Wearable Focused Ultrasound Neuromodulation and Electrophysiological Recording Patch 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 Chae, Daniel Song, Julieta Garcia, Rithvik Mittapalli, Adam Bush, Wynn Legon, Vincent Mysliwiec, Gregory A. Fonzo, Huiliang Wang

2025 · 10.1101/2025.10.19.683337

human healthyhuman patienthealthyothereeg megfmriother mriautonomic physiologyclinical scale

Abstract

The rise in sleep disease affecting the general population globally in the past decade has been detrimental to individually and socioeconomically. As of now, approaches often are temporary through medication, permanently using invasive implants with surgical complications or neuromodulation therapy. However, non-invasive, state-dependent neuromodulation during sleep is technically challenging, especially with the lack of flexibility, comfortability and robustness for sleep conditions. Here, we introduce a N on-invasive E lectrophysiological Recording and U ltrasound Neuromodulation Slee p P atch (NEUSLeeP) in delivering focused ultrasound stimulation to the subthalamic nucleus (STN) overnight with simultaneous stable polysomnography recording. Our sleep patch integrates a custom eight-channel concentric ring transducer array with real-time electroencephalography (EEG), electrooculography (EOG), electromyography (EMG) recording, and individualized line-of-sight targeting to sonicate deep brain areas while preserving mobility. Our platform operated safely and comfortably across the two-night’s sleep study. Stimulation of the left STN was delivered every 90 minutes throughout the night and was associated with a 25% increase in REM (Rapid Eye Movement) sleep duration and a 43 minutes reduction in REM sleep latency compared to a sham night in a study of 26 subjects. Blood Oxygen Level Dependent (BOLD) signal attenuation in functional Magnetic Resonance Imaging (fMRI) was localized primarily to a left ipsilateral basal-ganglia-midbrain-temporal circuit, consistent with selective network modulation rather than global arousal changes. Overall, NEUSLeeP demonstrates feasibility by (i) a light weight and wearable ultrasound neuromodulation and sleep recording during natural sleep; (ii) establishing a potential mechanism relating targeted ultrasound stimulation of STN/sleep networks to REM enhancement.

Abstract via europepmc.

Specieshuman

Subjects28 participants

Sessions per subject1

Randomisedno

Blindingsingle

Sham / controlinactive transducer

Auditory controlnot reported

Readout timingboth

Anaesthesianot applicable

Readoutseeg meg, fmri, other mri, autonomic physiology, clinical scale — overnight polysomnography (EEG/EOG/EMG) sleep staging; rsBOLD, ALFF and resting-state connectivity; ASL cerebral blood flow; heart rate variability; Hariri emotional face-matching fMRI task; SSS, ESS, STAI-S

Direction of effectmixed or unclearSTN-FUS night versus sham night was associated with longer REM duration and shorter REM latency, with no overall change in other sleep stages; fMRI after stimulation showed increased left-STN BOLD variability and left NAc CBF, which the authors interpret as probable STN inhibition without being able to separate excitatory from inhibitory input.

Adverse eventsnone observedSpO2, heart rate and blood pressure were monitored during stimulation with no significant changes, surface temperature and ex vivo skull heating were checked, and participants were followed up by email one week later; no adverse events are described.

Exposures

Exposure 1: Left STN-FUS with NEUSLeeP wearable concentric ring array (100 Hz PRF)

Target: subthalamic nucleus — “left STN”
Device: custom-built · custom 8-channel Concentric Ring Ultrasound Transducer Array (CRUTA) in NEUSLeeP patch, driven by Verasonics Vantage 64LE HIFU

Pulse timing
Waveform	pulsed
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 estimate	deratingsingle 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

Pulse Repetition Frequency 100 Hz, pulse duration 0.5 ms, duty cycle 5%, 30 s ON 30 s OFF. In the fMRI session 10 blocks (10 min); in the sleep study 5 min blocks delivered every 90 minutes for a total of five blocks overnight.

Flags from extraction

n_subjects — Text gives 28 sleep-study subjects (16 healthy, 12 insomnia), 26 analysed after two were excluded for noise, 16 enrolled in the fMRI session, and healthy n = 15 in Table 2; the total exposed is not unambiguous.
n_sessions_per_subject — Two consecutive nights (sham first, FUS second) plus a separate MRI-guided FUS fMRI session; 2 is the sleep-study count.
exposures[0].in_situ.pressure_kpa — Table 1 gives 0.68MPa derated at 72.2 mm; text elsewhere says derated threshold >0.70 MPa. Table value recorded.
exposures[0].free_field.pressure_kpa — Table 1 column is headed Free-Field; whether 0.90 MPa is peak or peak-to-peak is not stated.
exposures[0].timing.sonication_duration_s — Each block is 30 s ON and 30 s OFF; 30 s taken as the uninterrupted train. Burst duty is given as 50%.
randomised — Authors state stimulation order was not counterbalanced (sham night first), single-blinded, no parallel control; recorded as false.
exposures[0].device.family — Transducer is described as custom-made; maker not named. Driver is a Verasonics Vantage 64LE HIFU.