What Extreme Military Training Does to the Human Body — The Science Explained

Navy SEAL candidates go through about five and a half days of near-continuous activity on fewer than four hours of sleep, a stretch known as Hell Week, and reported attrition for the full BUD/S course runs from 70 to 85 percent depending on the source. The courses are built to find the limit of what a body and mind can do under stress. Researchers have followed trainees through them with blood draws, saliva samples, brain scans, and injury surveillance. The findings are less cinematic than the legend, and more informative.

Educational content, not medical advice. Talk to your doctor before changing any medication, treatment, or routine. Information here reflects current research but every person's situation differs.

Key takeaways

• Extreme training stacks several stressors at once (exertion, sleep loss, energy deficit, and psychological pressure), which makes it hard to attribute any one effect to any one cause.
• In an eight-week Ranger course, soldiers lost 13 percent of body mass while testosterone and IGF-I fell and cortisol rose.
• Cortisol is not always elevated. A field exercise with about two hours of sleep a night showed lower testosterone but no rise in morning cortisol.
• Sleep deprivation raises the reactivity of the amygdala, the brain's threat detector, which helps explain why emotional control erodes.
• Injuries and exertional rhabdomyolysis (muscle breakdown) are the main documented medical harms, and recruits have by far the highest rhabdomyolysis rate.

What "extreme" training is designed to do

Military courses serve two overlapping purposes. Training builds skills and fitness. Selection deliberately creates conditions hard enough that some people fail, so that those who finish have shown something about how they function under pressure. Many of the most dramatic programs are mainly selection.

Four examples recur in the research literature, and the studies cited here come from very different settings. Comparing them requires care.

Setting	What the research describes	Source
Navy SEAL BUD/S, including Hell Week	About five and a half days of near-continuous training on under four hours of sleep; reported attrition 70 to 85 percent	Press and academic summaries; figures vary by source
Army Ranger training (eight-week course studied)	50 men, average daily energy deficit of about 1,000 kcal, 13 percent body mass loss	Nindl et al., 2007
Military survival school (SERE)	Mock prisoner-of-war interrogation of Special Forces and other soldiers; stress hormones measured	Morgan et al., 2000
Australian Army field exercise	30 trainees, eight days, average 2.2 hours of sleep over four nights	Tait et al., 2022

The rest of this article follows the body systems in turn: stress hormones, sleep and the brain, energy and muscle, the gut, and the injuries that make up most of the medical cost.

Stress hormones: two alarms with different speeds

The body has two main stress alarms. The first is the sympathetic-adrenal-medullary system, which releases adrenaline and noradrenaline within seconds and drives the classic fight-or-flight response. The second is the HPA axis (hypothalamic-pituitary-adrenal axis), a slower chain in which the brain signals the pituitary gland, which signals the adrenal glands to release cortisol. Cortisol mobilizes fuel and modulates inflammation, and it keeps the body ready over minutes to hours. For background on this hormone, see cortisol and the stress response and the psychology of stress.

An analogy: the first system is a smoke detector that screams instantly, and the second is the building's central heating, which responds slowly and stays on until switched off. A well-regulated person gets a rapid alarm and a prompt shutoff. Chronic stress is the heating stuck at maximum.

The SERE studies: a hormone that tracks performance

Some of the best-known human data come from Charles Morgan and colleagues, who studied soldiers at a U.S. Army survival school. One paper measured neuropeptide Y (NPY), a chemical messenger that is thought to dampen the stress response. Writing in Biological Psychiatry in 2000, the team sampled 49 soldiers before and 24 hours after training and a separate group of 21 during a mock prisoner-of-war interrogation. NPY rose during interrogation and was higher in Special Forces soldiers than in other soldiers. By 24 hours afterward, Special Forces levels had returned to baseline, while the other soldiers remained below it. NPY correlated with how soldiers behaved during interrogation and tended to correlate negatively with reported dissociation (the feeling of detachment from oneself or surroundings).

News coverage reported that all participants produced high cortisol, and that Green Berets produced more NPY than Rangers or Marines. It is tempting to conclude that training builds this resilience. The study design cannot say so. A 2003 Yale Medicine article noted the obvious question: either Special Forces soldiers always released more NPY, or their training enhanced it. A single observational comparison cannot separate selection from adaptation, which is a recurring problem in this field.

What the hormone studies do and do not show

The popular version of this story says trainees "recalibrate" their stress response until cortisol barely moves. The reviewed studies do not establish that picture. They show that the pattern depends heavily on the program, its duration, and when samples were taken. The next section shows two quite different results.

The hormonal bill of severe training

Eight weeks of Ranger training

In 2007, Bradley Nindl and colleagues followed 50 male soldiers through an eight-week military course, described in Medicine & Science in Sports & Exercise. The soldiers averaged a daily energy deficit of about 1,000 kcal. By the end, body mass had fallen 13 percent, fat-free mass 6 percent, and fat mass 50 percent. Circulating total testosterone and IGF-I (insulin-like growth factor I, a growth-promoting hormone) fell significantly, and cortisol rose. IGF-I and cortisol, but not testosterone, tracked how much tissue the men lost.

The authors concluded that with severe weight loss, IGF-I and cortisol correlate more closely with soft-tissue changes than testosterone does. In plain terms, the body was running a prolonged fuel shortage and shifting toward breaking down tissue for energy. This is the textbook picture of a catabolic state, and it is consistent with the general account of how hormones work and of how sex hormones work.

An eight-day field exercise

A very different pattern appeared in a 2022 study in PLOS ONE by Jamie Tait and colleagues. They tracked 30 Australian Army combat engineer trainees (27 men and 3 women) through an eight-day field exercise over a 16-day observation window. During the field phase, trainees averaged 2.2 hours of sleep over four nights, including at least one night of total sleep deprivation, and took about 20,246 steps a day, roughly double their usual.

Testosterone on waking was 28 percent lower on day 9 than on day 1, and bedtime testosterone was 36 percent lower. Testosterone recovered within about four days. Morning cortisol, surprisingly, was not higher than baseline, and the authors concluded the exercise did not appear to raise catabolic activity by that measure. The bedtime testosterone-to-cortisol ratio was 42 percent lower on day 9 and had not returned to its starting level by the end of the study (Tait et al., via PMC).

Study	Duration and design	Testosterone	Cortisol
Nindl et al., 2007	8-week course, 50 men, about 1,000 kcal/day deficit	Significantly lower	Significantly higher
Tait et al., 2022	8-day field exercise, 30 trainees, 2.2 hours of sleep per night over four nights	28% lower on waking; recovered within about four days	Morning levels not higher than baseline

The contrast is the point. A long, calorie-starved course produced a classic catabolic profile. A short, sleep-starved exercise produced a testosterone dip that reversed quickly, without the cortisol spike many people expect. Cortisol also varies through the day, and a spike can be missed by a handful of saliva samples. Reading "extreme training raises cortisol" as a universal rule would be an error.

Sleep loss and the thinking brain

Sleep restriction is a signature feature of selection courses. Its effects are among the best documented of any stressor, and they bear directly on judgment.

The amygdala runs hot

In a 2007 study published in Current Biology, Seung-Schik Yoo, Matthew Walker, and colleagues kept 26 healthy adults aged 18 to 30 awake for 35 hours, while a control group slept normally at night. Everyone then viewed 100 images from neutral to very negative inside a functional MRI scanner. The amygdala, the brain's threat-detection hub, was hyperactive in the sleep-deprived group. Walker said the emotional centers were over 60 percent more reactive than after a normal night's sleep. The paper's title describes a "prefrontal amygdala disconnect." The press summary described the sleep-deprived brain as routing emotional responses through the locus coeruleus, a brainstem region that releases noradrenaline, instead of through the prefrontal cortex (UC Berkeley release).

The practical reading is that sleep-deprived people feel more threatened and are less able to talk themselves down. The sample was small and the task was passive viewing, so the result describes a mechanism, not a measured effect on battlefield decisions.

Reduced activity in attention circuits

Positron emission tomography (PET) studies by Maria Thomas and colleagues, published in the Journal of Sleep Research in 2000, scanned people during sleep deprivation using a tracer of glucose metabolism. They concluded that short-term deprivation produced global decreases in brain activity, with larger reductions in the cortico-thalamic network that supports attention and higher-order cognition. For an accessible overview of how performance degrades with time awake, see what happens to your body with no sleep for 24 to 72 hours.

Caffeine in the field

One of the few controlled experiments run inside real selection training tested caffeine. In a 2002 study in Psychopharmacology, Harris Lieberman and colleagues randomly assigned 68 Navy SEAL trainees to 100, 200, or 300 mg of caffeine or placebo during a 72-hour period of sleep deprivation. Reviews of the work report that the 200 and 300 mg doses improved vigilance and reaction time and reduced sleepiness. A companion study found those doses shortened target-sighting time without improving accuracy.

That is worth stating because it cuts against a common intuition. The benefit was measurable for vigilance and speed, not for precision, and it is a statement about a specific dose under specific conditions, not general advice. Those wanting broader context on the compound can read how caffeine works in the brain.

Energy deficit, muscle, and the cost of getting lighter

Strenuous load-bearing training can trigger real physiological adaptation, including more mitochondria (the cell structures that make most of a cell's energy) in muscle, as described in how mitochondria became cell powerhouses and in how ATP energy works in cells. Adaptation, however, requires fuel and recovery. When the energy deficit is severe, the net result can be tissue loss.

The Ranger course illustrates this. A 13 percent drop in body mass, half of it fat mass and some of it fat-free mass, is a substantial change over eight weeks. Soldiers can lose weight and still perform, but the hormonal profile in the Nindl study signals a body under strain, not one in a state of thriving.

How muscle normally grows and repairs is covered in how muscles work and how muscles grow. The key principle is that stimulus plus recovery produces adaptation, while stimulus without recovery produces breakdown. Extreme courses are built to withhold recovery on purpose, which is why they select for tolerance rather than optimize for growth.

The gut under siege

Intense exertion is thought to divert blood from the digestive tract and can change how the intestinal lining behaves. In a 2017 study in the American Journal of Physiology: Gastrointestinal and Liver Physiology, J. Philip Karl and colleagues followed 73 Norwegian Army soldiers through a four-day, roughly 51 km cross-country ski march carrying 45 kg packs. Intestinal permeability, a measure of how leaky the gut barrier is, rose 62 percent, and the change correlated with plasma lipopolysaccharide (a bacterial component) and the inflammatory signal IL-6. Gut microbiota composition and metabolites also shifted significantly, and the researchers concluded the changes were likely not due to diet alone (report in Medical Xpress).

The study did not test Ranger or SEAL courses, and it cannot show that a leaky gut caused any particular problem. It does indicate that the gut is one of the systems that responds to multi-stressor training. For the fundamentals, see the gut microbiome explained.

Learned helplessness, control, and what resilience means

Psychologists often describe military selection in terms of learned helplessness, the finding by Martin Seligman and Steven Maier in the 1960s that animals exposed to uncontrollable shocks later failed to escape even when escape was possible. The original experiments are described in the Seligman learned helplessness experiment.

The neuroscience has been revised since. In a 2016 paper in Psychological Review, titled "Learned Helplessness at Fifty," Maier and Seligman argued that the original theory had it backward. Passivity in the face of prolonged uncontrollable stress, they proposed, is not learned. It is the default response, driven by serotonergic activity in a brainstem region called the dorsal raphe nucleus. What is learned is control: when an organism discovers it can influence the outcome, the medial prefrontal cortex detects that control and inhibits the dorsal raphe, which blocks the passivity.

This has an interesting implication for training design. If resilience depends on detecting control, the controllability of a stressor matters as much as its intensity. A hardship that feels random and endless may be a very different experience from one the person believes they can influence. Selection courses that give recruits clear tasks, visible progress, and some agency would, on this model, be building something different from those that impose arbitrary punishment. That is an inference from a laboratory model, and the studies reviewed here did not test it in recruits.

Stress inoculation: two meanings

The phrase stress inoculation is used in two ways. In clinical psychology, it names a structured therapy usually described in three phases: conceptualizing the problem, acquiring and rehearsing coping skills, and applying them to real stressful situations. Military settings have used it. In casual writing, however, it often means simply "being exposed to hard things repeatedly makes you tougher." The second, looser claim does not follow from the first. A course that is merely punishing is not stress inoculation in the therapeutic sense, and the evidence that repeated hardship alone builds lasting resilience is weaker than popular accounts suggest.

Why these studies are hard to compare

A reader skimming the numbers above might want to line them up and draw one curve. The designs do not allow it, and the limits are worth spelling out.

Different samples. The Ranger study followed 50 men. The Australian field exercise followed 30 trainees, only three of them women. The skiing study followed 73 soldiers, two of them women. Findings in largely male groups may not transfer to mixed groups, and sex differences in hormone responses and injury patterns are part of the reason female and male trainees are tracked separately in surveillance reports.

Different sampling. The Ranger study compared hormone levels before and after a course. The field exercise sampled saliva at waking, 30 minutes after waking, and at bedtime on selected days. Cortisol follows a daily rhythm that peaks shortly after waking, so one sampling schedule can miss a spike that another would catch. Saliva and blood also do not measure exactly the same fraction of a hormone.

Different stressors. A course with a severe, sustained calorie deficit and a course with sleep deprivation but adequate food are different experiments. Because real training bundles stressors, a researcher usually cannot say whether a hormone change came from sleep loss, hunger, exertion, cold, or worry. Laboratory studies isolate one factor, such as 35 hours without sleep, but give up realism.

No clean control group. Most field studies follow volunteers or whole classes, and few have a matched group of similar people who skip the training. Without that, regular changes such as time of day, season, or the simple passage of weeks are hard to separate from the course itself. This is the same problem that limits the neuropeptide Y comparison, where Special Forces soldiers and other soldiers differed before training began.

Survivor effects. People who quit or are injured out often drop from the final measurements. If the people with the most extreme stress responses leave early, the averages for those who remain look more resilient than the starting group was. That makes a "trainees adapt" conclusion easier to reach than it should be.

None of this makes the studies unreliable. It means each one answers a narrow question, and the honest summary is a set of patterns that recur across programs, not a single law of how bodies respond to extreme demands. The recurring patterns are lower testosterone, high energy cost, impaired emotional regulation after sleep loss, and a steady stream of overuse injuries.

Does training "rewire the brain"?

Neuroplasticity, the brain's capacity to change its connections with experience, is well established (see neuroplasticity). Articles on military training frequently claim specific consequences, such as a larger motor cortex or stronger connectivity in attention circuits. The studies reviewed for this piece measure hormones, sleep effects, body composition, gut physiology, and injuries. They do not directly image recruits' brains before and after selection. The sleep-loss and stress findings above come from laboratory and field studies with small samples, and the plasticity claims are better treated as plausible hypotheses than as demonstrated results.

The same caution applies to the opposite claim, that extreme stress shrinks the hippocampus or damages the brain permanently. Chronic stress effects are documented in animal work and in clinical populations, but a few weeks of selection training is not the same exposure, and the studies cited here do not settle it.

The medical cost: injuries, rhabdomyolysis, and attrition

Hormones and neurons make for fascinating science, but the documented medical harm of training is mostly mundane. It comes from overuse and from heat.

Musculoskeletal injuries

Army Public Health Center material indicates that more than 80 percent of recruit injuries involve the lower body, mainly the knee, ankle, foot, and lower leg, reflecting the amount of marching and running. Military women have more than twice the risk of stress fractures of military men, according to Army Public Health Center fact sheets. Injury has consequences for who finishes: injured trainees were five times more likely to leave basic combat training than uninjured trainees, and four times more likely to leave one-station unit training. Bone is a living tissue that remodels under load, and stress fractures occur when loading outpaces repair, a process described in how bones remodel throughout life.

Exertional rhabdomyolysis

Exertional rhabdomyolysis is pathological muscle breakdown during strenuous exercise, in which muscle contents leak into the blood. The standard diagnostic criterion is creatine phosphokinase (a muscle enzyme) at least five times the upper limit of normal after recent exercise. A 2025 surveillance report in the Medical Surveillance Monthly Report covered 2020 through 2024 among active component members of the U.S. Armed Forces.

Measure (2024)	Value
Total cases	464
Overall crude rate	35.9 per 100,000 person-years
Recruits	410.4 per 100,000 person-years, more than 13 times officers and other enlisted members
Marine Corps / Army	84.7 / 53.8 per 100,000 person-years
Air Force, Navy, Coast Guard	11.8 to 12.7 per 100,000 person-years

Named risk factors include sickle cell trait, prior heat illness or heat stroke, and beginning high-intensity activity at an unaccustomed intensity or duration, especially in heat. Men, people under 20, non-Hispanic Black service members, and those in combat-specific occupations had higher rates (MSMR, 2025).

The recruit figure is the most telling number in this article. The group that has just begun unaccustomed training, not the elite operators, carries the highest risk. The principle is the same as for stress fractures: a body asked to do much more than it is used to, too quickly, breaks down. Heat adds to it, because the body is also managing temperature, as covered in how the body regulates temperature.

Why some people finish and others do not

Selection courses record who quits, but the reasons are tangled. Some candidates leave because of injury, some because of motivation, and some because of ability. The attrition figures reported for BUD/S, 70 to 85 percent in the sources reviewed, mix all of these. Researchers at the U.S. Naval Academy and Duke have studied whether psychological traits such as grit and resilience, along with a limited set of biomarkers, relate to performance and completion, which reflects how unsettled the question is. The sources reviewed here do not identify any single hormone or personality score that reliably predicts who will finish.

The same caution about interpretation applies to the NPY result described earlier. A higher level in people who completed or excelled in a program could reflect pre-existing biology, training, or both. Studies that sample the same people before and after, with comparison groups, are the way to separate these.

Common misconceptions

Claim	What the evidence reviewed here supports
Extreme training always raises cortisol	Not always. A sleep-restricted field exercise showed no rise in morning cortisol, while an eight-week Ranger course showed a rise
Trainees "recalibrate" their stress response	Not established. The NPY data are observational and cannot separate selection from adaptation
Sleep loss only makes you tired	It also raised amygdala reactivity in one study and reduced activity in attention networks in others
Learned helplessness is something people learn	The 2016 revision argues passivity is the default and control is what is learned
The gravest risks are psychological	The best-documented medical harms are injuries and exertional rhabdomyolysis
Elite courses produce the highest injury rates	Recruits had the highest rhabdomyolysis rate, 410.4 per 100,000 person-years in 2024

Frequently asked questions

Q: What happens to hormones during extreme military training? A: It depends on the program. In an eight-week Ranger course, testosterone and IGF-I fell significantly and cortisol rose, alongside a 13 percent loss of body mass. In an eight-day field exercise with about 2.2 hours of sleep per night, testosterone dropped but morning cortisol did not rise, and testosterone recovered within about four days.

Q: How does sleep deprivation affect the brain? A: In a 2007 study, 35 hours without sleep made the amygdala over 60 percent more reactive to negative images, according to the lead author, with signs of weaker prefrontal control. PET studies found reduced activity in networks supporting attention. In a SEAL trial, 200 and 300 mg of caffeine improved vigilance and reaction time during sleep deprivation.

Q: Why do some candidates cope with stress better than others? A: The research is not settled. Special Forces soldiers showed higher neuropeptide Y during mock interrogation, but the study could not tell whether training or pre-existing biology explained it. Researchers have also studied traits like grit, though no single measure reliably predicts completion.

Q: Does military training cause lasting brain changes? A: The sources reviewed here did not directly image recruits' brains, so specific claims, such as an enlarged motor cortex, remain hypotheses. Neuroplasticity itself is well established, and sleep loss has documented short-term effects on brain activity.

Q: What are the most common medical problems? A: Musculoskeletal injuries dominate, with over 80 percent of recruit injuries affecting the lower body, and women have more than twice the stress fracture risk. Exertional rhabdomyolysis occurred at 410.4 cases per 100,000 person-years among recruits in 2024, versus 35.9 overall.

Q: Is this a guide to training harder? A: No. The studies describe what happened to people in supervised programs with medical support. They are not a basis for self-directed extreme exertion, sleep restriction, or dieting. Anyone considering demanding physical programs should discuss it with a qualified clinician.

The takeaway

The most robust finding across these studies is that the body pays for extreme training in ways that depend on the details: a long calorie deficit produces a catabolic hormone profile, a short sleep-starved exercise produces a quick testosterone dip that reverses, and a multi-day march makes the gut leakier. The brain results from sleep loss, especially the heightened amygdala response, explain why judgment and emotional control suffer. Popular claims about rewired brains and recalibrated stress systems run ahead of the evidence, while the documented medical burden is injury and muscle breakdown, concentrated among recruits who are new to the load. The most useful lesson for anyone outside the military is that adaptation needs recovery, and that courses are designed to withhold it.

Related Reading

• How Does the Human Eye Actually Work?
• The Science of Sleep — Why You Need It and What Happens Without It
• The Science of Mental Health — What We Actually Know
• How Does DNA Work? The Molecular Blueprint of Life Explained
• How Does the Human Heart Actually Work?