.62,SD = 8.73 years), t(160) = .27, p > .05. Similarly, the distribution ofmen and women was not significantly different between right-HPand left-HP participants (x2(1) = .23, p > .05).

Table 1 depicts the means (SD) of participants on hostility andDBQ subscales, according to their HP group. As shown in Table 1,

Table 1Means and standard deviations (SD) of participants classified by hemisphericpreference (HP) on hostility and driving behavior subscales.

Variable Right-HP Left-HP

Mean SD Mean SD

Hostility 21.49 5.56 22.86* 5.55DBQ-aggression 1.44 1.19 1.77* 1.15DBQ-violations 1.42 1.00 1.65* 1.07DBQ-lapses 1.13 0.76 1.47* 0.98DBQ-mistakes 1.03 0.78 1.46* 1.10

Note: HP: hemispheric preference; DBQ: driving behavior questionnaire; *p < .05;(all tests are 2-tailed).

238 Y. Gidron et al. / Accident Analysis and Prevention 73 (2014) 236–241

left-HP participants scored significantly higher on hostility thanright-HP participants (t(163) = 2.97, p = .003). Concerning DBQsubscales, left-HP participants scored significantly higher onaggressive violations (t(149) = 2.02, p = .045), ordinary violations(t(149) = 2.26, p = .025), lapses (t(149) = 2.15, p = .033), and ondriving mistakes (t(96) = 2.47, p = .015) than right-HP participants.Fig. 1 depicts the relationship between HP and each of the DBQsubscales.

Hostility was, as expected, positively correlated with aggressiveviolations (r =0.215, p = .008), ordinary violations (r = .246, p = .002),driving lapses (r = .333, p < .001) and with driving mistakes(r = .326, p < .001).

We then tested whether hostility mediated the relationshipobserved between HP and DBQ subscales. Using a series of analysesof covariance, controlling for hostility, HP was no longersignificantly associated with aggressive violations (p = .147),ordinary violations (p = .107), driving lapses (p = .206) and onlytended to be associated with driving mistakes (p = .075). Incontrast, hostility remained significantly related to each DBQsubscale, when HP was statistically controlled for (p levels rangingfrom 0.024 to 0.000). Thus, hostility mediated the HP-DBQ

Fig. 2. The relationship between hemispheric pre

Fig. 1. Relationship between hemispheric preference (HP) and driver behaviorquestionnaire (DBQ) subscales.

relationships (see Fig. 2), while HP was not a mediator betweenhostility and DBQ subscales.

4. Discussion

This study may be among the first to begin to identify the neuralfactors possibly underlying the epidemiological relationshipsbetween psychological risk factors and dangerous driving. Weexamined the relationship between hemispheric preference (HP)and hostility, with self-reported dangerous driving, using validmeasures. We observed a positive correlation between hostility andself-reported dangerous driving behavior, in line with previousstudies (Gidron et al., 2003; Norris et al., 2000). Furthermore, left-HPwas positively correlated with both hostility and with self-reporteddangerous driving behavior. Importantly, hostility emerged as amediator in the relationship between left-HP and self-reporteddangerous driving behavior. A final series of analyses ruled out thepossibility that left-HP mediated the hostility–dangerous drivingrelationship. The mediating role of hostility is depicted as a model inFig. 2. This could mean that the possible effects of hostility on drivingstem froma more basic neuropsychologicalprocess of left-HP, whosepossible effects on driving operate via hostility.

Past studies on the relationship between anger and hostility andHP or hemispheric lateralization have been inconsistent. Somestudies found hostility to be related to right-HP (Demaree andHarrison, 1997) while others found hostility to be related with left-HP (e.g., Hewig et al., 2004). This may stem from the fact that hostilityand anger include an experiential component of negative affect,possibly linked to behavioral inhibition and to the right hemisphere,and a different behavioral component of antagonism (and anger-out)possibly linked to approach behavior and to the left hemisphere(Davidson, 2004; Costa et al., 1989). Indeed, one can view the righthemisphere as reflecting behavioral inhibition and the lefthemisphere as reflecting behavioral activation. In line with suchclassification, a more recent review proposes that factors such asanger (and perhaps hostility), as well as risk taking and positiveaffect, which reflect “approach oriented” and interpersonal activi-ties, are more mapped onto the left hemisphere. In contrast, anxietyand depression, which reflect more intra-personal inhibition orwithdrawal, are more mapped onto the right hemisphere(Hofman, 2008). The fact that also dangerous driving, which maybe manifested by “aggressive approach” behaviors between drivers,correlated with left-HP, supports and extends the validity of ourfinding that hostility is related to left-HP.

However, the main limitation of this study was its use of a cross-sectional correlation design. These findings need to be replicated ina prospective or experimental design, to verify whether HP andhostility predict and cause dangerous driving. An experimentalapproach in studying this topic would also help to clarify whetherHP has a causal role in driving hostility/anger and in itsmodulation, and this was the aim of Study 2.

ference (HP), hostility and dangerous driving.

Y. Gidron et al. / Accident Analysis and Prevention 73 (2014) 236–241 239

5. Study 2

Study 1 found that left-HP was related to hostility and to self-reported dangerous driving. Furthermore, hostility mediated therelationship between left-HP and self-reported dangerous driving.Given our results in Study 1 showing an association betweenhostility and left-HP, and since hostility includes approachoriented behaviors, hostility can be expected to be related to theleft hemisphere’s functions or activity, and partly be under itscontrol (Hewig et al., 2004; Hofman, 2008). Due to the phenome-non of inter-hemispheric inhibition (e.g., Sullivan, 2004), it ispossible that by activating the right hemisphere, levels of hostilityand driving related anger (possibly mediated by the lefthemisphere) may decline. Activating a hemisphere could be doneby electrical stimulation using repetitive transcranial magneticstimulation (rTMS). However, while relatively accurate and clearlystimulating the brain, such an approach is costly and logisticallynot feasible as a possible en mass intervention, given theprevalence of road hostility and dangerous driving behavior. Suchan approach may also carry certain health and ethical constraintsconcerning its safety and control over one’s behavior. Alternatively,exposing people to cognitive exercises was also found to activate ahemisphere as measured by an EEG, though this depended on thequality of individual performance (Papousek and Schulter, 2004).Furthermore, in a randomized controlled trial in early schizophre-nia, cognitive exercise therapy led to less loss of gray matter invarious cortical regions and this mediated the effects of suchexercises on improved cognition (Eack et al., 2010). Finally,Nombela et al. (2011) found that cognitive training for patientswith Parkinson’s Disease led to brain activity patters during thestroop test which resembled more healthy controls’ brain activitypatterns. Together, these studies show that cognitive exercises canactually yield important changes in the brain. Performinghemisphere-specific cognitive exercises is not at all costly and isunder people's full control. To preliminarily test this possibility anda causal relationship between HP and driving-anger, Study2 investigated the feasibility of exposing people to readingmaterial that aimed to activate one of the hemispheres (left,right) or both (control), and its subsequent effects on driving angerlevels. This preliminary study constituted merely a “proof ofconcept” study. Since we observed in Study 1 that left-HP waspositively associated with hostility, it was expected that anexercise aimed at right-hemisphere activation may reduce drivinganger compared to other conditions, via inter-hemisphericinhibition (Sullivan, 2004).

6. Methods

6.1. Participants

The sample initially included 657 students of various under-graduate programs at the Middle East Technical University inAnkara, Turkey. Student volunteers were recruited in classrooms,dormitories and in the campus in general. Participants (except thestudents helping in data collection) did not receive any financialreward. The only criteria for participating were: (1) willingness totake part, (2) a valid driving license and (3) student status. Weexcluded those without a driving license (N = 7). This resulted witha sample of 650 students, of whom 62.6% were male, 37.4% werefemale. Participants had a mean age of 28.8 (range 18–64 years,SD = 10.56 years). The mean estimated lifetime mileage was74,785.4 km (range 0–5,000,000 km, SD = 300,533.8 km) and meandriving experience time was 7.5 years (range 0–38, SD = 8.2 years).The average estimated annual mileage was 7124.8 km(range 0–150,000 km, SD = 14,713.1 km).

6.2. Measures

6.2.1. BackgroundThese included participants’ age, gender, dominant hand for

writing, as well as estimated lifetime and annual millage (in km)and years of driving experience. We included here the measure ofHP used in Study 1 (Wagner and Wells, 1985), as a covariatevariable (see below).

6.2.2. Driving angerThis was assessed with the Driving Anger Scale (DAS;

Deffenbacher et al., 1994), which originally includes 33 drivingrelated situations, initially derived from 53 such situations. In itsdevelopment, it showed high internal reliability (Cronbach’sAlpha = .90). In the present study, we derived from it threesubscales based on a study by Lajunen et al. (1998), which includedanger from others’ Reckless Driving (Cronbach’s Alpha = .91), DirectHostility (Cronbach’s Alpha = .83), and anger when Progress isImpeded (Cronbach’s Alpha = .84). The overall Cronbach’s Alpha ofthese three subscales in the present study was also high (0.91).

6.2.3. InterventionsThese interventions formed a “proof of concept” or feasibility

study, to examine whether it is possible to develop hemisphere-activating cognitive exercises (HACE) and whether they have anypreliminary effects on driving anger. Participants were randomlyassigned to one of three groups of HACE: left-HACE, right HACE or aneutral (control) exercise. In all three conditions, participants readone paragraph, always about the same topic and of approximatelythe same length (10 lines). The topic was about lakes. For the left-HACE, participants read a paragraph describing lakes from a logicaland analytic perspective: how they are formed, their shape and whatthey serve for people and animals. For the right-HACE, the paragraphdescribed lakes from a detailed visuo-spatial perspective: their form,the scenery, the water, birds flying around them, etc. Theseinstructions followed known differences between hemisphericfunctions, where the left one mediates more analytic thinking whilethe right one mediates more intuitive and visuo-spatial processes(e.g., Goel et al., 2007). Finally, the “hemispherically neutral” controlcondition included information about a lake without too manyanalytic/logical or too many visuo-spatial details, and represented abalance between the left- and right-HACE conditions.

6.2.4. Design and procedureThis study used a mixed experimental design, where Time (pre,

post intervention measure) was a within-subjects factor and Group(left-HACE, right-HACE or control) was a between subjects factor.Students entered the class, and were informed that this study ispart of their learning experience, but that their participation istotally voluntary. They then completed the baseline DAS andgeneral background information and were randomly assigned toeither the left-HACE, the right-HACE or neutral condition. Afterreading their respective paragraph (‘intervention’), they completedagain the DAS, as well as the HP test, used as a covariate.

6.2.5. Statistical analysisFirst, we examined group equality at baseline, using an analysis

of variance (ANOVA) for continuous data (e.g., age) and chi-squaretests for categorical data (e.g., gender). Second, we used a mixed-design ANOVA, where Time (pre, post intervention) was thewithin-subjects factor and Group (left-HACE, right-HACE orcontrol) was the between-subjects factor. Our main interest wasthe Time � Group interaction which was followed by simple-effects tests. In the main ANOVA, we controlled for annual millageand sex, and retested it with adding scores of HP as an additionalcovariate, to control for its effects.

Table 2Means and standard deviations (SD) of hemisphere-activatingcognitive exercise (HACE) groups on total driving anger scale (DAS)scores.

Group Mean SD

Left-HACEPre 3.30 .68Post 3.20 .78

Right-HACEPre 3.22 .70post 3.02* .82

ControlPre 3.33 .63Post 3.20 .73

SD: standard deviation; HACE: hemisphere-activating cognitiveexercises; *p < .05.

240 Y. Gidron et al. / Accident Analysis and Prevention 73 (2014) 236–241

7. Results

7.1. Equality of groups at baseline

First, no statistically significant differences were found oncontinuous scores between groups (for all ANOVAS, p > .05).Similarly, no group differences were found in distribution ofgender and of left/right handed participants (all ps > .05). Finally,left and right handed participants did not differ on DAS scores or onHP scores (all ps > .05). Hence, effects of these variables were notstatistically controlled for in the following analyses. Nevertheless,since annual mileage (reflecting road exposure) and gender usuallyhave a strong influence on driver behavior, we added them ascovariates, to be more rigorous.

7.2. Effects of hemisphere-activating cognitive exercises (HACE) ondriving anger

Using a mixed-design ANOVA, controlling for sex and annualmillage, we found a trend for a significant Time � Groupinteraction in relation to DAS scores (F(2,602) = 2.44, p = .088).When adding participants’ score on hemispheric preference (HP)as an additional covariate, this interaction became significant (F(2,579) = 3.59, p = .028). We then followed this interaction by asimple-effects analysis, which revealed that in the left-HACEgroup, Time had no effect (F(1,196) = 0, NS), in the control group,Time had no effect (F(1,198) = .82, p > .05), while only in the right-HACE group, Time had a statistically significant effect on drivinganger, controlling for sex and annual millage (F(1,204) = 3.99,p < .05). A similar pattern was seen when adding HP as anothercovariate. Table 2 depicts the means and SDs of DAS scores beforeand after the interventions and Fig. 3 shows this graphically.

We then examined which of the three DAS subscales wasaffected by the right-HACE. The Time � Group interaction onlytended to be significant for the DAS subscale of being irritated bySlow Driving (F(2,602) = 2.48, p = .084). In contrast, the Time �Group interaction was not significant for Reckless Driving (F(2,601) = 1.64, p = .195) or for Direct Hostile Behavior (F(2,598) = .947, p = .388). Indeed, concerning the Slow Drivingsubscale, only in the right-HACE group, did scores decrease overtime (p = .05), but not in the left-HACE group (p = .72) or in controls(p = .27). These results became more prominent when adding HP asa covariate.

8. Discussion

This preliminary study examined the feasibility to influencedrivers’ anger levels, by using an intervention based on aneuroscientific rationale. We show here that right-HACE led to

Fig. 3. Effects of hemisphere-activating cogn

statistically significant reductions in DAS scores, while left-HACEand neutral exercises did not. These results were limited to oneDAS-subscale, that of facing a slow driving vehicle. These findingsare partly in line with our findings in Study 1 and support ourhypothesis that right-HACE will reduce driving anger, possibly dueto inter-hemispheric inhibition of the left hemisphere (e.g.,Sullivan, 2004). However, these results only partly support thoseof Study 1, since if indeed the left hemisphere accounts for andmodulates anger and hostility, then the left-HACE group wouldhave reported temporary increases in DAS scores, but this did notoccur. Furthermore, these findings must be taken with greatcaution due to the use of very minimal HACE interventions, andsince assessment of driving anger was done only seconds after theintervention, with no follow-up. We shall now discus in greaterdetail the theoretical and practical applications of our findingsfrom both studies, and their limitations.

9. General Discussion

The two studies presented here investigated the role of HP inhostility/anger, a psychological risk factor of dangerous driving andTA. Study 1 showed that left-HP was related to both hostility and toself-reported dangerous driving, while Study 2 shows that it maybe partly possible to reduce driving anger by exposure to acognitive exercise (visuo-spatial description) which may activatethe right hemisphere. The fact that hostility emerged in Study 1 asa possible mediator in the relationship between left-HP anddangerous driving has several theoretical and preventive impli-cations. First, these findings suggest that hostility may be linked toa much more basic neural factor, namely left-HP, whose effects areeven manifested early in life (Davidson, 2004; Davidson and Fox1989). Second, these findings also inform us that the role ofhostility in driving may partly originate from such neural HP

itive exercises (HACE) on driving anger.

Y. Gidron et al. / Accident Analysis and Prevention 73 (2014) 236–241 241

tendencies. The fact that hostility mediated the relationshipbetween left-HP and self-reported dangerous driving behaviorsuggests that left-HP may be manifested partly by antagonisticbehaviors (e.g., general rudeness, non-cooperativeness) andcynical cognitions (e.g., “others are out there to get me”), possiblyleading to dangerous driving behavior (e.g., dangerous speeding,overtaking unsafely).

If replicated, these findings have implications for the identifi-cation and for possible prevention of dangerous driving behaviorby affecting psychological risk factors via neuro-modulation. First,measures of HP may be less prone to influences of socialdesirability than those of hostility since the former may be less“contaminated” by social values and judgment than measures ofhostility. In addition, HP can also be measured by neuropsycho-logical tests (the line bisection test) or by an electroencephalogram(Davidson, 2004), which are not self-report measures. Thus,pending replication of our findings with other measures of HP orhemispheric lateralization, it may be more beneficial to screen forHP rather than for hostility, to identify potentially dangerousdrivers (using both measures may even be better). Second, ourfindings have possible implications for neuro-modulation ofdangerous driving. Study 2 aimed to preliminarily examinewhether simple cognitive exercises, aimed at activating one ofthe hemispheres, could alter driving anger. We found that reading aparagraph containing visuo-spatial content, aimed at activatingthe right hemisphere, led to significant immediate reductions indriving anger, particularly towards slow drivers. These findingswere unaffected by drivers’ sex, annual millage or their own HP.These findings support those of Papousek and Schulter (2004),(Eack et al., 2010) and Nombela et al. (2011) showing that cognitiveexercises actually increase hemispheric activity or alter brainstructures, but raise multiple questions. First, can neuro-modula-tion, via repetitive transcranial magnetic stimulation (rTMS) or viaHACE, alter both psychological risk factors (e.g., hostility) andactual dangerous driving? Second, if replicated, what types anddoses of right-HACE are needed, and how long do their effectsremain? Future studies need to address these many issues. Oncethese questions are answered and pending replication of ourpreliminary results, performing daily HACE has no costs, and canbe done by drivers, empowering them and helping people take partin accident reduction. This could have serious implications forpublic health.

The two studies obviously have several limitations. First, weincluded only self-reported measures of HP and driving behavior.HP can be measured with more objective and still inexpensivemeasures including the line bisection method (e.g., van Vugt et al.,2000). Driving behavior may either be observed (e.g., by aninvestigator present in a participant’s vehicle), may be measuredby a simulator, or by a “black box” inside the vehicle and recordingdriving behaviors. Second, Study 1 was cross-sectional and itsfindings need to be replicated using a prospective design, and testwhether HP predicts dangerous driving and even TA with objectivemeasures of driving. Third, Study 2 used a very minimal HACEintervention, without verifying the neuronal effects of theinterventions, and with no follow-up. Thus, this requires replica-tion using longer HACE interventions and follow-ups, and objectivemeasures of neuronal activation and driving.

Nevertheless, both studies examined a novel neuroscientificapproach to understanding the role of psychological risk factors inTA, and Study 2 used an experimental design to test suchunderlying processes. Should these findings be replicated usingother methodologies, these would support the validity of the

hypothesis that left-HP underlies hostile driving and that byundergoing right-HACE, driving anger and dangerous drivingbehavior may partly be decreased.

References

Özkan, T., Lajunen, T., Chliaoutakis, J., Parker, D., Summala, H., 2006. Cross-culturaldifferences in driving behaviour: a comparison of six countries. Transport. Res.Part F: Traffic Psychol. Behav. 9, 227–242.

Barefoot, J.C., 1992. Developments in the measurement of hostility. In: Friedman, H.S. (Ed.), Hostility, Coping, and Health. American Psychological Association,Washington, DC, pp. 13–31.

Buss, A.H., Perry, M., 1992. The aggression questionnaire. J. Pers. Soc. Psychol. 63,452–459.

Costa, P.T., McCrea, R.R., Dembroski, T.M., 1989. Agreeableness versus antagonism:explication of a potential risk factor for CHD. In: Siegman, A.W., Dembroski, T.M.(Eds.), In Search of Coronary-Prone Behavior. Lawrence Erlbaum, Hillsdale, NJ,pp. 41–63.

Davidson, R.J., Fox, N.A., 1989. Frontal brain asymmetry predicts infants’ response tomaternal separation. J. Abnorm. Psychol. 98, 127–131.

Davidson, R.J., 2004. Well-being and affective style: neural substrates andbiobehavioural correlates. Philos. Trans. R Soc. Lond. B Biol. Sci. 359, 1395–1411.

Deffenbacher, J.L., Oetting, E.R., Lynch, R.S., 1994. Development of a driving angerscale. Psychol. Rep. 74, 83–91.

Deffenbacher, J.L., Filetti, L.B., Lynch, R.S., Dahlen, E.R., Oetting, E.R., 2002. Cognitive-behavioral treatment of high anger drivers. Behav. Res. Ther. 40, 895–910.

Demaree, H.A., Harrison, D.W., 1997. Physiological and neuropsychologicalcorrelates of hostility. Neuropsychologia 35 (10), 1405–1411 1997 Oct..

Eack, S.M., Hogarty, G.E., Cho, R.Y., Prasad, K.M., Greenwald, D.P., Hogarty, S.S.,Keshavan, M.S., 2010. Neuroprotective effects of cognitive enhancementtherapy against gray matter loss in early schizophrenia: results from a 2-yearrandomized controlled trial. Arch. Gen. Psychiatry 67, 674–682.

Güleç, H., Tamam, L., Yazıcı, M., Güleç, M.Y., Turhan, M., Karakufl, G., Stanford, M.S.,2008. Psychometric properties of the Turkish version of the BarrattImpulsiveness Scale-11. Klinik Psikofarmakol. Bül. 4, 251–258.

Gidron, Y., Davidson, K., Bata, I., 1999. The short-term effects of a hostility reductionintervention on male coronary heart disease patients. Health Psychol. 18, 416–420.

Gidron, Y., Davidson, K., Ilia, R., 2001. Development and cross-cultural and clinicalvalidation of a brief comprehensive scale for assessing hostility in medicalsettings. J. Behav. Med. 24, 1–15.

Gidron, Y., Gal, R., Syna Desevilya, H., 2003. Internal locus of control moderates theeffects of road-hostility on recalled driving behavior. Transport. Res. 6 (Part F),109–116.

Goel, V., Tierney, M., Sheesley, L., Bartolo, A., Vartanian, O., Grafman, J., 2007.Hemispheric specialization in human prefrontal cortex for resolving certain anduncertain inferences. Cereb. Cortex 17, 2245–2250.

Hewig, J., Hagemann, D., Seifert, J., Naumann, E., Bartussek, D., 2004. On the selectiverelation of frontal cortical asymmetry and anger-out versus anger-control. J.Pers. Soc. Psychol. 87, 926–939.

Hofman, D., 2008. The frontal laterality of emotion: a historical overview.Netherlands J. Psychol. 64, 112–118.

Lajunen, T., Parker, D., Stradling, S.G., 1998. Dimensions of driver anger, aggressiveand highway code violations and their mediation by safety orientation in UKdrivers. Transport. Res. Part F Traffic Psychol. Behav. 1, 107–121.

Lawton, R., Parker, D., Manstead, A.S.R., Stradling, S., 1997. The role of affect inpredicting social behaviours: the case of road traffic violations. J. Appl. Soc.Psychol. 27, 1258–1276.

Nombela, C., Bustillo, P.J., Castell, P.F., Sanchez, L., Medina, V., Herrero, M.T., 2011.Cognitive rehabilitation in Parkinson’s disease: evidence from neuroimaging.Front. Neurol. 2, 82.

Norris, F.H., Matthews, B.A., Riad, J.K., 2000. Characterological, situational, andbehavioral risk factors for motor vehicle accidents: a prospective examination.Accid. Anal. Prev. 32, 505–515.

Papousek, I., Schulter, G., 2004. Manipulation of frontal brain asymmetry bycognitive tasks. Brain Cogn. 54, 43–51.

Reason, J., Manstead, A., Stradling, S., Baxter, J., Campbell, K., 1990. Errors andviolations on the roads. Ergonomics 33, 1315–1332.

Spielberger, C.D., Johnson, E.H., Russel, S.F., Crane, R.J., Jacobs, G.A., Worden, T.J.,1985. The experience and expression of anger: construction and validation of ananger expression scale. In: Chesney, M.A., Rosneman, R.H. (Eds.), Anger andHostility in Cardiovascular and Behavioral Disorders. Hemisphere/McGrath Hill,New York, pp. 5–30.

Sullivan, R.M., 2004. Hemispheric asymmetry in stress processing in rat prefrontalcortex and the role of mesocortical dopamine. Stress 7, 131–143.

van Vugt, P., Fransen, I., Creten, W., Paquier, P., 2000. Line bisection performances of650 normal children. Neuropsychologia. 38 (6), 886–895.

Wagner, R.F., Wells, K.A., 1985. A refined neurobehavioral inventory of hemisphericpreference. J. Clin. Psychol. 41, 671–676.