200,000 years of evolution built a radiator on your forehead. OMIUS makes it work better.

What the evidence supports:

• The forehead and upper face have the highest sweat gland density (~360 glands/cm²), are rich in arteriovenous anastomoses (specialized heat-exchange vessels), and are 2–5× more thermosensitive than any other body region for triggering whole-body thermoregulatory responses.
• The head can dissipate approximately 130 W of heat during moderate exercise, far more than the brain produces (~15–20 W).
• Head and face cooling during exercise consistently improves thermal comfort and, in many studies, extends exercise capacity, even without changing core body temperature.
• Insulating headgear, particularly larger coverings like wool caps, reduces the head's heat dissipation capacity.
• Evaporative cooling systems offer practical advantages over ice for sustained per-cooling during exercise: they are lighter, continuously replenishable, and on AVA-rich surfaces like the forehead, they operate at moderate temperatures that keep vascular heat-exchange channels open.

What remains an open question:

• Whether cooled blood from the scalp surface reaches the brain through emissary veins in quantities sufficient to lower brain temperature (the selective brain cooling hypothesis) is actively debated. The anatomical pathway exists and flow reversal during hyperthermia has been documented, but direct measurements of deep brain temperature have not shown a cooling effect from face fanning.
• Biophysical models predict that external cooling can only affect the outer ~1.5–3.6 mm of cortex. Whether this thin superficial gradient is functionally meaningful has not been tested in exercising humans.
• The prolactin suppression observed with face cooling may reflect direct cortical cooling, trigeminal nerve signaling, or both. The evidence does not yet distinguish between these explanations.

The forehead is uniquely designed for cooling

Three independent lines of evidence converge on the forehead as one of the body's most favorable surfaces for heat dissipation:

• Highest sweat gland density — Up to 360 glands/cm², far exceeding the torso or limbs.
• Most thermosensitive skin — In open-loop human experiments, the face was 2–5× more thermosensitive than any other body segment for sweating and thermal discomfort.
• Hairless and vascularized — Direct evaporation with minimal insulation.

The Anatomy Is Real

In 1985, Cabanac & Brinnel used Doppler recordings to show that during hyperthermia, blood in the mastoid and parietal emissary veins flows from skin to brain. In cold conditions, that flow slowed, disappeared, or reversed. The skull is not a sealed thermal barrier. Under heat stress, venous blood from the surface of the head can be routed inward through valveless emissary veins.

Zenker & Kubik later broadened that picture anatomically, showing that the scalp, face, dura, and cranial venous plexuses form a wider transport network — and that subarachnoid arterial branches in humans have diameters comparable to vessels found in the carotid rete of animals where selective brain cooling is well established.

The Debate: How Much Cooling Gets Through?

The existence of the anatomical route is not disputed. What is disputed is whether the thermal transfer through it is large enough to meaningfully cool the brain.

The case for selective brain cooling in humans was championed by Cabanac and colleagues over several decades. Their evidence rests primarily on the observation that tympanic temperature — measured at the eardrum, close to the brain — drops below core body temperature during hyperthermia with face cooling. They interpret this as evidence that cooled venous blood from the head surface reaches the intracranium and lowers brain temperature.

However, prominent researchers have challenged this interpretation. Nybo & Secher (2011) measured jugular venous blood temperature directly and found it was always hotter than arterial blood during exercise — even with face fanning — suggesting the brain is adding heat, not being selectively cooled. Simon (2007) reached a similar conclusion, arguing that tympanic temperature is not suited to indicate selective brain cooling in humans.

Perhaps the most striking data point comes from Shiraki et al. (1988), who obtained direct brain temperature measurements in a patient with an implanted drainage catheter. Face fanning did reduce tympanic temperature — but brain temperature was unaffected. It tracked esophageal temperature instead.

What the Measurements May Be Missing

There is a possibility that both sides of the debate are partially right — and that the disagreement stems from what is being measured rather than what is happening.

The evidence against selective brain cooling comes primarily from two kinds of measurement: jugular venous blood temperature and deep brain temperature. Jugular venous blood is a composite outflow from the entire brain, mixing drainage from deep structures heavily perfused by arterial blood with drainage from superficial cortical regions closer to the skull. A modest cooling effect at the cortical surface would be diluted into the jugular average and might not register.

Similarly, Shiraki's direct measurement was taken in the lateral ventricle — deep brain, surrounded by cerebrospinal fluid, thermally dominated by the ~750 mL/min of arterial blood perfusing the brain. A gradient of 0.2–0.3°C at the cortical surface a few millimeters from the skull and emissary veins would not necessarily appear in ventricular readings.

Biophysical modeling supports this concern quantitatively. Zhu et al. (2006) calculated the “temperature shielding length” of cerebral blood flow — the depth beyond which perfusion overwhelms any surface cooling effect. For normal adult gray matter perfusion, that shielding length is approximately 3.6 mm. In other words, external cooling is physically predicted to affect only the outer few millimeters of cortex before being washed out by arterial blood flow. Deeper measurements — jugular venous, ventricular, or even large MRS voxels — would not detect a gradient confined to this thin zone.

Nelson & Nunneley (1998), the most frequently cited modeling paper against head cooling, reached a similar conclusion: surface cooling affects only the most superficial cerebrum (≤1.5 mm). The authors judged this too thin to matter clinically — but for the present question, the point is that even their model predicts a gradient exists; it simply falls below the resolution of conventional measurements.

Direct human evidence confirms that the brain is not thermally uniform. Stone et al. (1997) measured temperature at multiple depths in the temporal lobe of 39 neurosurgical patients and found a 1.4°C drop between 2 cm and 1 cm depth(P < 0.001), with temperatures falling further toward the cortical surface. This was under operative conditions (exposed brain), but it demonstrates that the cortical surface is a thermally distinct compartment from deeper structures — a fact that any deep-brain measurement would miss entirely.

In other words, the existing measurements may be answering the wrong question. They convincingly show that face cooling does not lower deep brain temperature. But they do not rule out a thermal gradient at the cortical surface — and the cortical surface is where much of what matters for exercise performance actually happens.

This is speculative — no study has directly measured cortical surface temperature gradients during head cooling in exercising humans with intact skulls. But it is a physiologically coherent hypothesis that is consistent with the anatomical evidence (emissary veins connect surface vasculature to the dura and cortical venous drainage), the biophysical modeling (Zhu 2006, Nelson & Nunneley 1998), the neurosurgical gradient data (Stone 1997), and the performance data (head and face cooling consistently improve exercise capacity in ways that may not be fully explained by changes in core temperature alone). Future research using high-resolution MR thermometry could potentially resolve this question.

The OMIUS Headband: What Has Been Tested

One peer-reviewed study has directly evaluated the OMIUS headband. Jolicoeur Desroches et al. (2024) tested 10 trained male runners in a crossover design at 35°C and 56% humidity, comparing the OMIUS headband against a sham. The headband produced significant forehead cooling (p = 0.007) and significantly improved forehead thermal comfort (p = 0.008) during 70 minutes of submaximal running. However, in a subsequent 5-km time trial, no statistically significant performance improvement was found.

Several design features of the study are important for interpreting this result. The headband was not re-wetted during the time trial, a limitation the authors themselves acknowledged. Baseline rectal temperatures were 0.11°C higher in the OMIUS condition before headband placement, a pre-existing difference unrelated to the headbands.

And with 10 participants, the study's minimum detectable effect was approximately 5%, larger than any legal ergogenic aid has demonstrated in running TT research. The observed trend was 2.2% (24 seconds), consistent with other cooling interventions but below the study's detection threshold. The study supports local cooling and improved comfort; performance effects under that protocol remain uncertain and warrant larger trials.

Why This Pattern Matters

The recurring finding across these studies is striking: head and face cooling improves performance without changing core temperature. This creates an apparent paradox: how can cooling the face improve endurance by 50% if it doesn't cool the body?

Three explanations are possible, and they are not mutually exclusive:

First, thermoregulatory signaling. As discussed in Section 01, the face has 2–5× the thermosensitivity of other body regions. Cooling it reduces perceived thermal strain, which in turn modulates voluntary pacing and extends exercise tolerance (Cheung 2010). This mechanism is well established.

Second, cortical surface cooling. As discussed in Section 02, existing measurements of brain temperature (jugular venous blood, ventricular probes) capture deep brain temperature but may miss thermal gradients at the cortical surface, where the motor cortex, prefrontal cortex, and insular cortex reside. The Ansley prolactin finding is noteworthy in this context: prolactin release is driven by hypothalamic serotonergic pathways, and its abolition by head cooling without any change in core temperature is consistent with, though does not prove, a direct thermal effect on brain tissue. One possible interpretation is that subtle cooling of the cortical surface, invisible to jugular or ventricular measurements, could play a role. However, a purely neural explanation is at least equally plausible: Mündel et al. (2006) showed that face cooling abolishes the prolactin response even during passive heating without exercise, suggesting the effect may be mediated by facial skin thermoreceptors signaling through the trigeminal nerve to the hypothalamus, rather than by a cortical thermal gradient. Both mechanisms remain open, and they are not mutually exclusive.

Third, combined effects. Real-world head cooling likely operates through multiple pathways simultaneously: real heat removal from the head surface, thermoregulatory signaling from facial thermoreceptors, and possibly a degree of cortical cooling that current measurement methods cannot resolve.

For OMIUS, the practical conclusion is the same regardless of which pathway dominates: head and face cooling during exercise can produce meaningful performance benefits in some settings, and amplifying the forehead's natural evaporative capacity is a physiologically sound way to deliver that cooling.

The AVA Question

Stanford researcher Craig Heller identified a specific mechanism relevant to cooling at AVA-rich sites like the palms, soles, and upper face. Arteriovenous anastomoses, the specialized shunts that allow rapid heat exchange, close when local skin temperature drops below approximately 24°C. When ice is applied directly to these sites, the very vessels designed for high-volume heat transfer constrict, reducing the body's ability to move heat from the core to the skin surface.

This does not mean ice is counterproductive everywhere on the body. An ice vest on the torso — where AVA density is low and the cooling mechanism is primarily conductive heat absorption rather than vascular heat exchange — operates through a different pathway. But on AVA-rich surfaces like the forehead, the distinction matters: intense cold can shut down the local heat-exchange architecture that makes these sites so effective in the first place.

The Thermodynamics: Not as Simple as It Looks

A common comparison is that evaporating water removes far more heat per gram than melting ice. The raw numbers support this: the latent heat of vaporization (~2,430 J/g) dwarfs the latent heat of fusion (~334 J/g) plus warming of meltwater to skin temperature (~138 J/g), giving ice a total of roughly 472 J/g through the melting-and-warming phase alone. On that basis, evaporation removes about 5× more heat per gram.

But this comparison is incomplete. Ice that melts produces water, and if that meltwater then evaporates from the skin, the total cooling potential of 1 gram of ice is approximately 2,900 J/g (fusion + warming + evaporation), which is actually slightly more than 1 gram of room-temperature water evaporated (~2,480 J/g).

In practice, however, meltwater from ice vests and ice packs rarely evaporates in situ. It drips off, soaks into clothing, or pools, meaning the evaporative phase is largely lost. This is where the practical advantage of purpose-built evaporative systems becomes real: a high-surface-area material like the OMIUS graphite matrix is designed specifically to hold water in contact with skin and expose it to airflow, maximizing the fraction that actually evaporates rather than running off.

The real advantages of evaporative per-cooling over ice are therefore less about raw thermodynamics and more about the use case: evaporative systems are lighter, continuously replenishable from aid stations, and on AVA-rich surfaces like the forehead they operate at moderate temperatures that keep vascular heat-exchange channels open rather than triggering vasoconstriction.

Amplified Evaporation:The hydrophilic graphite matrix wicks sweat away from the skin and spreads it across a larger surface area, increasing the rate of evaporative cooling. Evaporation is the body's most powerful heat-removal pathway during exercise, and the forehead has the highest sweat gland density of any body region. OMIUS increases the effective evaporative surface rather than trapping sweat under insulating fabric.

Conductive Heat Transfer via AVAs:The forehead is rich in arteriovenous anastomoses — the same specialized heat-exchange vessels identified in the palms, soles, and upper face. OMIUS's thermally conductive graphite contact points are designed to interface directly with this AVA-rich skin, drawing heat conductively without dropping skin temperature below the ~24°C threshold where AVAs constrict — intended to sustain heat extraction within the physiological range where vascular heat exchange remains active.

Thermoregulatory Signaling:Cotter & Taylor showed that the face is 2–5× more thermosensitive than any other body region for triggering whole-body sweating and thermal discomfort responses. Cooling the forehead does not just remove heat locally — it sends a disproportionately strong signal to the thermoregulatory system that the body is managing its heat load.

A Note on Deeper Cooling

As discussed in Section 02, some researchers have proposed that cooled venous blood from the head surface may travel inward through emissary veins to cool brain-adjacent circulation, a mechanism known as selective brain cooling. The anatomical route exists; biophysical models predict a thin cortical gradient from external cooling (Zhu et al. 2006), and the hypothesis has not been ruled out by existing measurements, which target deep brain structures rather than the cortical surface. If this mechanism is active, it would represent an additional benefit of forehead cooling beyond the three established pathways above. But the core case for OMIUS does not depend on it. Evaporation, conduction, and thermoregulatory signaling are each independently supported and together provide a strong physiological rationale for enhanced forehead cooling during exercise in the heat.

The human forehead is not an accident of anatomy. It sits at the convergence of three independent physiological systems (high-density sweat glands, AVA-rich vasculature, and disproportionate thermoregulatory influence) that together make it one of the most effective heat-dissipation surfaces on the body. Whether it also serves as a gateway for cooling the brain's cortical surface is plausible and supported by the underlying anatomy and biophysics, but remains unproven in exercising humans.

What is clear is that head and face cooling during exercise works. Across dozens of studies and a 2025 meta-analysis of 63 controlled trials, cooling this region consistently reduces thermal discomfort and, in many protocols, extends exercise capacity, often without measurably changing core body temperature. The mechanisms behind this effect likely include real heat removal from the head surface, thermoregulatory signaling from facial thermoreceptors, and possibly a degree of superficial cortical cooling that current measurement methods have not been designed to detect.

The scientific debate around selective brain cooling has been active for over four decades, and this article does not pretend to resolve it. What we have tried to do is present both sides honestly: the anatomical evidence for an inward cooling pathway, the direct measurements that challenge it, the biophysical models that predict a thin cortical gradient, and the performance data that do not fit neatly into either camp. We believe that engaging openly with scientific uncertainty is more credible than ignoring it, and more useful to the athletes, coaches, and medical professionals who are evaluating this technology.

OMIUS is built around a simple premise: the forehead's cooling capacity should be amplified, not blocked. Conventional headgear insulates. OMIUS is designed to do the opposite: maximizing evaporative surface area through thermally conductive graphite, sustaining vascular heat exchange by operating within the temperature range where AVAs remain open, and leveraging the face's outsized influence on thermoregulatory drive. These mechanisms are consistent with established physiology. Whether future research reveals additional benefits at the cortical level, the foundational case stands on what is already known.