← Explore

Skin-attached bioadhesive patch enabling ultrasound deep brain stimulation and real-time electrophysiological monitoring for REM sleep enhancement

Kai Wing Kevin Tang, Benjamin Baird, William D. Moscoso-Barrera, Mengxia Yu, Mengmeng Yao, Jinmo Jeong, Ilya Pyatnitskiy, Anakaren Romero Lozano, Jiachen Wang, Ju-Chun Hsieh, Tony Sungjin Chae, Daniel Song, Julieta Garcia, Rithvik Mittapalli, Adam Bush, Wynn Legon, Vincent Mysliwiec, Gregory A. Fonzo, Huiliang Wang

Nature Communications 2026 · 10.1038/s41467-026-73787-6

human healthyhealthyotherfmrieeg megautonomic physiologybehaviourclinical scale

Abstract

Wearable bioelectronic interfaces capable of simultaneous neural sensing and targeted deep brain modulation remain limited by the lack of non-invasive technologies with sufficient spatial precision and mechanical stability for continuous operation. Here we report NEUSLeeP, a flexible skin-attached bioadhesive patch integrating electrophysiological sensing with transcranial focused ultrasound neuromodulation. The system incorporates a tunable concentric-ring ultrasound array, conformal hydrogel electrophysiological electrodes, and compliant interconnects within a soft substrate optimized for stable overnight operation. This integrated architecture enables spatially selective modulation and concurrent electrophysiological monitoring of deep brain structures, specifically the subthalamic nucleus during natural sleep. In a 28-participant study, NEUSLeeP demonstrated robust monitoring of sleep performance with precise-targeted neuromodulation of STN resulting in an increase in REM duration by 4.6% and reducing REM latency by 24%. This work establishes an ultrasound bioelectronic platform for non-invasive, spatiotemporally precise modulation and monitoring of deep neural circuits for neuroscience and bioelectronic medicine. ClinicalTrials.gov identifier: NCT07190287.

Abstract via europepmc.

Specieshuman
Subjects28 participants
Sessions per subject1
Randomisedno
Blindingsingle
Sham / controlinactive transducer
Auditory controlmasking sound
Readout timingboth
Anaesthesianot applicable
Readoutsfmri, eeg meg, autonomic physiology, behaviour, clinical scalePolysomnography (EEG/EOG/EMG) sleep staging; PSQI, Epworth Sleepiness Scale, Stanford Sleepiness Scale, State-Trait Anxiety Inventory; heart rate variability (ECG); emotional face-matching (Hariri) task response time; resting-state and task fMRI (BOLD, ALFF, pCASL CBF, functional connectivity)
Direction of effectinhibitoryA PRF of 100 Hz was selected throughout the study specifically to promote inhibition of the STN (based on a preliminary pilot study suggesting a PRF-dependent bimodal effect, with 10 Hz excitatory and 100 Hz inhibitory); STN-FUS was associated with increased REM duration and decreased REM latency, interpreted as downstream disinhibition of REM-promoting circuits secondary to STN inhibition.
Adverse eventsnone observedBlood oxygen saturation levels, heart rate, and blood pressure were measured throughout the session, showing no significant changes with STN-FUS by CRUTA on health vitals; follow-up emails one week after the session confirmed no detrimental health effects.

Exposures

Exposure 1: STN-FUS with NEUSLeeP/CRUTA (target engagement study and overnight REM sleep study)

Target: subthalamic nucleus — “left subthalamic nucleus (STN)
Device: custom-built · Concentric Ring Ultrasound Transducer Array (CRUTA)

Pulse timing
Waveformpulsed
Fundamental frequency (kHz)650✓✓
Pulse duration (ms)0.5✓✓
Pulse repetition frequency (Hz)100✓✓
Duty cycle (%)5pulse duration × PRF gives 5%✓✓
Sonication duration (s)30✓✓
Pressure and intensity, by domain
Free-field pressure (kPa)900✓✓
Free-field Isppa (W/cm²)26.9✓✓
Free-field Ispta (W/cm²)0.674✓✓
In-situ estimatederatingsingle value
In-situ pressure (kPa)680✓✓
In-situ Isppa (W/cm²)15.4✓✓
In-situ Ispta (W/cm²)0.385✓✓
Protocol, in the paper’s words

In an initial single-session target-engagement study (n=16), participants received 10 min of FUS delivered as 10 blocks of 30 s ON / 30 s OFF (Pressure 0.90 MPa, PRF 100 Hz, pulse duration 0.5 ms), before/after resting-state fMRI. In the subsequent overnight sleep study, participants underwent two consecutive overnight sessions (sham night 1, FUS night 2); on the FUS night, stimulation (30 s ON/30 s OFF for 5 min) was delivered per block every 90 min for five blocks overnight. Table 1 also reports a 'Burst Duty' of 50% (the 30 s ON/30 s OFF cycling) in addition to the 5% 'Pulse Duty' derived from the 0.5 ms pulse within the 10 ms PRF period. A separate preliminary pilot study (cited as ref 68, fMRI-compatible BrainSonix device) tested 10 Hz vs 100 Hz PRF at 5% duty cycle and ~719 mW/cm2 derated ISPTA, finding a PRF-dependent bimodal (excitatory at 10 Hz, inhibitory at 100 Hz) effect; frequency and other parameters for that pilot device are not given in the main text.

Flags from extraction

  • n_subjects28 subjects were recorded (healthy n=16, insomnia n=12) but 2 were excluded from analysis for excessive noise (subjects #1 and #25), leaving 26 analysed ('the observed effects of NEUSLeeP enabled STN-FUS in 26 subjects'); enrolled/recorded total (28) is reported here per convention. It is also unclear whether the 16 healthy participants overlap with the separately described 16-participant STN target-engagement sub-study (mean age reported as 25.7±6.3 years there vs. 22.7±5.4 years for the healthy REM-study group).
  • n_sessions_per_subjectthe value of 2 reflects the two consecutive overnight sessions (sham + FUS) of the main REM sleep study; a subset of participants may have also undergone an earlier single-session STN target-engagement scan (10 min FUS with pre/post fMRI), which would add a session not counted here.
  • exposures[0].timing.sonication_duration_sthe paper reports the sonication structure at several nested timescales (30 s ON/OFF block cycling; 5 or 10 min per experimental block; 10 min total for the engagement study); 30 s (the length of one uninterrupted ON period) was recorded as sonication_duration_s per the definition, with the other durations placed in protocol_description.
  • exposures[0].timing.duty_cycle_pctTable 1 reports both a 'Pulse Duty' of 5% (matching 0.5 ms pulse duration at 100 Hz PRF) and a separate 'Burst Duty' of 50% (the 30 s ON/30 s OFF block cycling); 5% (Pulse Duty) was recorded as duty_cycle_pct and the 50% Burst Duty is described in protocol_description.
  • exposures[0].device.manufacturerCRUTA is described as a custom-made transducer array built by the authors; no external manufacturer is named for the transducer itself (the driving electronics, Vantage 64LE, are from Verasonics).
  • condition_otherthe insomnia subgroup was screened by PSQI score only and explicitly described as not having received a formal clinical diagnosis; classified under 'other' rather than as a distinct clinical condition.