The effects of hypohydration and thirst on cognitive performance and mental fatigue

A popular belief within exercise and sports science is that hypohydration and its associated perceptions negatively influence cognitive performance. Such detriment is thought to manifest when losses in body mass (BM) are approximately 2%, however, several recent publications report that most cognitive domains are preserved when this threshold is exceeded. Given these equivocal findings, this thesis explores the relationship between both hydration and thirst, and resultant cognitive function. This will be achieved by highlighting the conceptual intricacies concerning the evaluation of cognitive performance and hypohydration, and exploring the prospective mechanisms that may be enacted to preserve cognitive capability when this physiological stressor is encountered. As such, this thesis aims to i) determine whether hydration status is important for cognitive performance, specifically the appropriateness of recommendations to avoid an “excessive” (2%) loss in BM, ii) examine compensatory (neuro)physiological responses that may arise during hypohydration, and iii) investigate the existence of additional associations implicated with hypohydration and cognitive function. Prior to empirically investigating the hypotheses, it was important to first ascertain the current state of the literature concerning hypohydration and cognitive function. Thus, a meta-analysis was completed to determine whether this physiological stressor is likely to be of consequence to cognitive performance. Additionally, because current fluid recommendations encourage one to avoid moderate/ excessive dehydration (≥2% loss in BM), on the provision that such deficits might impair cognitive performance, this theoretical critical water deficit was also investigated. A systematic search revealed 10 studies (14 trials), where cognitive performance data were extracted and aligned to respective cognitive domains (complex attention, executive function, learning and memory, and perceptual-motor function). Overall, cognitive performance was not found to be impaired by hypohydration (g = -0.177; 95% CI = -0.532-0.179; P = 0.33). There was also no significant impairment of the underlying cognitive domains that were able to be examined (complex attention, executive function, learning and memory) (all P > 0.24), independent of the incurred fluid loss (less than or greater than 2% loss in BM); although results were not always homogenous (I2 ranging between 0% and 93%). Collectively, these data suggest that hypohydration may not compromise cognitive function, nor any of the investigated subdomains more than if euhydration were maintained. Recommendations to avoid moderate hypohydration on the basis of maintaining optimal cognitive function are therefore not substantiated by this meta-analysis.Following the systematic literature search, it was apparent that poor methodological control was often implemented within studies investigating this topic, with resultant data often the product of several conflating factors, each of which could prospectively influence cognitive function. Therefore, study 1 was undertaken to investigate the separate effects of hydration and cognitive function when implementing strong methodological controls to limit the influence of confounding factors that might arise during exercise-induced dehydration. Here, 15 participants (12 males and three females, means ± standard deviation (SD); age 27 ± 6 years; peak oxygen consumption (VO2PEAK) 49.0 ± 6.2 mL/kg/min) completed a 90 min self-paced simulated military march in the heat (40.7 ± 1.2 degrees Celsius (°C) and 35 ± 7% relative humidity (RH)), either maintaining euhydration by consuming fluid ad libitum or becoming hypohydrated through fluid restriction. A cognitive testing battery that evaluated complex attention, executive function, learning and memory, and perceptual-motor function was administered pre-exercise and following a rest period after the cessation of the march, which was sufficient to return core temperature (TC) back to pre-exercise levels (55 ± 7 min). Subjective estimates of performance were also quantified using a visual analogue scale (VAS) that explored thirst, concentration, alertness, lethargy, and task difficulty. Aspects of memory and executive function were not comparable to pre-exercise data (both P ≤ 0.05), while a shift in the speed-accuracy trade-off was apparent in the switching attention task (executive function), with accuracy decreasing (P < 0.01), while reaction time was supplemented (P = 0.03). Despite BM losses of ~2.3%, performance in each of the measured cognitive domains was comparable between the conditions post-march (all P ≥ 0.26). When hypohydrated, subjective estimates of thirst were significantly greater post-exercise (P < 0.01), and moderate effect sizes were found for lethargy and task difficulty post-exercise (both d < 0.553). These results suggest that maintaining euhydration predominantly preserves cognitive function, although, this does not produce superior cognitive performance compared with fluid restriction following an identical exercise task. As such, cognitive performance appears to remain largely stable, despite losses in BM exceeding 2%. In addition to the paucity of methodological rigor identified among the literature from the meta-analysis, a lack of data pertaining to learning was also evident. Indeed, the learning and memory subdomain in the meta-analysis was largely informed by acute tests of memory function. Therefore, a novel study into the effects of hypohydration and motor sequence learning was conducted to determine whether this physiological stressor may compromise motor learning capability. This study also investigated whether neural activity increases during hypohydration, and considered whether this may be a plausible mechanism for equalising learning capacity. 20 males and 10 females were pseudo-randomised to either a control (10 males and five females, means ± SD; age 33 ± 10 years; VO2PEAK 40.1 ± 7.0 mL/kg/min) or hypohydrated (10 males and five females, means ± SD; age 34 ± 10 years; VO2PEAK 42.3 ± 7.6 mL/kg/min) group. Motor sequence learning was quantified in each group using a serial reaction time task (SRTT). Prior to practicing the SRTT, the hypohydrated group completed an exercise induced dehydration regime to elicit fluid losses of greater than 2% of BM, which was followed by a rest period of sufficient duration to allow TC to return to pre-exercise levels (46 ± 16 min). The control group did not complete exercise prior to SRTT practice. All participants returned following a retention interval (302 ± 30 min) in a euhydrated state as confirmed by urine specific gravity and BM, and completed a retention and transfer test of the SRTT. Functional near-infrared spectroscopy (fNIRS) was used throughout the SRTT to measure cerebral haemodynamics and neural activity. Speed and accuracy were obtained from the SRTT and collated into the inverse efficiency score (IES). Despite losses in BM of ~2.4%, IES performance in the hypohydrated group was similar to the control group (P = 0.36). The overall percentage improvement in learning over the practice trials was deemed similar between the groups, differing by only 2% (P = 0.91). All metrics of cerebral haemodynamics significantly changed with respect to a baseline collection taken immediately prior to practice (all P ≤ 0.01), however, oxyhaemoglobin (O2Hb) was elevated in the hypohydrated group during the initial practice block (P = 0.02). IES performance in both the retention and transfer tests were similar between the groups (P = 0.36), and there were no differences in cerebral haemodynamics between the conditions during these trials (all P ≥ 0.46). These results suggest that SRTT practice was not influenced by hydration status, with compensatory neural activity likely implemented to equalise behavioural performance when fluid deficits approximate to 2% of BM. This compensation does not appear to have behavioural or neurophysiological ramifications during subsequent executions in comparable or novel contexts when in a euhydrated state. The preceding studies show that hypohydration of at least 2% of BM is likely to be of little consequence to one’s cognitive capability. Study 3 was conducted to explore alternative associations between hydration and cognition, and investigated whether mental fatigue may be exacerbated by the primordial emotion thirst, as these states appear to share a common neural representation. In a crossover design, thirst was monitored in 15 males (means ± SD, age 29 ± 7 years; VO2PEAK 49.9 ± 6.1 mL/kg/min) during 60 min of cycling in normothermic conditions (means ± SD, 25.4 ± 0.8 °C, 60 ± 6% RH). Participants either had unlimited access to water to consume to the dictates of thirst (sated), or fluid was withheld and replaced with periodic salt water mouth rinses (thirst). Following thirst manipulation, a stroop task was completed for 60 min to evoke mental fatigue. Prefrontal cortex (PFC) O2Hb and deoxyhaemoglobin (HHb) were monitored throughout the prolonged cognitive task using fNIRS, and subjective perceptions of fatigue were reported through a VAS. Participants were instructed to complete as many iterations of the task as possible within the allotted time, and behavioural performance was quantified through the total number of stroop task iterations completed, and by the collated accuracy and response time outcome, the IES. Subjective thirst was greater in the thirst condition prior to commencing the mentally fatiguing task (P < 0.01). In addition to completing fewer stroop task iterations (P = 0.05), IES was also significantly poorer during thirst in the latter portion of the prolonged cognitive task (P ≤ 0.03). Each condition produced increased O2Hb in response to the prolonged stroop task. However, these differed temporally, being earlier during thirst. Subjective estimates of fatigue increased as time on task lengthened (P < 0.01), but were similar between the conditions (P = 0.83). Thirst resulted in the completion of less cognitive work, and greater compromises in IES during the mentally fatiguing task. This is likely the product of an inability to sustain the synergy between neuroanatomical facilitation and inhibitory systems for a prolonged period of time, probably due to exacerbated inhibitory input produced by thirst. These data provide novel insight into the relationship between thirst and mental fatigue, and suggest that drinking to the dictates of thirst may be pertinent to sustaining prolonged cognitive task performance. Collectively, the findings of the meta-analysis and three completed studies indicate that hydration status is unlikely to be of significance to the modulation of cognitive performance, nor any specific underlying cognitive domains (complex attention, executive function, learning and memory, and perceptual-motor function) at least to the threshold examined in this thesis (~2.5%). Furthermore, there appears to be little evidence to support the recommendation that athletes should aim to avoid “excessive” (2% loss in BM) dehydration, at least with respect to preserving cognition. Instead, cognitive reserve is likely ensured by several compensatory responses that arise alongside hypohydration. Specifically, regional neural redistribution, indirect actioning of neurotransmitters or glucose from dehydration related hormones, and shifts in arousal. Seemingly, recommendations concerning hydration and cognitive performance need to be revised to more accurately reflect contemporary findings. Where, instead of adopting models based on the quantification of BM, consideration should instead be given to targeting the avoidance of thirst, as this appears to have a greater modulating effect on cognitive performance, particularly under contexts of sustained cognitive work.