In daily traffic, a bicyclist is more vulnerable and injured more often than a driver in a motor-driven vehicle in Sweden (Rizzi et al., 2020). Bicycling while intoxicated with alcohol is associated with an increased risk of sustaining severe or even fatal injuries; the risk increases with increases in blood alcohol concentration (BAC) level and is 20% higher with an alcohol level of 0.8 g/l and higher (Li et al., 2001). In general, there is greater risk of being involved in an accident when intoxicated with alcohol (Asbridge et al., 2014; Martínez-Ruiz et al., 2013).

In Europe, between the years 2010 and 2018, there were around 19 450 bicycle fatalities (Adminaité-Fodor & Jost, 2020), and every year in Sweden approximately 23 000 people seek care at an emergency hospital due to a bicycle-related incident (Schyllander & Ekman, 2013). The lack of reliable data makes it hard to establish how many of these fatalities and incidents involved alcohol. An increase of alcohol-related accidents is also related to the time of day (de Waard et al., 2016), and to gender, as males compared to females, are more likely to be injured when intoxicated (Airaksinen et al., 2018).

Drinking alcohol impairs bicycling (Hartung et al., 2015; Hartung et al., 2015; Kovacsova et al., 2016; van Lunteren et al., 1970). Although it is difficult to determine the exact impact that alcohol intoxication has on accidents, and the frequency of such events, it seems reasonable to assume that bicycling performance worsens with alcohol consumption (Andersson et al., 2023).

In Sweden, there is no alcohol limit for riding a bicycle, and this includes the increasingly popular power-driven bicycles. However, by law reckless behavior in traffic is forbidden. In this paper, we demonstrate the relation between intoxication by alcohol and bicycling performance. According to social norms, it is also acceptable in Sweden to bicycle home after a party or a night out with friends when intoxicated with alcohol (Wallén Warner et al., 2017). Furthermore, while bicycling is a legal way of transportation when intoxicated, private cars are not an option (the blood alcohol level limit for motor vehicle drivers is 0.2‰ in Sweden).

One aspect to consider when investigating bicycling performance during alcohol intoxication is how the bicycle is handled by the bicyclist, i.e. the stability of the individual when bicycling. The literature on bicycling stability (balance) reveals that speed is important (Cain et al., 2016). At lower speeds, novice and skilled bicyclists show similar balance performances, but at higher speeds, skilled bicyclists use more lean control and less steer control.Hartung et al. (2015) studied bicycling performance and found severe coordinative faults when participants bicycled under the influence of alcohol. In addition, they found differences between men and women in terms of the BAC level at which severe faults occurred.

The literature also suggests that there is an interesting and complex interplay between alcohol intoxication and cognitive performance (Cash et al., 2015; Greenstein et al., 2010; Mintzer, 2007; Park et al., 2011; Ralevski et al., 2012; Weiss & Marksteiner, 2007). Overall, the results suggest that cognitive functions are usually negatively affected by alcohol intoxication, but not always (Hoffman et al., 2015; Spinola et al., 2017).

Self-ratings of ability enhance our understanding of the more self-regulated aspects of bicycling while intoxicated. Bicycling ability is negatively affected by alcohol intoxication (Hartung et al., 2015) and by alcohol-induced hangover (Hartung et al., 2015) . It is clear that alcohol intoxication is associated with more severe injuries, increased hospital resources and a higher mortality rate (Sethi et al., 2016). Individuals also take riskier decisions when intoxicated (George et al., 2005).

In order to further explore how alcohol intoxication affects bicycling performance, several factors need to be examined. It is difficult to predict which and how different individual characteristics affect stability when intoxicated with alcohol as there is no existing research on the topic. Based on the understanding of the effects of i) familiarity on performance, ii) skill on the task, and iii) traits that are strongly related to how people perform on tasks in general, three relevant characteristics of bicycling during alcohol intoxication were chosen for the present study, i.e. bicycling experience, physical fitness and, sensation seeking personality. It is not evident from the literature whether these three characteristics play an active part in bicycling performance during alcohol intoxication.

The present study aimed to investigate if individual characteristics of bicycling experience, physical fitness, or sensation seeking scores, have an effect on bicycling stability, cognitive performance, or self-rated bicycling ability at different levels of alcohol intoxication. The exploratory approach was mainly driven by the aim to increase our understanding of how individual characteristics are related to bicycling during alcohol intoxication. This study is a follow up on the work byAndersson et al. (2023), where personal characteristics were not analyzed in the experiment. Hence, the experiment and therefore the method section is the same for the two articles, but data from the same experiment was analyzed with another perspective, i.e. with the focus on group characteristics.

A total of 29 people participated in the study. These individuals had completed the health declaration and did not show any signs of harmful alcohol use or dependence symptoms as measured by the Swedish version of the World Health Organization's AUDIT (Socialstyrelsen, 2022). Of these 29 participants, 18 were given alcohol during the experiment (intoxicated group), while 11 randomized participants completed the bicycling test sober (sober group).

The participants were recruited via a Facebook advertisement and personal contacts. The potential participants were directed to a dedicated VTI (Swedish National Road and Transport Research Institute) webpage, where the required criteria for participation (age gender, health, etc.) were checked as the first step of inclusion in the study. All the requirement criteria, which were approved by the ethics committee, were met by those who participated in the study.

The participant screening criteria included: a minimum age of 20 years old, no history of alcohol or substance addiction, good health for bicycling, and moderate alcohol consumption. The participant criteria also included living within the town of Falun (and Borlänge) (to limit costs of taxis, used to drive participants home after the experiment had ended), the sober participants did not have to live within the town of Falun. Only non-pregnant women could participate in the experiment (to avoid possible fetal damage associated with alcohol consumption) and women were offered a pregnancy test if required. Other requirements included previous experience of both bicycling and drinking alcohol.

The selection of participants included several steps, starting with the initial screening mentioned above. The screening procedure was repeated when each participant arrived at the test facility. The AUDIT scale was also completed and checked prior to testing. The participants in the intoxicated group needed to be able to stay onsite for approximately 4 hours, so that their post-experiment safety and sobriety could be monitored. The sober group were free to leave as soon as all post-test evaluations were completed.

The participants were reimbursed for their participation with approximately €150 for participation in the intoxicated group and €75 for participation in the sober group. All participants were also offered a cost-free taxi ride home.

This experiment utilized a between-participant design (intoxicated versus sober participants) to investigate the effects of alcohol intoxication on participants' bicycling performance. The characteristics of interest were: (i) bicycling experience, (ii) physical fitness and (iii) sensation seeking.

The independent variable was the repeated measurement variable of time, i.e. the breath alcohol concentration (BrAC) level was measured five times by the Dräger Alcotest® 6810 and 6820 devices for alcohol screening (see section 2.5 for details). The dependent variables were stability (yaw and roll rate), cognitive performance (n-back score) and self-rated bicycling ability. Speed and physical exertion might affect the results as well; the experimental set-up controlled for these aspects by keeping the speed and physical exertion constant over sessions.

The experiment consisted of a 25-minute loop that was repeated five times, whereby the participants bicycled on the treadmill in 10-minute stints, with a 15-minute non-bicycling period. Only 8 minutes of the 10 minutes of stability measures data were analyzed. The starting phase and the ending phase were controlled by the experiment leader; the experiment leader stabilized the participant by holding the gantry until the target speed was reached. The same procedure was used before the treadmill was stopped. The 8-minute period was free from any form of stability support. The 15-minute non-bicycling period comprised: (1) a BrAC assessment (four times, two times with both instruments); (2) an n-back test; (3) a subjective bicycling ability rating; and (4) time for refreshments and alcoholic beverages, if necessary, as well as a rest period.

Participants initially underwent a familiarization period of bicycling on the treadmill. Participants then performed in one experimental condition, when they all were sober. This was followed by four repetitions of the intoxication experimental condition, for those in the intoxicated group. The level of intoxication was increased throughout the experiment, up to a level of 0.8‰ BrAC. The BrAC level was chosen for pragmatical reasons.

The experiment was carried out in the sports laboratory at Lugnets Idrottsvetenskapligt Institut (LIVI), in co-operation with Dala Sports Academy, in Falun, during July and October/November 2019 and an addition of control participants in May 2022. The participants bicycled on a motor-driven treadmill (Saturn 450/300 RS; H/P/Cosmos Sports & Medical, Nussdorf, Germany), which had a surface area of 13.5 m² and could be tilted to increase physical workload (if necessary). The treadmill's general specifications comprised a running surface of 450 × 300 cm, a speed range of 0–40 km/h, an elevation of -5% to +25% (-2.8 to 14.0 degrees), a 30-kW motor system (40.8 HP), and a reinforced thick rubber running belt that is also suitable for use with ski rollers, ski poles, spike shoes, and bicycles.

The participants bicycled on a mid-range mountain bike (Biltema Yosemite X-dirt bicycle), with relatively large wheel dimensions for stability (27.5 × 2.8-inch wheels). The brakes were disconnected on the test bicycle, ready for use on the treadmill. The bicycle was equipped with a V-Box data acquisition system (Racelogic 3i) and gyro sensor (Racelogic IMU02). The participants wore a pulse sensor (to monitor their physical effort), which was connected to the treadmill system. The participants also wore a safety harness, which was attached to the gantry above the treadmill (Figure 1). The harness attachment point was connected to a kill-switch, to enhance safety. All participants wore a bicycle helmet.

The breathalyzer equipment employed in this study comprised two instruments: a Dräger 6810 (older model) and a Dräger 6820 (newer model). Both models comply with the standards required by the Swedish police (EU Police approval code: EN15964) and had valid calibration approval. The literature reveals that the Dräger 6810 is reliable, and no significant differences exist between BrAC and BAC if the instruments are used according to the manufacturer's instructions (Jurič et al., 2018).

The 3 (groups) by 5 (assessment times) mixed ANOVA on stability for roll (HBRR at 25 degrees) revealed a main effect and an interaction effect (Table 4). Pairwise comparisons showed that the sober group became more stable over time, and that both high and low bicycling experience groups became less stable (Table 3). There was no significant difference between low and high bicycling experience groups at any time. The 3 (groups) by 5 (assessment times) mixed ANOVA on stability for yaw (HBRR at 25 degrees) revealed a main effect of time and an interaction effect (Table 4). Pairwise comparisons showed that the sober group became more stable over time, and that both high and low bicycling experience groups became less stable (Figure 2). The 3 (groups) by 5 (assessment times) mixed ANOVA on the cognitive performance task revealed only a tendency of a main effect of time, indicating that cognitive performance improved overall, over time (Table 3). No differences between groups with high or low on bicycling experience. The 3 (groups) by 5 (assessment times) mixed ANOVA on self-rated bicycling abilities revealed a significant interaction effect. Pairwise comparisons showed that the sober participants' ratings of bicycling ability increased over time, and especially the group with low experience of bicycling ratings decreased over time (Table 4).

The correlation analyses (which included only intoxicated participants/groups at Time 1 and Time 5) revealed that bicycling experience scores did not correlate with any of the dependent measures (stability (roll and yaw), cognitive performance, or self-rated bicycling ability) at time 1 (when sober) or time 5 (when intoxicated).

Taken together (bicycling experience): results show that the stability of the sober group increased over time, and the stability decreased in both the high and low group on the measures of roll rate and yaw rate. The results showed that cognitive performance was relatively unaffected. The participants in the sober group self-rated bicycle ability ratings increased over time meanwhile the self-rated bicycle ability ratings decreased mainly in the low bicycling experience groups.

The analysis on the 3 (group) by 5 (assessment times) mixed ANOVA) on physical fitness and stability for roll (HBRR at 25 degrees) revealed an interaction effect (Table 4). Pairwise comparisons showed that the sober group became more stable over time, and that both high and low physical fitness groups became less stable (Table 3). Especially the group with low ratings on self-rated physical fitness became less stable over time from time 2 and onwards. The 3 (groups) by 5 (assessment times) mixed ANOVA on stability for yaw (HBRR at 25 degrees) revealed a main effect of time and an interaction effect. Pairwise comparisons showed that the sober group became more stable over time, and that both high and low physical experience groups became less stable. There was no significant difference between low and high physical fitness groups. The 3 (groups) by 5 (assessment times) mixed ANOVA on the cognitive performance task revealed a main effect of time, indicating that cognitive performance improved over time (Figure 3). The 3 (groups) by 5 (assessment times) mixed ANOVA on self-rated bicycling abilities revealed a significant main effect of group and an interaction effect. Pairwise comparisons revealed that the more physically fit participants' self-ratings of bicycling abilities were higher at all assessment times compared to the less physically fit participants, and even higher compared to the sober group at time 1 but not after that.

The correlation analyses (which included only intoxicated participants/groups at Time 1 and Time 5) revealed that physical fitness correlated with self-rated bicycling ability only at time 5, i.e. the participants who had high values on physical fitness rated themselves as having higher bicycling ability when intoxicated at time 5 (r = 0.61; p < 0.01), but not at time 1 (when sober).

Taken together (physical fitness): results show that the stability of the sober group increased, and the stability decreased in both the high and low physical fitness groups on the measure of roll and yaw (HHRR at 25 degrees). The results showed that the cognitive performance improved over time. The results from self-rated bicycle ability ratings showed a main effect of group and especially the low group on physical fitness rated a decreased ability. Finally, the physically fit participants rated their ability higher compared to the less physically fit participants but only when alcohol intoxicated (at time 5).

The analysis on stability for roll (HBRR at 25 degrees) revealed a main effect of time and an interaction effect i.e. an increase of stability for the sober group and a decrease in stability for especially the group with low on sensation seeking scores. The 3 (groups) by 5 (assessment times) mixed ANOVA on stability for yaw (HBRR at 25 degrees) revealed a main effect of time and an interaction effect. Pairwise comparisons revealed that the sober group became more stable over time compared to the low and high groups on sensation seeking that became more and more unstable. The decrease in stability for the intoxicated groups were larger than the increase in stability for the sober group. The 3 (groups) by 5 (assessment times) mixed ANOVA on the cognitive performance task revealed only a main effect of time, indicating that cognitive performance improved over time. Especially the sober group improved, and the intoxicated group's results were relatively unaffected. The 3 (groups) by 5 (assessment times) mixed ANOVA on self-rated bicycling abilities revealed a significant main effect of group and an interaction effect. Pairwise comparisons showed that the sober participants' self-rated bicycling abilities increased over time. Those who scored high for sensation seeking were relatively stable over time and the participants with low sensation seeking scores showed a decrease in ability ratings over time (Figure 4). The sober group were significantly different from the low sensation seeking scoring group (from time 3 and onwards).

The correlation analyses revealed that SSS scores correlated only with self-rated bicycling ability, i.e. the participants who had high sensation seeking scores rated themselves as having a higher bicycling ability when intoxicated at time 5 (r = 0.68; p < 0.01), but not at time 1 (when sober).

Taken together (sensation seeking scores): results show that the stability of the sober group increased, and the stability decreased in both the high and low group in the measure of yaw and the group with low scores on sensation seeking when roll was analyzed. The results showed a main effect of time and that the cognitive performance improved over time. The self-rated bicycling ability rating revealed a significant interaction effect where those who scored high on sensation seeking were relatively stable over time meanwhile the participants with low sensation seeking scores showed a decrease in ability ratings over time.

Five limitations of the experiment need to be addressed to avoid overemphasized conclusions. First, the task at hand was bicycling on a treadmill on an unfamiliar bicycle. This is not the same as bicycling on a road with your own bicycle. All participants were unfamiliar with the task, and stability was hence more difficult. The participants felt uncomfortable at the beginning of the task indicating that more practice sessions should be longer in future.

Second, the participants were divided into high and low groups for each characteristic posteriori. The basic idea was to investigate and collect data on characteristics that could be related to different aspects of bicycling on a treadmill when intoxicated with alcohol, but the groups were not chosen a priori, i.e. the participants were divided into high and low groups based only on their subjective statements after completion of the experiment. A more controlled selection of participants would of course be more appropriate, especially if participants had high values for some characteristics and low values for others.

Third, the learning effect was larger than anticipated. Future studies should attempt to reduce the learning curve to allow for enhancement of the effects of interest.

Fourth, the small group sizes could be discussed, even if it is argued that this did not affect the results in a substantial way. This limitation would be more problematic if the analyzes revealed tendencies for differences rather than clear significant differences, i.e. when it is unclear if a significant effect would be obtained with more participants. That was not the case here; although the number of participants was small, significant effects were obtained.

A fifth limitation is that the group size is rather small and homogenous and that the findings can't be generalized to bicyclists in general. Further research is needed.

Bicycling during alcohol intoxication needs to be studied in more realistic settings. This experiment has revealed several interesting results, but the interpretations should not be overemphasized in light of the exploratory approach and the limitations discussed above. From an experimental perspective, the lessons learned are also noteworthy. The participants' learning curves should be considered more carefully, even if the experiment is performed in a more realistic setting. The cognitive task chosen should be also compared to other cognitive tasks. This study suggests that cognitive performance was affected by alcohol intoxication overall. Self-rated bicycle ability ratings interacted with group characteristics, despite stability measures being similar for all intoxicated groups. This experiment does not suggest that the cognitive performance task used here affects self-regulated behavior, i.e. all intoxicated groups cognitive performance were negatively affected but self-ratings varied due to group characteristics. The question is however how self-rating of bicycle abilities is related to self-regulating behavior (Hittner & Swickert, 2006; Spinola et al., 2017). This finding needs to be investigated further.