StimulHeat: Engineered Thermal Stimulation

• StimulHeat is a class of thermal stimulation methods that actively control heat flow to modulate perception and cellular responses.
• These systems integrate diverse modalities like thermoelectrics, optical sources, ultrasound, and convection to generate rapid thermal transients.
• Research in StimulHeat spans VR feedback, neural modulation, and trans-scale bio-signaling, emphasizing precise control and safety.

Searching arXiv for papers on “StimulHeat” and closely related thermal stimulation systems. StimulHeat denotes a class of thermal stimulation paradigms in which controlled heat transfer is used as an active signal rather than as a passive environmental condition. Across the cited literature, the term is associated with systems that modulate perception, comfort, affect, neural activity, cellular signaling, or material response by driving thermal energy through thermoelectric devices, optical radiation, airborne ultrasound, convection, biochemical heat release, or electromagnetic absorption. The underlying premise is that thermal stimulation can be engineered across scales—from molecular reactions and protein heating to body-worn interfaces and virtual reality controllers—provided that the relevant transport coefficients, control variables, and perceptual or physiological thresholds are explicitly modeled and monitored (Mesnage et al., 5 Sep 2025).

1. Conceptual scope and physical basis

At its broadest, StimulHeat is anchored in the distinction between temperature as a state variable and heat flow as a control variable. The controller-integrated VR device "StimulHeat: a Low-Energy Wearable Thermal Feedback Device Using Peltier Elements with Heat Flow Controlled Loop for Hand Interactions in Virtual Reality" formalizes this distinction by treating the thermoelectric device as a current-controlled heat pump/heater and by offering both heat-flow and temperature control modes (Mesnage et al., 5 Sep 2025). In parallel, the review on trans-scale thermal signaling defines thermogenesis as “a physiological activity of internally releasing heat” and frames local heat release from biochemical reactions and supramolecular reconfiguration as a biologically consequential signal that propagates across molecules, cells, tissues, and organisms (Kurisaki et al., 2023).

Several governing relations recur across the literature. Heat conduction is repeatedly described through Fourier’s law, $\vec{q} = -k \nabla T$, and through diffusion equations of the form $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$ or closely related variants (Kurisaki et al., 2023). Thermoelectric implementations use absorbed-side or cold-side heat expressions in which current, electrical resistance, Seebeck coupling, and interfacial temperature difference jointly determine instantaneous heat flow (Mesnage et al., 5 Sep 2025). Non-contact systems based on ultrasound or optical stimulation also reduce to volumetric or boundary heat-source models in which dissipated acoustic or optical power is mapped into transient tissue heating (Iwabuchi et al., 2024). This shared formalism is one reason the same label can span such different apparatuses.

A second cross-cutting principle is temporal asymmetry. Thermal stimulation is effective not only when it changes absolute temperature, but also when it controls the rate of change. The photo-thermal neural excitation literature proposes a temperature-rate model in which rapid temperature changes induce a depolarizing transmembrane current proportional to $dT/dt$ (Farah et al., 2012). The non-contact cold-sensation work likewise exploits asymmetry by alternating perceptible cooling and imperceptible warming on the same skin area while keeping average skin temperature nearly constant (Xu et al., 2023). This suggests that StimulHeat is often less about maintaining a static thermal state than about shaping transient thermal trajectories with specific perceptual or biophysical consequences.

2. Device architectures and control modalities

Human-scale StimulHeat systems span contact thermoelectrics, optical emitters, airborne ultrasound, and hybrid convection-radiation displays. The clip-on Valve Index accessory uses an ET-071-08-15 thermoelectric device, a continuous bidirectional current driver, BLE control, and selectable heat-flow or temperature loops in a palm-contact geometry (Mesnage et al., 5 Sep 2025). The dual-sided wearable Peltier architecture uses eight TES1-4902 modules mounted to 270-degree rotation servos so that preheated or precooled faces are mechanically swapped to the skin, allowing perceptual switching to be limited by motor motion rather than by the thermoelectric ramp itself (Kang et al., 2024). The in-ear Heatables prototype instead uses 810 nm and 950 nm LED channels, pulse-width modulated at 100 Hz with a 59% duty cycle, to deliver localized optical heating in the auditory canal (Zitz et al., 3 Jun 2025).

Non-contact StimulHeat implementations emphasize low residual heat and rapid reversibility. Airborne ultrasound can generate both temperature rise and acoustic-radiation-pressure touch in the same focus; a 12-array AUTD3 system produced a palm temperature increase of $5.4\,^\circ\mathrm{C}$ after $5.0\,\mathrm{s}$ in static focus mode and $4.5\,^\circ\mathrm{C}$ after $5.0\,\mathrm{s}$ with 50 Hz amplitude modulation at duty ratio $0.9$ (Iwabuchi et al., 2024). A different non-contact architecture integrates convection via cold air and radiation via visible light; by alternating a cooling-only phase with a cooling-plus-light-warming phase, it maintains near-constant skin temperature while eliciting a persistent cold percept (Xu et al., 2023). HeatFlicker extends non-contact thermal output into apparent motion by pairing visible-LED heating with asymmetric fingertip vibration that produces a pulling illusion, so that heat from a fixed source appears to flicker and shift directionally (Ito et al., 2024).

The choice of control variable is consequential. In the Valve Index system, heat flow is treated as a directly controllable quantity because of the continuous relationship between injected current and thermoelectric transport (Mesnage et al., 5 Sep 2025). In the dual-sided wearable, by contrast, response speed is achieved mechanically: the system avoids waiting for the Peltier interface to traverse a long thermal time constant and instead rotates a preconditioned face into contact (Kang et al., 2024). The non-contact cold display adopts yet another strategy, using calibrated duty-to-rate mappings for cold air and visible light so that cooling and warming are matched over each sub-second cycle (Xu et al., 2023). These designs are technically distinct, but each treats thermal stimulation as a closed-loop or quasi-closed-loop signal-delivery problem rather than as uncontrolled heating.

3. Perception, immersion, and personal comfort

A substantial branch of StimulHeat research concerns perceptual engineering in VR, XR, and personal comfort systems. The controller-mounted StimulHeat user study tested heat-flow mode during repeated grasp interactions and reported 190 correct HOT/COLD classifications out of 192 such trials, with only 3 false positives in 96 NEUTRAL trials (Mesnage et al., 5 Sep 2025). This is a narrow but important result: the delivered cue was discriminable at the controller interface without making the clip-on shell perceptually intrusive. The paper also states that thermal perception is not entirely determined by surface temperature and identifies material thermal effusivity as relevant to perceived warm/cool onset (Mesnage et al., 5 Sep 2025).

Other studies use different actuators but address the same question of how localized thermal input modulates experience. The VR crossmodal study found that a spot projector was more efficient than a fan heater to create a warm sensation, and that fan plus water spray was more efficient than fan alone to create cold sensation; visual cues did not significantly alter thermal perception except in the extremely cold condition combining snow visualization with fan plus water spray (Helfenstein-Didier et al., 2023). HeatFlicker adds a motion component rather than an intensity component, attempting to reproduce the unsteady thermal character of a campfire through synchronized thermal, tactile, and auditory cues (Ito et al., 2024).

Personal comfort systems generalize StimulHeat beyond VR. Heatables, an in-ear NIR-IR-LED device evaluated in a placebo-controlled study with 24 participants exposed for 150 minutes in an approximately $17.5\,^\circ\mathrm{C}$ office, significantly increased perceived temperature at the ear, wrist, and ambient locations, with ambient and wrist increases reported as approximately $+1.2\,^\circ\mathrm{C}$ and with significant gains in whole-body warmth, comfort, and thermal acceptability (Zitz et al., 3 Jun 2025). The heated-garment study in a $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$0 climate chamber reached a different use case but a similar conclusion: segmented local heating improved overall TSV by 1.50 and TCV by 1.53 during the self-adjustment phase, while preferred surface temperatures differed widely across segments and participants (Ju et al., 2022). The common implication is that localized thermal stimulation can shift global thermal appraisal without conditioning an entire room or suit uniformly.

A frequent misconception is that thermal comfort or thermal realism reduces to maximizing local temperature rise. The literature does not support that simplification. The non-contact cold display shows that continuous cold can be perceived without significantly altering skin temperature, provided the cooling rate and cooling time ratio fall in specific ranges (Xu et al., 2023). Heatables shows whole-body comfort gains from a binaural electrical input of approximately 960 mW rather than from large body-surface temperature changes (Zitz et al., 3 Jun 2025). This suggests that StimulHeat systems often operate by biasing perceptual and thermoregulatory inference, not merely by imposing large thermal loads.

4. Neural and affective modulation

Thermal stimulation is also used as a neural or affective control signal. The EEG study on imagined temperature sensations tested 15 healthy right-handed adults and showed sustained mu-ERD during both real thermal stimulation and imagined temperature sensation over the full 4 s task interval, localized contralaterally over the sensorimotor hand area, prominently at C3 and neighboring electrodes (Belichenko et al., 22 Aug 2025). Direct comparisons between real and imagined conditions yielded slightly larger ERD for real stimulation, but the differences were not significant: TS_hot vs ITS_hot had $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$1, and TS_cold vs ITS_cold had $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$2 (Belichenko et al., 22 Aug 2025). Because both TS and ITS differed robustly from rest, the paper argues that temperature imagery is a viable non-motor BCI modality.

The same study explicitly frames this possibility as a “StimulHeat” system with two complementary control axes: real thermal stimulation as a calibration or assisted channel, and imagined temperature sensation as an autonomous channel (Belichenko et al., 22 Aug 2025). Concrete decoding features proposed in the paper include mu-band power suppression over the contralateral central cluster, activation latency beginning around 0.5 s after cue onset, and sustained engagement over the 4 s window (Belichenko et al., 22 Aug 2025). A plausible implication is that StimulHeat in this context refers not to a wearable heater but to a neurophysiological command space built from thermally grounded sensorimotor imagery.

Thermal stimulation also modulates affect. In a controlled study with 120 naive participants exposed to constant upper-back temperatures of $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$3, $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$4, or $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$5 while viewing IAPS images, warmth produced a main effect on arousal, $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$6, and a main effect on valence, $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$7 (Hojatmadani et al., 2023). Warm participants reported higher arousal than cool participants in some positive or neutral low-to-medium-arousal image categories, while valence was decreased in warm conditions relative to neutral (Hojatmadani et al., 2023). The Geneva Emotion Wheel did not show significant temperature effects. This directly counters a common intuition that mild warmth is uniformly pleasant; in this experiment, warmth elevated arousal yet slightly reduced valence relative to neutral.

The neural-stimulation literature extends these observations into direct excitability control. Localized photothermal heating using 50 ms, 650 nm laser pulses focused onto carbon particles or dye at the membrane triggered action potentials in leech neurons with 70–90% reliability and produced transmembrane currents in Xenopus oocytes, supporting a non-voltage-gated-channel mechanism (Migliori et al., 2012). The complementary temperature-rate model formalizes the hypothesis as $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$8 and shows that stimulation thresholds predicted from this model agree with both PAINTS cortical-culture data and earlier INS auditory-neuron data (Farah et al., 2012). In this branch of the literature, StimulHeat denotes rapid, localized thermal transients that act as electrically effective perturbations.

5. Cellular, molecular, and trans-scale signaling

At the cellular and molecular scales, StimulHeat becomes a problem of thermogenesis, transport coefficients, and localized response. The review on trans-scale thermal signaling identifies ATP/GTP hydrolysis and multimeric complex formation or disassembly as microscopic heat sources and highlights simulation toolkits such as QM/MM, SF2MD, hMC/MD, WE, PaCS, equilibrium MD, and NEMD for quantifying their consequences (Kurisaki et al., 2023). Specific quantitative examples include actin ATP hydrolysis with $$\partial T / \partial t = \alpha \nabla^2 T + Q/(\rho c_p)$$9 kcal/mol in simulation, myosin II pathways with activation barriers of approximately 10 or 20 kcal/mol depending on the mechanism, and Ras SF2MD simulations in which approximately 5 kcal/mol is stored in the P-loop while kinetic-energy directionality is lost within approximately 1 ps (Kurisaki et al., 2023). The article’s core claim is that microscopic heat release couples to mesoscopic biological processes through thermal conductivity, interfacial thermal conductance, heat capacity, density, and geometry.

Direct experimental thermal manipulation of cells corroborates this view. The diamond heater-thermometer study used a single polycrystalline CVD diamond particle containing SiV centers as both heater and thermometer; local heating of 11–12 $dT/dt$0 above ambient next to HeLa cells and mouse hippocampal neurons altered intracellular calcium dynamics (Romshin et al., 2022). In HeLa cells, a long-term increase in Fluo-4 NW fluorescence by about three times over about 30 s was observed, while in neurons heating caused a calcium surge of about 30% with duration about 0.4 ms (Romshin et al., 2022). The heated footprint was reported as approximately $dT/dt$1, and the modeled cell-surface temperature was approximately $dT/dt$2 when the heater was at $dT/dt$3 and positioned about $dT/dt$4 away (Romshin et al., 2022). This is a canonical StimulHeat implementation at subcellular resolution.

Mitochondrial thermogenesis supplies an endogenous version of the same idea. Using a diamond thermometer based on SiV zero-phonon-line shifts, heat release from isolated mouse brain mitochondria during CCCP-induced uncoupling was measured directly; local temperature rose by $dT/dt$5–$dT/dt$6 above ambient, with an absolute maximum of $dT/dt$7 (Romshin et al., 2022). Burst amplitudes saturated with duration according to a simple law with $dT/dt$8 and $dT/dt$9 (Romshin et al., 2022). The thermocouple remained near ambient, indicating local rather than bulk heating.

THz stimulation extends the concept to protein-scale hotspots. The THz protein-interaction study models proteins as thermal nanosensors whose resonant absorption can raise local temperature and thereby alter the opening probability of thermally gated ion channels (Elayan et al., 2024). For lysozyme-like parameters, COMSOL simulations showed that increasing incident power from 200 nW to 400 nW for a 1 ms pulse increased temperature by about $5.4\,^\circ\mathrm{C}$0 across a $5.4\,^\circ\mathrm{C}$1 patch, and the paper uses a denaturation-avoidance constraint of $5.4\,^\circ\mathrm{C}$2 (Elayan et al., 2024). The mechanistic continuity with larger-scale StimulHeat systems lies in the fact that both domains treat thermal control as programmable energy deposition constrained by geometry, transport, and safety.

6. Measurement, constraints, and future directions

Measurement and validation are central because StimulHeat systems are often judged by local transients rather than by equilibrium temperatures. The MR-compatible animal temperature-regulation system uses rectal thermistor feedback, Peltier-controlled circulating water, and PID control to stabilize animal body temperature with standard deviations below a tenth of a degree after convergence (Verghese et al., 2023). A mouse converged to new setpoints in about 20 min, and PRESS MR spectroscopic thermometry reconstructed brain temperatures of $5.4\,^\circ\mathrm{C}$3 and $5.4\,^\circ\mathrm{C}$4, within less than $5.4\,^\circ\mathrm{C}$5 of rectal setpoint (Verghese et al., 2023). Although this is a regulation platform rather than a perceptual interface, it illustrates a recurring StimulHeat requirement: thermal output must be embedded in a metrology stack that is itself compatible with the surrounding sensing modality.

Safety findings vary by modality. The transcranial photobiomodulation model at 810 nm and 100 mW/cm² reported that effective light penetration is about 1 cm, reaching the cortex, while cortical temperature rise remained less than $5.4\,^\circ\mathrm{C}$6 in steady state and less than $5.4\,^\circ\mathrm{C}$7 in the first minute (Ibrahimi et al., 2022). By contrast, the airborne ultrasound haptics study reported palm temperature increases up to $5.4\,^\circ\mathrm{C}$8 after 30.0 s in static focus mode, prompting explicit discussion of discomfort and burn thresholds around $5.4\,^\circ\mathrm{C}$9–$5.0\,\mathrm{s}$0 (Iwabuchi et al., 2024). The dual-sided wearable Peltier work targets a warm-side temperature of $5.0\,\mathrm{s}$1 and reports a dual-sided lifetime of 206.3 s at 2.0 V, the selected optimum for reaching the target while preserving cool-side longevity (Kang et al., 2024). Across the literature, then, “safe thermal stimulation” is not a single bound but a modality-specific operating envelope defined jointly by absolute temperature, rate of change, exposure duration, geometry, and residual heat.

A second recurring limitation is confounding. The EEG imagined-temperature study notes that block order was not counterbalanced and that flames or ice visuals during real stimulation could introduce attentional or expectancy effects (Belichenko et al., 22 Aug 2025). Heatables did not report irradiance or exact optical output split between 810 nm and 950 nm channels, limiting dosimetric interpretation (Zitz et al., 3 Jun 2025). HeatFlicker currently rests on a preliminary observation with three participants and no quantitative evaluation of illusion strength or direction agreement (Ito et al., 2024). The persistent-cold display required offline thermal-camera calibration rather than an integrated skin-temperature feedback sensor (Xu et al., 2023). These are not trivial omissions, because StimulHeat claims often hinge on whether a response is truly thermal, crossmodal, or expectancy-driven.

Future directions are correspondingly technical. The trans-scale thermal-signaling review calls for improved estimates of thermal conductivity and interfacial conductance in diverse biomolecules and shapes, explicitly to connect molecular dynamics to hydrodynamics (Kurisaki et al., 2023). The Valve Index StimulHeat device points toward enhanced heat sinking and more advanced control algorithms for sustained cooling (Mesnage et al., 5 Sep 2025). The EEG study motivates individualized mu-band tuning, richer spatial features, and ITS-specific calibration for imagined hot-versus-cold decoding (Belichenko et al., 22 Aug 2025). The non-contact cold display suggests integrated real-time skin thermometry and more adaptive control under ambient variability (Xu et al., 2023). Taken together, these efforts suggest that StimulHeat is evolving toward quantitatively instrumented thermal actuation in which heat generation, transport, perception, and physiology are modeled as a single coupled system rather than as separate engineering layers.