Mountain ultra-marathons (MUMs) have become increasingly popular in the last decade and they are considered a model to investigate acute responses to extreme load and stress (Millet and Millet, 2012). As previous studies demonstrate, during and immediately after running, MUMs induce cardiac (Maufrais et al., 2016) and neuromuscular (Saugy et al., 2013) fatigue, lung functions alterations as pulmonary ventilations (Vernillo et al., 2015) as well as impairment in postural control (Degache et al., 2014). In parallel, an increase in total body water (Knechtle et al., 2008), in volumes of athletes' body extremities (e.g., feet) (Bracher et al., 2012) and the development of peripheral oedema have been identified (Zavorsky et al., 2014). Blood biomarkers changes also confirm this response, showing higher C-reactive protein concentrations (Millet et al., 2011a; Saugy et al., 2013) and decreased hematocrit levels in runners right after MUMs (Robach et al., 2014; Vitiello et al., 2015).

Besides these physiological changes, pioneer studies demonstrate the impact of extreme sport load also on the human brain. Disorders of serum-sodium occur commonly in athletes participating in endurance sports leading to threatening conditions for athletes health as cerebral edema and pulmonary edema (Ayus et al., 2000). The investigation of brain water diffusivity, allowing the measurement of the intercellular water and the assessment of cerebral edema, can therefore give a great contribution to the investigation of the effects of extreme sports on the brain (Hagen et al., 2007). Rapid diagnosis and appropriate therapy of these athletes is required to prevent severe complications or death.

In the present study, we investigated changes occurring during and after ultra-marathon in brain water diffusivity aiming at investigating the presence of cerebral edemas. In particular, by experience and observation, 48 h is the period post- MUM where the edemas (especially in the calves) are maximal.

Using brain magnetic resonance imaging (MRI) apparent diffusion coefficient (ADC) combined with physiological parameters, the present work aims to (1) investigate the effects of MUM on brain water diffusivity and (2) further explore non-invasively the interaction between MUM-induced body and brain water responses on master runners. This allows us to capture the complete picture of brain and physiological changes generated by endurance running.

We performed a serial investigation of MUM athletes at baseline, upon arrival and after 48 h including brain MRI and multiple blood parameters. Brain MRI apparent diffusion coefficient (ADC) derived from diffusion weighed imaging (DWI) was used, which measures the diffusion of water notably in the inter-cellular compartment. As we assume that there will be no significant neuronal axonal growing or shrinking within the short observational interval of less than 1 week, ADC provides an indirect in-vivo estimation of the inter-cellular compartment of the human brain. Additionally, we assessed multiple blood parameter covering inflammation, osmolality and electrolytes, and correlated brain ADC with these physiological parameters.

The current investigation assessed the effect of extreme effort (both extreme distance and elevation changes) in mountain ultra-marathons (MUMs) athletes on brain water diffusion and blood physiological parameters in master runners. The whole brain ADC significantly increased from baseline to arrival and then significantly decreased at recovery, to lower values than at baseline. Linked changes in physiological electrolytes concentration are also present during and after MUM. Our findings identify the presence of a brain-body matrix suggesting that together with changes at the body level, modifications at the brain level can be also demonstrated upon arrival, and that even 48–72 h after arrival, the brain has not yet completely recovered to the baseline state.

ADC measures the diffusion of water notably in the inter-cellular space (Duong et al., 2001; Sehy et al., 2002). Two are the main origins for an increase in ADC values: The first is the neuronal and axonal degeneration (Simon and Kliot, 2014) that consists in progressive structural atrophy occurring over years, for example during neurodegenerative diseases (Erickson et al., 2013). This scenario is unlikely in the current setting, since the ADC changes appear in a range of few days.

The second option consists in rapid changes in the inter-cellular space. The typical example is acute stroke (Albers, 1998; van Everdingen et al., 1998) when neurons swell in the acute phase, inter-cellular space is therefore reduced and consequently ADC values decrease (Marks et al., 1996; Birenbaum et al., 2011). In the context of MUM, changes in water and electrolyte concentration may rapidly modify the neural inter-cellular space (Sotak, 2004). Due to the speed of these alterations, this is the most plausible scenario and can be considered parallel to the one occurring at the physical level already reported by previous studies on MUMs.

In fact, after ultra-marathon an increase in total body water (Mattson, 2014), in volumes of athletes' body extremities (e.g., feet) (Bracher et al., 2012) and the development of peripheral oedema have been identified. Following this line of research, our results suggest for the first time the occurrence of an overall increase in inter-cellular water also in the brain of runners immediately after ultra-marathon. Few days after MUM, when the athletes recover, extra-cellular space increases in the whole brain from baseline to arrival and then it decreases overshooting the baseline and not yet completely recovering to the baseline equilibrium state after 48 h.

To further explore this brain-body interaction after extreme sport, we assessed multiple parameters in the peripheral blood. First, we observed a significant increase in CRP (c-reactive protein) upon arrival and a partial recovery after 48–72 h. The increase in inflammation demonstrated after MUMs (Millet et al., 2011a; Saugy et al., 2013) can also explain the increase in total body water because inflammation was shown to be associated with tissue oedema (Proske and Morgan, 2001). Our results are therefore in line with previous works (Kasapis and Thompson, 2005; Stewart et al., 2007) showing that extreme exercise produces a short-term, systemic inflammatory reaction, whereas a long- term “anti-inflammatory” response can be expected. Confirming this interpretation, urine osmolality follows the CRP pattern also increasing and then partially recovering. This reflects an exchange in total body water that can also be explained by dehydration and hydration mechanisms during and after sport (Shirreffs and Maughan, 1998; Oppliger et al., 2005; Popkin et al., 2010).

Moreover, concerning electrolytes, the changes over time can also be compared to the brain ones. In particular, the hematocrit, which measures the percentage of serum vs. corpuscular components of the blood, first decreased and then partially recovered. Coherently with our previous interpretation on ADC results and with previous studies on MUMs the large decrease in HCT from T1 to T2 may reflect a large hemodilution: Expansion in plasma volume and in total extracellular water (Robach et al., 2014; Vitiello et al., 2015).

The pattern of CRP, urinary osmolality and HCT mirrors our previous neuroimaging results reflecting an increase in extra-cellular space from T1 to T2. On the contrary the ADC overshooting beyond the baseline (from T2 to T3) seems to be a peculiar pattern of the brain, not reflected by electrolytes.

On the other side, the electrolytes calcium, chloride, potassium and sodium had different and opposite evolutions over time. While calcium and potassium had a similar pattern decreasing from T1 to T2 and then increasing without reaching the baseline level at T3, chloride presented an opposite pattern, increasing at arrival (T2) and not decreasing anymore. Sodium remained always constant. Our study extents findings of previous studies on MUMs (Millet et al., 2011b; Millet and Millet, 2012; Saugy et al., 2013), showing specific modifications in electrolytes concentration after 48 h of intense sport activity. In particular, an increase in plasma volume due to an increase in the total exchangeable potassium has been already previously reported after endurance events (Milledge et al., 1982; Maughan, 1985; Fellmann et al., 1999; Knechtle et al., 2008).

This complex pattern of the various electrolytes over time (plus the other electrolytes not assessed here) may explain the stability of the serum osmolality.

After measuring brain water diffusion parameters and blood biomarkers, we performed additional correlation analyses to find a direct link between electrolytes variations and brain ADC changes. For T2–T1, an increase in cerebral ADC is associated to higher levels of plasma osmolality and cholesterol. For T3–T2 a decrease in ADC level is negatively associated to CRP, chloride, sodium and osmolality and positively to cholesterol.

We want to highlight that the results concerning osmolality are of great interest confirming our previous interpretation. They suggest that body water might directly influence the change in cerebral ADC, increasing first and then decreasing. They establish therefore a direct link between general brain structural changes and physiological biomarkers modifications produced by extreme sport.

In conclusion, our findings show that MUMs induce in the brain of runners increased inter-cellular water in the brain and increased extra-cellular water in the body.

Finally, we want to highlight the absence of differences for handiness and footedness and after post hoc analyses of ADC. Our interpretation is that during marathon races and extreme sports in general, ADC changes in the brain reflect overall body hemodynamics changes rather than specific changes in brain functions.

Major strengths of the current investigation include the prospective study design, and the rigorous assessment of both brain MRI and blood samples before, upon arrival and during recovery of extreme mountain marathon athletes. Major limitations include the small sample size and the indirect assessment of brain modifications. Concerning the small sample, this can be a reason why the correlations blood biomarkers-ADC values didn't survive multiple comparisons corrections. Future studies with higher number of participants are recommended. Moreover, the assessed ADC values indirectly assess the diffusion of water dominantly in the inter-cellular compartment of the brain, and serum electrolytes may only indirectly indicated modifications at the brain level due to the blood brain barrier. Furthermore, ADC changes in the brain can be influenced by sleep deprivation during the race, for which we didn't perform any standardized assessment. However, in the absence of more precise and non-invasive assessments at the level of the brain, the observed indirect methods provide at least a feasible approach to assess modifications of the brain in response to extreme physical activity. Concerning the blood sampling, some limitations must be acknowledged. One limitation was the timing of the blood sampling. The samples were not taken at exactly the same time because all the subjects did not run at the same speed, even if the sampling was made at identical moment of the Tor. Moreover, intra-individual biological variations exist which can have an impact on the results. The post-exercise liquid intake could also lead to biological modifications due to a possible hemodilution but as the hematocrit didn't change a lot, we conclude that the impact of this potentially confounding factor is very low. The subjects were allowed to consume food and fluid ad libitum during the race, but the intake was not monitored due to logistical constraints and the difficulty to access the runners on mountainous path along the MUM (Maufrais et al., 2016). Furthermore, the exact weather conditions during the race were not monitored. Finally, having diet, exercise, caffeine intake values before even before T1 and after T3 would be a great improvement for our study. Unfortunately, most runners were not locals and came the day of the race and left the day after the study. It was very difficult to control for these factors.

The current investigation assessed the effect of extreme effort (both extreme distance and elevation changes) in MUMs athletes on brain water diffusion and blood physiological parameters. Extreme physical activity leads to a significant increase of the inter-cellular compartment of the brain, which is not at the equilibrium state even 48–72 h after the race.

Conceived and designed the experiments: MV, SH, PC. Performed the experiments: MV, PC. Data analyses: DZ, SH, and MV. Blood analyses: CL. Manuscript writing: DZ, SH, and MV. Manuscript editing: PC, GM, GG, and CL. Ethical Commitee and whole project organization: GG, GM, MV, and PC.