Soccer is considered the most popular sport in the world (Reilly, 1997), with elite competitive matches requiring an officiating team [referee, assistants × 2 and within the Union of European Football Associations (UEFA) region 2 × goal line officials] to apply the laws of the game (Castagna et al., 2007). Football is characterized by variable intermittent activity and recovery, inherently displaying more complex physiological requirements than sports with continuous exercise (Drust et al., 2000). The physiological demands of refereeing are similar to that of a midfield soccer player, with total distance covered and high speed running consistent between player and referee (Weston et al., 2011).

Increased physical fatigue in a referee can elicit an inability to maintain proximity to key incidents within game play (varying rule infringements), with increases in the infringement/distance nexus, known to impair decision making ability (Catterall et al., 1993). Therefore, repeated sprint ability is of high importance for referees to maintain proximity to such incidents (Galanti et al., 2008). Compounding issues regarding proximity, high intensity exercise per se is known to have negative effects on cognition (McMorris and Graydon, 1997), with such cognitive decrements associated with altered arousal and narrowed attention (Brisswalter et al., 2002). Thus, the interplay between the physical demands of refereeing and decision making is important, yet, requires further elucidation.

The UEFA Champions League and The UEFA Europa League, played principally in Europe between September and May, ensures referees often perform their roles within varying challenging environmental conditions [Madrid (~30°C) and Moscow (~ –5°C)] within the UEFA region and season. Indeed, intermittent sprint exercise in hot conditions (exercise-heat stress) is known to increase thermoregulatory and physiological strain/load (Duffield et al., 2009), reducing physiological performance (i.e., distance covered), and likely further disturbs arousal levels (Bandelow et al., 2010), when compared to similar exercise in temperate conditions. Conversely, cold temperatures can cause severe discomfort and lead to a drop in body temperature (Parsons, 2003), which, may also effect cognitive performance due to distraction (Mäkinen et al., 2006). However, metabolic heat production during soccer specific exercise may help overcome the distraction and severe discomfort associated with cold temperatures (Brotherhood, 2008), although, whether this ameliorates the well-reported negative effects on cognitive function (Mäkinen et al., 2006; Racinais et al., 2008) has yet to be confirmed. Despite evidence that both heat (Racinais et al., 2008; Simmons et al., 2008) and cold stress (Mäkinen et al., 2006) detrimentally effect cognitive capacity and performance, this has not been studied in soccer referees, who are constantly required to make correct decisions on infringements and are regularly exposed to environmental extremes within the UEFA region and season, whilst, performing soccer specific exercise. Referee decisions can often affect the outcome of a match and thus it is important to understand the interaction between the ambient environment and the ability of soccer referees to make correct decisions.

Due to inherent difficulties, and likely confounding nature in the inability to standardize the external environment (both weather and games factors players, tactics, results, etc.) and physiological demands within an actual match the tripartite research questions of interest here (decision making, physiological demands/capacity and environmental fluctuations) would be best explored by a treadmill based, environmental chamber delivered protocol. The seminal work of Drust et al. (2000) is currently the most used protocol and will be utilized within the present study.

The aims of the present study were to investigate the effect of hot and cold conditions, typical of the UEFA region and season, on decision making and physiological responses of soccer referees during a treadmill based intermittent sprint protocol (Drust et al., 2000). It was hypothesized that hot temperatures would increase the fatigue that is associated with the exercise protocol and thus cause a detrimental effect on decision making. Cold conditions would have no significant effect on core temperature, but skin temperature (T_sk) would significantly decrease, resulting in a negligible effect on decision making ability.

The aim of this investigation was to examine whether environmental conditions typical of the UEFA region and season, could have an effect on decision making and physiological responses within participants during a soccer specific intermittent sprint protocol (Drust et al., 2000). The main finding was that HOT caused no detriment to decision making performance, thus rejecting the hypothesis that heat stress would cause decrement to referee cognitive performance. Similar results were observed for COLD as hypothesized.

The novel finding from the present study is that thermal challenges (HOT and COLD) do not negatively affect referees decision making ability—negative cognitive responses were not increased during exposure to hot conditions. Moreover, positive cognitive responses suffered no decrement during exercise in the heat. Similarly, no difference in core temperature was observed between conditions [CONT and HOT conditions (p = 0.33); COLD and HOT conditions (p = 1.00)]. A lack of difference in T_re between conditions may be attributable to a number of variables; it may be that a lower resting T_re prior to exercise in HOT may have masked any difference in T_re increase rate across conditions. Moreover, it may be that the exercise protocol was not strenuous enough to induce sufficient thermoregulatory strain in the hot condition. Finally, it may be HOT was only warm conditions, again providing insufficient thermoregulatory load. However, T_sk was significantly higher in HOT than CONT (p < 0.001) and COLD (p < 0.001).

The absence of significant differences in cognitive performance between conditions within the present study (see Figure 3) are likely attributable to the lack of significant differences in T_re observed [CONT (38.05 ± 0.08) vs. HOT (37.85 ± 0.06); p = 0.33; COLD (37.7 ± 0.09) vs. HOT (p = 1.00) (see Figure 2A)]. Exercise-heat stress induced increases in core temperature results in a reduction in cerebral blood flow (Fujii et al., 2008; Brothers et al., 2009; Hayashi et al., 2011), suppressed arousal (Nielsen et al., 2001) and reductions in cognitive capacity (Bandelow et al., 2010). Moreover, reductions in cerebral blood flow have been shown to correlate with exercise fatigue (Nybo and Nielsen, 2001; Thomas and Stephane, 2008). This interplay between core body temperature and cognitive processes (Nielsen et al., 2001; Nybo and Nielsen, 2001; Bandelow et al., 2010), which are not significantly disturbed within the utilized experimental design (likely due to the exercise protocol used (Drust et al., 2000), as discussed later in text), may provide explanation for the lack of variance in cognitive performance across and between conditions in the present study.

The utilized “Drust Protocol” (Drust et al., 2000) provides an instrument which is designed to mimic the movement patterns (various speeds and durations of movements and associated rest periods) of out-field soccer players. As soccer players and referees display similar types of movement (walk, jog, run, cruise and sprint) and cover similar distances [referees 9–13 km (Reilly and Gregson, 2006); players ~ 9–12 km (Di Salvo et al., 2013)] over the duration of a 90 min match, the Drust Protocol (Drust et al., 2000) was selected for the present study. This protocol is conducted on a motorized treadmill, using fixed speeds which are not individualized to each participant's physiological capacity, e.g., peak sprint/treadmill speed. Lack of individualization in physiological load, likely produced differential internal and external loads within subjects. The mean average HR across conditions in the present study (COLD: 138 ± 5; CONT: 146 ± 6; HOT: 151 ± 5), compared to 165 bpm from soccer match play (Catterall et al., 1993) and 167 bpm from an appropriately valid and reliable treadmill based soccer simulation (Aldous et al., 2013) respectively, suggests that physiological load in the present study was not indicative, thus not valid, in light of the values typically seen in soccer match play (Catterall et al., 1993; Castagna et al., 2007; Krustrup et al., 2009; Bradley and Noakes, 2013). Further concern is raised by the highest mean RPE recorded throughout HOT (30°C) in the present study being 15 (HARD), whereas repeated sprint protocols typically elicit RPE >18 (Drust et al., 2005).

Additionally, match analysis data (Reilly and Gregson, 2006) provides commentary on the distances covered by soccer match officials, the types of movement, and the difference in performance between the first and second half. It was reported that soccer referees from different national leagues (English Premier League & Danish SuperLiga) (Catterall et al., 1993; Krustrup and Bangsbo, 2001) as well as elites soccer players (Mohr et al., 2003; Bradley and Noakes, 2013), all covered less distance running at high intensities and less overall distance, during the second half of matches. As the Drust protocol utilizes fixed speeds on a motorized treadmill there is no allowance for this bi-phasic performance between halves, and thus not reflective of actual match play performance in soccer referees (Catterall et al., 1993; Krustrup and Bangsbo, 2001) or players (Mohr et al., 2003; Bradley and Noakes, 2013). The HOT condition final T_re (90 min) (38.4°C ± 0.4) was lower than that reported in soccer match play data for both professional soccer players (38.5°C) and recreational soccer players (39.0°C) (90 min); even though these results were collected in temperate conditions (16°C) (Edwards and Clark, 2006). This lack of appropriate specific physiological load (i.e., reflective of a referee/player during a game), could be pivotal in the lack of cognitive impairment seen in the present study, for example the core temperature cognition nexus previously described, which was hypothesized to occur within HOT. As the movement patterns of soccer referees and players are similar (Weston et al., 2011) such comparisons are possible, and demonstrate that the physiological load experienced during the Drust Protocol was not sufficient, nor valid when compared to match play data (Mohr et al., 2003; Barros et al., 2007; Bradley and Noakes, 2013).

A further point of interest that may explain the disparity in continuity between the present study and that of Drust et al. (2000) may be the difference in fitness between the two studies. The study by Drust et al. (2000) used university standard football players where as the present study used both semi-professional footballer players and referees. If those participants recruited for the present were fitter than those from Drust et al. (2000) due to their superior level of soccer participation it may be that the Drust protocol is valid for use on amateur soccer players but not semi-professional/professional. In the present study there were no significant differences in decision making ability during and between all conditions, likely due to the validity issues outlined above regarding the employed protocol (Drust et al., 2000).

A recently published protocol utilizing a non-motorized treadmill based intermittent soccer-specific performance test (iSPT) (Aldous et al., 2013) provides the individualized variance (e.g., reduction in distance covered between 1st and 2nd halves) observed in real-life match play in soccer players (Mohr et al., 2003; Barros et al., 2007; Bradley and Noakes, 2013) and referees (Catterall et al., 1993; Krustrup and Bangsbo, 2001), which was absent in the utilized protocol (Drust et al., 2000). Data from the iSPT displays reductions in total distance covered and distance covered at high intensities during the second 45 min of a 90 min protocol (Aldous et al., 2013), similar to the field data observed in elite referees (Krustrup and Bangsbo, 2001; Reilly and Gregson, 2006) and players (Mohr et al., 2003; Barros et al., 2007; Bradley and Noakes, 2013). All variables associated with internal and external load in response to iSPT (Aldous et al., 2013), were shown to be reliable, and valid in comparison to match play data. Therefore, it is recommended that the present experimental design be repeated with the protocol employed (Drust et al., 2000) replaced with iSPT (Aldous et al., 2013), or a similar appropriately constructed protocol (Williams et al., 2010).

The interaction between skin blood flow (T_sk) and T_re must not be overlooked. Alterations in skin blood flow may explain why T_re did not display noticeable change between conditions. It is plausible that increases in skin blood flow and thus increases in skin temperature provided sufficient thermoregulatory homeostasis via heat evaporation and radiation that T_re did not increase in the hot condition—particularly if either the hot condition was not hot enough, the exercise protocol was not strenuous enough or a combination of both. This provides explanation for the non-significant difference in T_re between conditions as well as the statistically significant difference in T_sk between conditions.

As hypothesized, there was no effect on cognition in the COLD condition. Whilst performing the exercise protocol, metabolic heat would be produced by the working muscles (Brotherhood, 2008) and this may serve to maintain deep body temperature (T_re). Whilst maintaining T_re, there will be a decline in temperature in the peripheral tissue (Tipton, 2006), which was observed in the present investigation, as T_sk was significantly lower in the COLD condition compared to both CONT and HOT. Though the experimental hypothesis was proven correct with regard to COLD, the formulation of the employed protocol (Drust et al., 2000), as discussed previously, may confound the validity of the observed data and its interpretation. Unpublished data from our laboratories suggest that the effects of extreme environmental conditions on decision making of non-exercising participants (e.g., goal line officials employed within the UEFA region) causes a reduction in cognitive performance in a similar COLD condition, with said performance unaffected in a warmer conditions, as observed in other previous research (Mäkinen et al., 2006). Evidently, the interaction of exercise, cognition and environmental stress is multifaceted, and requires further elucidation, specifically regarding environments of greater stress than those used within the present study.

Aside from the environment mediated challenge to cognition processes, it is possible that cognitive function may be impaired during by exercise per se (McMorris and Graydon, 1997). Previous research has found that neural activation associated with a task, whilst disturbed during task (e.g., exercise), rapidly returns to baseline levels after exercise (Magnié et al., 2000). Such in-task disturbance is likely due to a large part of the brain being associated with basic sensory processes and motor outputs, and thus, must come at the expense of other neural tasks (Dietrich and Sparling, 2004). Therefore, future work, utilizing an appropriate protocol as suggested (Williams et al., 2010; Aldous et al., 2013), should seek to assess decision making ability during the exercise task itself. To enhance ecological validity further the use of a soccer specific decision making tool should be considered for future research. The tool used in the present study (PSCHE Software) was developed for use in a clinical setting. And although it provides useable, quantitative data it has not been employed extensively in a sporting context. It may be that a soccer specific tool more sensitive than the PSYCHE software would highlight differences in decision making that were not observed in the present study.

In conclusion, intermittent sprint exercise in any condition did not have a significant effect on decision making ability of the participants. The exercise protocol utilized (Drust et al., 2000), as described, likely did not provide a truly indicative simulation of the internal and external load of soccer match play, and thus referee movement patterns. This lack of reflective internal and external loads resulted in the tripartite relationship between environmental UEFA region and season specific environmental stress, cognition processes and exercise not being securely and fully elucidated within the utilized experimental design. Future work should look to assess cognition during exercise/game play in conjunction with the use of an appropriately constructed protocol (e.g., Aldous et al., 2013) in a variety of environmental conditions to enhance ecological/external validity to not only the UEFA region, but that of Federation Internationale de Football Association (FIFA).

This was a Union of European Football Associations (UEFA) funded project.