Every cold plunge protocol you have ever read specifies two things: temperature and duration. Drop to 10°C. Stay for two minutes. Maybe three. These are the dials everyone adjusts, the numbers everyone debates. But almost no protocol specifies a third variable: how deep you immerse. A 2021 study using polysomnography suggests that ice bath depth may affect your sleep architecture more than either of them.
Conducted at the French National Institute of Sport (INSEP), the study found that athletes who immersed their whole body in cold water produced more slow-wave sleep in the first three hours of the night than those who sat in the same water, at the same temperature, for the same duration, but only up to the waist. Same cold. Same clock. Different depth. Different sleep.
The bridge between cold water and slow-wave sleep
Your body’s core temperature follows a circadian rhythm, peaking in the late afternoon, declining through the evening, bottoming out in the early hours. Sleep onset, and particularly the onset of deep slow-wave sleep, tracks this decline. The steeper and smoother the drop, the more readily the brain enters slow-wave stages. A 2019 meta-analysis in Sleep Medicine Reviews confirmed the principle from the warm side: passive body heating one to two hours before bed shortened sleep onset latency by roughly ten minutes, because warming the periphery accelerated the subsequent core-temperature drop. Cold immersion works the same mechanism from the opposite direction, forcing a rapid core-temperature decline during and immediately after the bath.
What the INSEP study actually found
Mathieu Nedelec, a researcher in recovery and sleep at INSEP, led the centrepiece study with his colleagues. Their 2021 study remains the only published trial directly comparing immersion depth’s effect on sleep architecture using polysomnography: electrodes tracking brain waves, muscle activity, and eye movement through the entire night.
Twelve trained male runners completed a simulated trail run in heat, then underwent one of three recovery conditions in randomised order: whole-body cold water immersion at 13.3°C for ten minutes (water to the neck), partial immersion at the same temperature and duration (water to the iliac crest, roughly hip level), or a seated control with no immersion.
Both immersion conditions reduced sleep disruptions compared with no immersion — fewer arousals, better continuity. But the depth-dependent effects separated the two groups. Whole-body immersion produced a lower nocturnal core temperature and a higher proportion of slow-wave sleep in the first 180 minutes of the night compared with partial immersion. Only whole-body immersion significantly decreased limb movements during sleep. Nedelec’s group concluded that whole-body immersion “may be particularly useful in the athlete’s recovery process,” with the sleep-architecture improvements concentrated in exactly the period where the body does its deepest restorative work.
What the study does not prove
One study. Twelve participants, all male, all trained runners, all measured after exercise in heat. Whether the same depth effect appears in women, in untrained individuals, or in people immersing without prior exercise has not been tested with polysomnography. The effect sizes are meaningful but the sample is small. No one should treat this as settled science.
What it does provide is a clear, well-controlled signal that depth is an independent variable, not just a proxy for ‘more cold’, with its own consequences for sleep. And when you examine the mechanisms, the signal becomes considerably more persuasive.
The head-cooling surprise
Why depth matters becomes clearest not in the sleep study itself but in a separate line of research that explains the physiology underneath it.
A 2006 study from Gordon Giesbrecht’s Laboratory for Exercise and Environmental Medicine at the University of Manitoba measured what happens when you add head submersion to cold water immersion. Your head represents roughly 7% of total body surface area. By the arithmetic of heat exchange alone, submerging it should increase cooling by a modest, proportional amount.
It does not. Head submersion increased the core cooling rate by 42%.
That figure, wildly out of proportion to the surface area involved, points to something beyond simple thermal conduction. Giesbrecht’s team attributed the effect to a redistribution of blood flow triggered by stimulation of the trigeminal nerve receptors on the face and scalp. When the head is submerged, the body does not just lose heat faster through the skin; it redirects warm blood from the core toward the cooling surface in a way that amplifies the thermal decline far beyond what the physics of skin exposure would predict.
For the sleep question, here is what that means. If core-temperature decline drives slow-wave sleep, and head immersion accelerates core cooling by 42%, then the difference between neck-deep and shoulder-deep is not trivial. Those last few centimetres of water, covering the upper trapezius, the base of the neck, and potentially contacting the lower face, may contribute more to the sleep benefit than all the water below the ribcage combined.
The dive reflex: a second depth-dependent mechanism
Core cooling is not the only thing that changes when water reaches the face.
When cold water contacts the forehead and the area around the eyes and nose, it triggers a trigeminal-vagal response known as the diving reflex. Heart rate drops, peripheral vessels constrict, and cardiac vagal activity (a marker of parasympathetic dominance) increases. A 2023 systematic review in Psychophysiology confirmed that the response produces moderate to large increases in cardiac vagal activity during exposure.
This matters for sleep because the transition from wakefulness to sleep is, among other things, a transition from sympathetic to parasympathetic dominance. A cold plunge morning or night that activates the dive reflex primes the autonomic nervous system for the state it needs to enter at bedtime. You don’t need to hold your head underwater for ten minutes to trigger it. A few seconds of deliberate face submersion, or even splashing cold water across the forehead at the end of a session, is sufficient based on the available evidence.
A practical depth hierarchy
A clear hierarchy of immersion depth emerges from the evidence, and the returns aren’t linear. The highest-value regions are at the top.
Head and face contact sits at the apex. Even brief submersion activates the dive reflex and, if sustained, unlocks the outsized core-cooling effect documented by Giesbrecht’s lab. No other depth increment matters as much. It is also the least comfortable, which is why most people skip it.
Neck and shoulders come next. Submerging to the neck ensures that the large blood vessels of the cervical region are fully exposed to cold, maximising convective heat loss from the blood returning to the core. Chauvineau’s whole-body condition immersed participants to this level, and it was this condition that produced the superior slow-wave sleep.
Chest-level immersion covers a large surface area and adds hydrostatic pressure, the physical compression of water against the torso that aids venous return and fluid redistribution, a mechanism that Shona Halson, a recovery scientist at Australian Catholic University, has noted is absent from showers. But the incremental cooling contribution per centimetre of depth is smaller here than at the neck and shoulders. Better than waist-deep; not as consequential as shoulder-deep.
Waist-level immersion is where Chauvineau’s partial condition sat, and where many commercial tubs and DIY setups leave people in practice, because seating position raises the body and water volume is limited. This level still improves sleep continuity versus no immersion, but the ice baths and sleep four-week threshold advantage was absent.
Limbs only — feet and lower legs — offer minimal core-cooling stimulus. Pleasant, possibly useful for localised inflammation, but not enough to shift sleep architecture.
What this means for your practice — and your tub
If you are cold-plunging for sleep, your waterline matters at least as much as your thermometer.
Brief face contact is the highest-value addition. A deliberate dip, face to the waterline for five to ten seconds at the end of your immersion, activates the dive reflex and may trigger the accelerated cooling response. A few seconds of discomfort for what the physiology suggests is a disproportionate return.
At Latitude Zero, a surf resort in the Mentawai Islands where we installed cold plunge systems, the recommended post-surf protocol calls for immersion to the shoulders at minimum — the threshold the INSEP data identified as the boundary between partial and whole-body benefit. Anything shallower, and guests consistently report reduced effects: less of the calm, settled feeling in the hours afterward, and less of the deep sleep quality they came to associate with their evening plunge. The evidence is observational, not polysomnographic, but across hundreds of sessions the pattern aligns with what Nedelec’s group measured in the lab.
A cold plunge vessel’s internal depth, seating height, and water volume determine whether a user can realistically immerse to the shoulders. A tub that looks generous but seats the user on a raised bench, or holds too little water to reach the collarbones, is designing out the most evidence-supported depth range. For anyone specifying a cold plunge installation, at home or in a hospitality setting, internal water depth isn’t an aesthetic decision. It’s a protocol decision.
A note on cold showers and sleep: a shower delivers neither consistent temperature (a cold shower in northern England is a genuinely different stimulus from one in Bali, where mains water runs tepid year-round) nor hydrostatic pressure, nor any realistic opportunity for face immersion. Better than nothing for the temperature signal, but not a substitute for immersion if sleep architecture is the goal.
The waterline you have been ignoring
For years, the cold plunge world has refined how cold and how long while largely ignoring how deep. The evidence is early: one well-designed polysomnography study, a handful of mechanistic investigations, a meta-analysis confirming the depth pattern across recovery outcomes. It is not a closed case, and the next few years of research will either strengthen the signal or complicate it. But the mechanisms are coherent, the practical cost of acting on them is zero, and the direction is clear.
Set your waterline next time you set your chiller temperature. Shoulders at minimum. Face if you can manage it. The best time to cold plunge may depend on your goals, but the depth you choose shapes whether you reach them. Everyone watches the thermometer. The depth of the water may be doing more of the work.
