For decades, the only way to electrically stimulate a deep brain structure was to drill through the skull and implant electrodes. A new device from the University of Texas at Austin does something closer to the opposite: a flexible patch stuck to the skin delivers focused ultrasound seven centimeters (2.8 inches) into the brain, enough to reach the subthalamic nucleus, and in a 28-person trial it measurably changed sleep. Time to REM sleep fell by 43 minutes, and REM duration increased by about 16 minutes, with no changes to other sleep stages.
The device, named NEUSLeeP, for noninvasive electrophysiological recording and ultrasound neuromodulation sleep patch, combines two technologies in one wearable. A concentric-ring ultrasound transducer array, eight channels and just over three millimeters (0.12 inches) thick, generates focused beams that can be steered in depth by adjusting the phase of each ring. Around it sit soft hydrogel electrodes, four EEG channels plus eye-movement and muscle channels, that record sleep stages while the stimulation runs. The whole assembly weighs about 103 grams (3.6 ounces) and adheres to the skin with a bioadhesive elastomer.
The trial targeted the subthalamic nucleus, a structure about 72 millimeters (2.8 inches) below the scalp in the adult brain, best known as a target for deep brain stimulation in Parkinson’s disease. The choice was deliberate: STN is deep, which makes it a demanding test of whether noninvasive ultrasound can reach structures previously accessible only to implanted electrodes. The stimulation used 100-hertz pulses, a frequency that in earlier pilot work appeared inhibitory, mirroring the effect of clinical STN deep brain stimulation on sleep.
The headline results come from a within-subject comparison: REM latency fell from an average of 177 minutes to 135 minutes, a 24% reduction, and REM as a share of total sleep rose from 16.3% to 20.9%. The trial was not a test of whether ultrasound could cure insomnia, but a proof that targeted neuromodulation of a deep structure through the intact skull can shift a well-defined physiological state. The authors note that the REM increase of about 4.6% of total sleep exceeds what invasive STN stimulation achieved in earlier studies, a striking point for a device that touches only the skin.
The engineering is as notable as the physiology. The bioadhesive, a composite of silicone elastomer and a polymer called PEIE, achieves skin adhesion roughly 20 times stronger than the base material alone, and doubles as an acoustic couplant that lets ultrasound pass through with minimal loss. The hydrogel electrodes have about half the impedance of commercial EEG electrodes in the band that matters for sleep staging. In a separate validation, the patch’s sleep staging agreed with a standard 32-channel clinical headcap at a Cohen’s kappa of 0.88, near-perfect agreement, using only six electrodes. The patch left negligible skin irritation after eight hours of wear.
The mechanism deserves care. Ultrasound neuromodulation is real but not fully understood; leading hypotheses involve mechanical effects on ion channels and cell membranes rather than direct electrical excitation. The 100-hertz inhibitory effect seen in this study is consistent with the idea that the patch is influencing the same circuits that deep brain stimulation reaches, but the biological chain from acoustic pressure to changed sleep architecture is not yet fully mapped. The derated acoustic pressure, 0.9 megapascals (130 psi), sits above the safety threshold commonly cited for transcranial ultrasound, and the trial was conducted under an approved protocol with DARPA oversight.
The choice of REM sleep as the target outcome is notable. REM has clear physiological markers, which makes it measurable, and many people struggle to reach or maintain it, from shift workers to older adults to patients with depression, whose REM patterns are characteristically disturbed. The trial does not claim to treat any of those conditions; it demonstrates a technique. But a noninvasive wearable that can push a deep brain structure in a direction that measurably changes sleep architecture gives researchers a new experimental tool, and potentially a therapeutic one, for a class of problems that has been remarkably resistant to noninvasive intervention.
There is also a practical argument. Polysomnography, the gold standard for sleep measurement, requires a laboratory, a technician, and a 32-channel electrode cap. The NEUSLeeP patch performs comparable sleep staging with six electrodes and no laboratory, which means sleep research could move out of the clinic and into the home, and sleep recording could become a continuous, low-cost monitoring tool rather than a one-night snapshot. A wearable that both measures and modulates sleep is more than a lab curiosity.
The limitations are those of a pilot study: 28 participants, a single target structure, and results that establish feasibility rather than clinical benefit. Whether the effect persists with repeated use, whether it transfers to people with insomnia rather than healthy sleepers, and whether deeper or more lateral structures can be reached all remain open questions. What the trial demonstrates is that the toolbox of noninvasive brain modulation has grown: a patch on the temple can now do, at least in miniature, what once required an operating room.
Sources: Tang, K.W.K., Baird, B., Moscoso-Barrera, W.D. et al. Skin-attached bioadhesive patch enabling ultrasound deep brain stimulation and real-time electrophysiological monitoring for REM sleep enhancement. Nature Communications 17, 5570 (2026). DOI: 10.1038/s41467-026-73787-6.