A common practice used by participants in weight-category sports is to achieve an acute loss of body mass before competition. This might be to compete at a lower weight category (boxing, martial arts) or to comply with weight penalties used in some sports (e.g., horse racing). Athletes in weight-category sports tend to display higher morning urine osmolality values than those observed in the general population, perhaps reflecting their attempts to maintain a low body mass through dehydration (18). A survey of Australian jockeys found that 81% restricted food and fluid intake in the 24 h before competition, with exercise/sauna-induced sweating and diuretic use also regularly used to reduce body mass (12). These practices all result in a progressive loss of body water, causing an increase in plasma osmolality. This in turn produces a movement of water from interstitial and intracellular spaces to defend blood volume, thus helping to maintain blood flow to the exercising muscles and to the skin (13). Although this response is well defined in the periphery, it is also clear that an increase in extracellular osmolality can exert an influence on total brain (3) and brain cell (1) volumes, potentially also influencing the permeability of the blood-brain barrier (25).
At present, there are limited data on the effect of dehydration on the central nervous system (CNS) in humans. Recent work has used magnetic resonance imaging (MRI) to determine the effects on brain volume of dehydration induced by exercise under conditions of restricted heat loss (5,10) and during a 16-h period of fluid restriction (7). Interestingly, although changes in brain volume were reported after fluid restriction protocols, no change was observed after exercise-induced dehydration. The two previous reports investigating the response to exercise have reported marked changes in ventricular volume, with the magnitude of change loosely related to the degree of hypohydration induced (5,10). The reasons for the apparent discrepancy between these passive and active forms of dehydration are unclear, particularly because the magnitude of total body water loss, expressed as a percentage reduction in body mass, was greater in the exercise studies (passive = 1.6% ± 1.0%, active = 2.3% ± 0.2% and 2.2% ± 0.5%). It is possible that the gradual loss of fluid over a prolonged timescale with fluid restriction produces a different response from that which occurs with rapid loss of fluid caused by exercise in the heat, but this can only be speculated. Animal data suggest that any loss of fluid from the brain results in the rapid movement of osmolytes from the cerebrospinal fluid (CSF) and serum into the brain to defend brain volume and limit further fluid losses (8). Water loss and electrolyte uptake are likely to occur almost simultaneously, with the magnitude of brain water losses demonstrated to be only 35% of that anticipated in the 30 min after osmotic opening of the blood-brain barrier (3). It is possible that redistribution of fluids to restore brain volume might have confounded the results of some studies investigating the effect of exercise and thermally induced fluid losses (5,10,14).
In some sports, there can be a long interval between the weigh-in and the start of competition (e.g., professional boxing). This movement of fluids might protect against potential changes in brain volume resulting from the dehydration practices used by competitors in these sports. There are situations, however, where competition begins only a few minutes after a weigh-in (horse racing, amateur boxing/martial arts): this, coupled with any hypohydration accrued during the activity itself, might place these individuals at an increased risk of traumatic brain injury. A marked reduction in the size of the brain has the potential to increase trauma experienced during falls and collisions because of greater movement of the brain in the skull after impact. Apart from the risk of intracranial bleeding, collisions between the surface of the brain and the walls of the interior skull might cause the cortical tissue to be deformed, compressed, or stretched, and this has been implicated in the etiology of concussion (16). Clearly, the high risk of falls or blows to the head in some sports can place competitors at risk of traumatic brain injury. Whether practices routinely used to achieve a desired competition weight can affect brain volume is therefore an interesting question. This is true from a purely mechanistic standpoint but might also have implications for sports medicine because the number of concussions and other head injuries sustained by competitors in weight-category sports appears to be greater than reported in other sports (2,23).
The aim of this study was to examine the effect of exercise-induced hypohydration on brain, ventricular, and CSF volumes. Previous studies examining changes in brain volume with dehydration have reported inconsistent results, particularly in the changes observed in ventricular volume. These investigations have used varying levels of dehydration, which could explain some of the inconsistency in this response. In addition, there has yet to be any examination of the potential cell-volume regulation response in humans that has been reported in animals to defend brain volume in the presence of an osmotic challenge. To achieve this, we made every effort to complete the first measurements as soon as possible after the cessation of exercise. In addition, responses were observed over a 2-h period after exercise where no fluid was provided to examine a time course of change.
MATERIALS AND METHODS
Eight physically active men (mean ± SD: age = 26 ± 4 yr, height = 1.79 ± 0.09 m, body mass = 79.7 ± 11.2 kg, V˙O2peak = 4.1 ± 0.3 L·min−1, % body fat = 12.2% ± 1.6%) were recruited to take part in this investigation. All subjects were physically active and habituated to the sensation of strenuous exercise, but none was accustomed to exercise in a warm environment at the time of the study. Before volunteering, all subjects received written details outlining the nature and the purpose of the study. After questions regarding the protocol, a written statement of consent was signed. The protocol was approved by the ethical advisory committees of the Loughborough University and the University of Nottingham.
All subjects completed a preliminary test to measure peak oxygen uptake (V˙O2peak), a familiarization trial, and a single experimental trial. V˙O2peak was determined using a discontinuous, incremental exercise test to volitional exhaustion on a cycle ergometer (Monark Ergomedic 824E, Varberg, Sweden). The familiarization trial involved the completion of the same exercise protocol as the experimental trial. The familiarization and the experimental trials were separated by at least 7 d to limit the development of heat acclimation. Subjects were instructed to record dietary intake and physical activity during the day before the familiarization trial and to replicate this before the subsequent experimental trial. No strenuous exercise or alcohol consumption was permitted for 24 h before each visit to the laboratory.
The trials took place in the morning after an overnight fast, other than the ingestion of 500 mL of plain water 90 min before commencing exercise. On arrival at the laboratory, subjects first emptied their bladder, and a sample of urine was retained for measurement of osmolality (Gonotec Osmomat 030; YSI, Farnborough, UK). Subjects then sat in a comfortable environment (22°C-24°C) for 20 min before entering a climatic chamber maintained at a temperature of 35.1°C ± 0.4°C and relative humidity of 57% ± 3%, where nude body mass was measured to the nearest 10 g (Adam CFW150 digital scale, Milton Keynes, UK). The subject then began a series of 10-min periods of cycle exercise at an intensity corresponding to 60% of V˙O2peak. A 5-min rest period separated successive blocks of exercise, during which the subject toweled dry and the nude body mass was recorded. Sweat losses were determined for each period of exercise through changes in body mass, and this pattern of activity and rest continued until the subject had lost approximately 3% of their initial body mass. Subjects then returned to a comfortable environment and were not permitted to ingest any fluid for 2 h after the end of exercise.
Because of limitations associated with the monitoring of core temperature when exposed to MRI, this was measured during the familiarization trial. A radiotelemetry pill (HQ Inc., Palmetto, FL) was ingested 10 h before this trial, before an overnight fast, to enable intestinal temperature to be determined. The subjective feelings of thirst, alertness, and fatigue were assessed using a series of 100-mm visual analog scales before exercise, immediately after exercise, and during the 2-h recovery period.
To determine the normal variation and/or test-retest error in measured brain and CSF volumes over the same time frame as the experimental trials, two subjects, who did not participate in the subsequent trials, remained at rest in a comfortable environment over a 3-h period. During the experimental trial, measurements of brain, ventricular, and CSF volumes were undertaken before exercise, upon attainment of 3% body mass loss, and 1 and 2 h after the end of exercise. Anatomical MPRAGE and Vista scans were performed at each time point using a 3-T magnetic resonance scanner (Philips Achieva, AE Eindhoven, The Netherlands). Imaging parameters were as follows: MPRAGE, echo time = 2.3 ms, repetition time = 7.7 ms, voxel size = 0.8 × 0.8 × 0.8 mm, 256 × 256 matrix, field of view = 205 × 205 × 147 mm; Vista (this scan specifically highlights the CSF and was used in volumetric method 2 described below), echo time = 380 ms, repetition time = 2500 ms, voxel size = 0.8 × 0.8 × 0.8 mm, 312 × 312 matrix, field of view = 250 × 250 × 180.
The brain in all scans was first extracted using the brain extraction tool (BET) skull stripping tool (19). Brain-extracted images were then inspected to ensure that any residual nonbrain tissue (neck tissue, nasal sinus, skull, etc.) had been removed. All input scans were transformed to the center position of the set to ensure a homogeneous amount of blurring related to interpolation, using robust linear registration (15). Percentage changes in brain, ventricular, and CSF volumes were then estimated using two methods.
Volumetric method 1.
Changes in brain volume were determined from the MPRAGE scans using SIENA (version 2.6; FMRIB Software Library, Oxford University, UK), a fully automated analysis tool designed for the longitudinal assessment of brain atrophy (20,21). SIENA segments the brain images to determine the brain/CSF edge points, with the perpendicular displacement of these edge points used to estimate percentage brain volume change between scans. Regional brain volume changes were determined using the voxelwise extension of SIENA: the edge displacement images generated for each subject were spatially dilated, transformed to a standard space, masked with a standard brain edge image (MNI152), and smoothed by a Gaussian kernel of 5 mm. Changes in ventricular volume were determined by manually removing the nonventricular CSF before tissue-type segmentation undertaken using FAST (FMRIB Software Library) to determine the brain/nonbrain boundaries (27).
Volumetric method 2.
We first thresholded the preexercise Vista scan at an intensity level commensurate with that of CSF (8e + 5 in this case). This reference scan was then precisely nonlinearly registered to the subsequent scans with subvoxel precision. The resulting images were thresholded at the same level, and the sum of the voxel intensity values above the threshold was computed. This integral measure can then be used to compute differences across subject or before and after exercise. This approach was intended to alleviate the influence of changes in contrast, brightness, and homogeneity across scans and the artifacts of partial voluming. A proxy for brain size was also obtained by linearly registering the reference scan to the ICBM standard atlas, with the product of the scaling factors along each axis used as a covariate in the statistical analysis.
Blood handling and analysis.
Blood samples (5 mL) were drawn by venous puncture from a superficial antecubital vein before the start of exercise, at the end of the dehydration protocol, and after 1 and 2 h of recovery. Subjects were seated for at least 15 min before each collection. Blood samples were drawn directly into dry syringes: a 1-mL aliquot was dispensed into tubes containing K2EDTA, with the remaining 4 mL placed into a plain tube. Hemoglobin (in duplicate by the cyanmethemoglobin method) and hematocrit (in triplicate by microcentrifugation) values were used to estimate percentage changes in blood, plasma, and red cell volumes relative to the resting sample (6). The portion of the sample added to a plain tube was left to clot at room temperature for 60 min before being centrifuged to yield serum. This was kept at 4°C for the analysis in duplicate of serum sodium, potassium (Corning 410C, New York), and chloride (Jenway PCLM 3, Essex, UK) concentrations and serum osmolality (Gonotec Osmomat 030; YSI).
Data are presented as means ± SD unless otherwise stated. To identify differences in normally distributed results, we used one-way repeated-measures ANOVA. Pairwise differences were evaluated using paired t-tests with the Holm-Bonferroni adjustment for multiple comparisons. Statistical significance was accepted at P < 0.05. Cohen's d effect sizes (ES) for the changes in brain, CSF, and ventricular volumes were also determined [(Volpre± − Volpost)/SD]. Regional changes in brain volume were determined by permutation-based, voxelwise nonparametric testing using Randomise (FMRIB Software Library). Statistical inference was based on voxel-level thresholding at corrected P values of 0.05, adjusted for multiple comparisons.
Preexercise urine osmolality was 495 ± 116 and 449 ± 145 mOsm·kg−1 during the familiarization and experimental trials, respectively (P = 0.438). Total body mass loss was 2.31 ± 0.29 kg during the familiarization trial and 2.30 ± 0.27 kg in the experimental trial, representing a 2.9% ± 0.1% reduction in body mass. Subjects took 67.4 ± 10.2 min of exercise to reach this level of hypohydration, with a total of 107 ± 15 min spent inside the climatic chamber.
Although a full control trial was not included in the present investigation, normal variation in measured brain and CSF volumes was monitored over the same time frame as the experimental trials in two subjects who did not participate in the main study. There was no change in brain (difference from scan 1: scan 2 = 0.0% ± 0.0%, scan 3 = 0.0% ± 0.0%) or CSF (difference from scan 1: scan 2 = 1.0% ± 0.3%, scan 3 = 0.2% ± 0.0%) volumes over the 3-h rest period. Because of the limited sample size, it is difficult to draw definitive conclusions regarding the normal variation and/or measurement error at rest, but these values fall within the reported test-retest error of SIENA (0.2% brain volume change ).
The change in brain and CSF volumes resulting from exercise/heat-induced hypohydration determined using SIENA is presented in Figure 1. Brain volume changes were unaffected by hypohydration (0.2% ± 0.4%; ES = 0.2, P = 0.310). A modest reduction in ventricular volume was observed after exercise (4.0% ± 1.8%; ES = 4.6, P < 0.001), with evidence of a gradual restoration over the 2-h recovery period. This response is highlighted in Figure 2, with significant morphometric edge flow of the brain/CSF boundary surrounding the ventricles indicating some degree of shrinkage between the preexercise and postexercise scans and the preexercise and 2-h recovery scans (P < 0.05). Using analysis method 2 (Vista scan), percent changes in brain and CSF volumes were also estimated (Fig. 3). Again, brain volume was not influenced by hypohydration (0.0% ± 0.4%, P = 0.805), but in a similar manner to the ventricular volume data presented above, a small reduction in CSF volume was apparent after exercise (3.1% ± 1.9%, P = 0.003). This response was maintained throughout the 2-h rest period. There was a moderate relationship between the change in brain volume calculated using analysis methods 1 and 2 (r2 = 0.611, P = 0.022), with SIENA estimating a slightly greater change in brain volume after exercise in seven of eight subjects.
Blood and plasma volumes were 9.0% ± 2.5% and 13.7% ± 3.9% lower, respectively, immediately after exercise, with some recovery of both volumes apparent during the recovery period (P < 0.001; Fig. 4). Serum sodium concentration increased from 144 ± 1 to 148 ± 1 mmol·L−1 during exercise (P = 0.003), but there was no change in serum potassium (P = 0.062) or chloride (P = 0.210) concentrations (Table 1). Compared with preexercise levels, serum osmolality was increased at the end of exercise (+10 ± 2 mOsm·kg−1, P < 0.001) and remained elevated throughout the 2-h postexercise period (Table 1).
Core temperature data obtained using a radiotelemetry pill are presented in Figure 5. Preexercise core temperature was 37.1°C ± 0.3°C, with exercise resulting in an increase of 2.2°C ± 0.4°C. Core temperature at the end of exercise was 39.3°C ± 0.5°C but returned to preexercise levels after 1 h of recovery. Subjective feelings related to thirst and hunger, tiredness and ability to concentrate, and headache are presented in Table 2. There was a marked increase in perceived thirst reported immediately after exercise, with this maintained throughout the 2-h postexercise period (P < 0.001). There was no change in perceived feelings of tiredness (P = 0.201) or the ability to concentrate (P = 0.378), but the subjects reported a drop in alertness (P = 0.037) and perceived energy (P = 0.021) after exercise. Subjects reported no symptoms of headache before exercise, but feelings of head soreness were apparent after exercise and during recovery (P = 0.011).
The results of this study demonstrate that brain volume remains unchanged in response to moderate levels of acute hypohydration induced by exercise in a warm environment, but there do appear to be small reductions in ventricular and CSF volumes. Osmotically driven movement of fluid from the CNS to the circulation appears to be minimal after exercise/heat-induced hypohydration, perhaps suggesting that mechanisms are in place to defend brain and CSF volumes in spite of a substantial increase in extracellular osmolality. The protocol used in the present investigation resulted in the loss of 2.9% ± 0.1% of the subjects' initial body mass: this level of hypohydration would be sufficient to reduce exercise capacity, particularly when exercise is performed in a warm environment (11).
The white and the gray matter of the brain consist of approximately 70% and 82% water, respectively (22), and brain volume is therefore susceptible to changes induced by fluctuations in water content. The water content of the brain is dictated largely by the solute content of the brain tissue relative to that of the extracerebral space, and a change in plasma osmolality can result in a net flux of water into the circulation. Several animal studies have reported marked changes in brain volume in response to peripheral infusions of hypertonic NaCl and mannitol that resulted in a state of hypernatremia (1,3,4,9). Similarly, in the presence of hyponatremia, water can move from the circulation into the brain (along an osmotic gradient) resulting in cerebral edema (8). In the rat, acute hypernatremia results in a 7% reduction in total brain volume within 30-90 min, with the fluid being drawn primarily from the extracellular water compartments (4). These authors also reported that the degree of water loss from the brain in response to the osmotic challenge was not as great as predicted on the basis of ideal osmotic behavior, and they attributed this to a net influx of electrolytes from the CSF and the plasma (1,3). This volume regulation acts as a defense mechanism that limits the degree of brain shrinkage (1,8). It should be noted, however, that the hypernatremia induced in the nephrectomized rats studied by Cserr et al. (4) was particularly severe, with plasma sodium concentrations increasing by 30 mmol·L−1. This is far beyond the extent of change that would occur in healthy humans during prolonged exercise, when sodium concentration seldom increases by more than about 5-6 mmol·L−1. In a group of 90 marathon runners, for example, the mean increase in plasma sodium concentration over the course of the race was 6 mmol·L−1 (26).
At present, there are limited data on the effects of changes in fluid balance on the CNS in humans, but three recent studies using scanning techniques have examined changes in brain and ventricular volumes with hypohydration (5,7,10). A 16-h period of fluid restriction, resulting in a body mass loss of 1.6% ± 1.0%, produced a significant reduction in total brain volume. Brain volume was subsequently restored by rehydration with plain water (7). Using various degrees of exercise and heat exposure, Dickson et al. (5) and Kempton et al. (10) reported no reduction in total brain volume, but they found marked changes in ventricular volume with levels of hypohydration between 1.7% and 2.9%. The striking outcome of both these exercise studies was the marked variation in ventricular volume change observed after exercise (changes from −15% to +42%), with the magnitude and direction of change related to the degree of hypohydration induced. It should be noted that the extent of fluid loss induced in these studies was highly variable, with the subjects who attained body mass losses of more than 2.5% displaying ventricular expansion, a response which is difficult to explain. Subjects in this study all lost between 2.8% and 3.1% of body mass, but none of these subjects showed an increase in ventricular volume. The present data suggest that brain volume is tightly maintained in the presence of marked reductions in total body water, and these distinct and variable changes in ventricular or CSF volumes were not apparent.
The present study found a 3.1% ± 1.9% reduction in CSF volume in the presence of moderate levels of hypohydration. Ventricular volume was also reduced immediately after exercise-induced hypohydration, with a gradual restoration of this loss apparent over the 2-h recovery period. Although these differences were statistically significant, it is unlikely that a change of this magnitude is physiologically relevant, with the change in CSF volume representing a fluid loss of only 4-5 mL from this compartment compared with the larger changes in blood (9% ± 3%) and plasma (14% ± 4%) volumes. This represents a loss of about 410-600 mL of water from the vascular space on the basis of the assumption that blood volume was 5582 ± 786 mL in these subjects (blood volume estimates based on 70 mL·kg−1 body mass ). Regional changes in brain and ventricular volumes were highlighted using the voxelwise extension of SIENA. Both methods used to estimate differences in brain volume reported no global volume change, but there was a suggestion of some regional redistribution of fluid within the brain after exercise and heat-induced hypohydration (Fig. 2). The physiological importance of these changes are unknown at present, but this may be implicated in the decrements in several aspects of cognitive function reported after strenuous exercise in warm conditions.
The technical challenges associated with the measurement of changes in brain, ventricular, and CSF volumes using MRI techniques should be recognized. The CSF is produced by the ependymal cells of the choroid plexus found throughout the ventricular system and provides basic mechanical and immunological protection to the brain inside the skull as well as a transport medium for nutrients and neurotransmitters. CSF flows from the lateral ventricles into the third ventricle and then the fourth ventricle via the cerebral aqueduct in the brainstem. The aqueduct between ventricles is small and presents a common source of error when determining ventricular and CSF volumes. The application of two separate volumetric analysis techniques, with both producing similar outcomes, should provide additional confidence in the data presented. It is difficult to explain the marked discrepancy in changes in ventricular volume reported between this study and those published previously (5,10), but it is possible that the variable levels of hypohydration attained in the earlier studies might be a factor, along with differences in the time frame between the end of exercise and the completion of the MRI scan.
In this study, serum osmolality increased by 10 ± 2 mOsm·kg−1 (P < 0.001) during exercise, with mean values of 294 ± 2 mOsm·kg−1 observed at the end of exercise. This is similar to the response observed in a previous study investigating changes in serum S100β concentrations during intermittent exercise in the heat, where serum osmolality increased by 12 ± 3 mOsm·kg−1 (25). Although this change is small relative to that produced by mannitol or glycerol infusion (which can result in serum osmolality values in excess of 310 mOsm·kg−1), the resulting osmotic gradient between the periphery and the CNS should have been sufficient to produce a net movement of fluid from the brain to the circulation. As highlighted previously, the degree of brain water loss observed in animals exposed to acute hyperosmotic states was not as great as predicted on the basis of ideal osmotic behavior (3). The accumulation of solutes in the brain, including electrolytes drawn from the CSF and the plasma, occurs during acute hyperosmotic states to limit fluid losses. This volume regulation occurs rapidly and acts to restrict brain shrinkage (1,8). It appears that this mechanism is in operation when individuals are exposed to moderate hypohydration resulting from exercise in warm/humid conditions.
This movement of fluids might protect against potential changes in brain volume resulting from the dehydration practices used by competitors in some sports where a significant time interval exists between the weigh-in and the start of competition (e.g., professional boxing). There are cases, however, where competition begins shortly after a weigh-in (horse racing, amateur boxing/martial arts): this, coupled with any hypohydration accrued during the activity itself, could place these individuals at an increased risk of traumatic brain injury. The convulsive theory of concussion suggests that an abrupt increase or a sudden arrest of head movement results in an angular acceleration of the head that causes turbulent rotation of the mass of brain within the cranium (16). Sudden mechanical loading of the head might increase the chances of a tissue-deforming collision between the brain cortex and the walls of the skull. Any reduction in the volume of the brain or in the volume of CSF surrounding the brain might increase trauma experienced during falls and collisions because of a greater degree of brain movement within the skull.
It is important to recognize the limitations of this study and of other MRI-based studies that have examined changes in brain volume. It is clear that mechanisms operate to limit fluid losses from the brain, and the magnitude of brain water losses has been demonstrated to be only about 35% of that expected 30 min after infusion of hyperosmotic solutions (3). Subjects were positioned in the MRI magnet within 5 min after the end of exercise, and the scanning procedure took around 10 min to complete. Despite efforts to ensure a rapid turnaround, water loss and electrolyte uptake are likely to occur simultaneously, potentially causing part of any response to be missed. Animal studies investigating the effect of dehydration by 10% of initial body mass also report no change in brain volume, but these measurements were taken once core temperature had returned to basal levels (14). Again, it is probable that there was a redistribution of fluids during this time to restore brain volume. It is also important to acknowledge the likely influence of posture change on the results presented. Moving from a standing to supine position results in a marked redistribution of fluid within the body, with blood and plasma volumes increasing by 8% and 17%, respectively, after 60 min in the supine position (17). It is possible that moving from a seated position while at rest and during exercise to a supine position in the scanner could have distorted the results, but this same artifact should be present in all scans of this nature unless the posture of the subjects is first stabilized and then maintained throughout the study period.
In conclusion, these data demonstrate that brain volume remains unchanged in response to moderate levels of hypohydration, but there does appear to be a small transient reduction in CSF volume. Osmotically driven movement of fluid from the CNS to the circulation appears to be minimal after exercise/heat-induced hypohydration, perhaps suggesting that mechanisms are in place to defend brain volume.
The authors acknowledge the Gatorade Sport Science Institute (GSSI), Barrington, Illinois, for providing financial support for this study.
The results of this study do not constitute endorsement by the American College of Sports Medicine.
1. Ayus JC, Armstrong DL, Arieff AI. Effects of hypernatraemia in the central nervous system and its therapy in rats and rabbits. J Physiol
2. Bledsoe GH, Li G, Levy F. Injury risk in professional boxing. South Med J
3. Cserr HF, DePasquale M, Patlak CS. Regulation of brain water and electrolytes during acute hyperosmolality in rats. Am J Physiol
4. Cserr HF, DePasquale M, Nicholson C, Patlak CS, Pettigrew KD. Extracellular volume decreases while cell volume is maintained by ion uptake in rat brain during acute hypernatremia. J Physiol
5. Dickson JM, Weavers HM, Mitchell N, et al. The effects of dehydration on brain volume-preliminary results. Int J Sports Med
6. Dill DB, Costill DL. Calculation of percentage changes in volumes of blood, plasma, and red cells in dehydration. J Appl Physiol
7. Duning T, Kloska S, Steinstrater O, Kugel H, Heindel W, Knecht S. Dehydration confounds the assessment of brain atrophy. Neurology
8. Gullans SR, Verbalis JG. Control of brain volume during hyperosmolar and hypoosmolar conditions. Annu Rev Med
9. Holliday MA, Kalayci MN, Harrah J. Factors that limit brain volume changes in responses to acute and sustained hyper- and hyponatremia. J Clin Invest
10. Kempton MJ, Ettinger U, Schmechtig A, et al. Effects of acute dehydration on brain morphology in healthy humans. Hum Brain Mapp
11. Maughan RJ, Shirreffs SM. Development of individual hydration strategies for athletes. Int J Sport Nutr Exerc Metab
12. Moore JM, Timperio AF, Crawford DA, Burns CM, Cameron-Smith D. Weight management and weight loss strategies of professional jockeys. Int J Sport Nutr Exerc Metab
13. Nose H, Mack GW, Shi X, Nadel ER. Shift in body fluid compartments after dehydration in humans. J Appl Physiol
14. Nose H, Morimoto T, Ogura K. Distribution of water losses among fluid compartments of tissues under thermal dehydration in the rat. Jpn J Physiol
15. Rueckert D, Sonoda LI, Hayes C, Hill DLG, Leach MO, Hawkes DJ. Non-rigid registration using free-form deformations: application to breast MR images. IEEE Trans Med Imaging
16. Shaw NA. The neurophysiology of concussion. Prog Neurobiol
17. Shirreffs SM, Maughan RJ. The effect of posture change on blood volume, serum potassium and whole body electrical impedance. Eur J Appl Physiol
18. Shirreffs SM, Maughan RJ. Urine osmolality and conductivity as indices of hydration status in athletes in the heat. Med Sci Sports Exerc
19. Smith SM. Fast robust automated brain extraction. Hum Brain Mapp
20. Smith SM, Zhang Y, Jenkinson M, et al. Accurate, robust and automated longitudinal and cross-sectional brain change analysis. Neuroimage
21. Smith SM, Jenkinson M, Woolrich MW, et al Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage
22. Toffs PS. PD: proton density of tissue water. In: Toffs PS, editor. Quantitative MRI of the Brain: Measuring Changes Caused by Disease
. Malden (MA): Wiley; 2000. p. 85-110.
23. Turner M, McCrory P, Halley W. Injuries in professional horse racing in Great Britain and the Republic of Ireland during 1992-2000. Br J Sports Med
24. Wasserman LR, Yoh T, Rashkoff IA. Blood volume determination: comparison of T-1824 and P32 labelled red cell methods. J Lab Clin Med
25. Watson P, Black KE, Clark SC, Maughan RJ. Exercise in the heat: effect of fluid ingestion on blood-brain barrier permeability. Med Sci Sports Exerc
26. Whiting PH, Maughan RJ, Miller JDB. Dehydration and serum biochemical changes in marathon runners. Eur J Appl Physiol
27. Zhang Y, Brady M, Smith S. Segmentation of brain MR images through a hidden Markov random field model and the expectation maximization algorithm. IEEE Trans Med Imaging