Low‐intensity focused ultrasound neuromodulation for drug‐resistant epilepsy: A randomized, sham‐controlled crossover trial

Chien‐Chen Chou, Yen‐Cheng Shih, Yi‐Hsiu Chen, Po‐Tso Lin, Chun‐Fu Lin, Po‐Chun Chu, Hao‐Li Liu, Cheng‐Chia Lee, Hsiang‐Yu Yu

Epilepsia 2026 · 10.1002/epi.70404

human patientepilepsyclinical scaleother mri

Abstract

Objective This study was undertaken to evaluate the safety, feasibility, and efficacy of low-intensity focused ultrasound (LIFU) as a noninvasive neuromodulation technique in patients with drug-resistant epilepsy (DRE). Methods In this pilot, single-blind, randomized sham-controlled crossover trial, 12 patients with DRE underwent both LIFU and sham stimulation in a randomized sequence, targeting the seizure onset zone (SOZ), each followed by a 4-week observation period. Seizure frequency was analyzed as proportional change from baseline using linear mixed-effects models. An open-label extension phase evaluated longitudinal seizure outcomes following LIFU. Safety assessments included neurological examinations, magnetic resonance imaging (MRI), and neuropsychological and quality of life (QOL) measures. Results During the crossover phase, LIFU did not significantly reduce seizure frequency compared with sham (estimate = .49, 95% confidence interval [CI] = -.04 to 1.01, p = .068). No period or sequence effects were observed. Conversely, during the open-label extension phase, seizure frequency demonstrated a significant longitudinal reduction following LIFU (β = -14.0 percentage points per month, 95% CI = -22.2 to -5.8, p = .001). Post-LIFU MRI showed no structural lesions. No significant changes were observed in anxiety, depression, or QOL scores. There were only transient mild-to-moderate adverse events reported, without serious complications. Significance LIFU neuromodulation targeting the SOZ is safe and well tolerated. Although the primary crossover analysis did not demonstrate a statistically significant antiseizure effect, delayed seizure reduction during the extended follow-up suggests potential sustained neuromodulatory effects. Larger parallel-group trials with longer observation periods are warranted.

Abstract via europepmc.

Specieshuman

Subjects12 participants

Sessions per subject1

Randomisedyes

Blindingsingle

Sham / controlinactive transducer

Auditory controlsound only sham

Readout timingoffline

Anaesthesianot applicable

Readoutsclinical scale, other mri — Daily seizure diaries (seizure frequency); Beck Anxiety Inventory (BAI); Beck Depression Inventory (BDI); Quality of Life in Epilepsy Inventory-31 (QOLIE-31); structural brain MRI safety assessment

Direction of effectmixed or unclearThe primary sham-controlled crossover analysis did not show a significant reduction in seizure frequency with LIFU versus sham (p=.068), but an uncontrolled open-label extension phase after crossover showed a significant longitudinal decrease in seizure frequency over time following LIFU.

Adverse eventsobservedNo serious adverse events were observed. Mild headache was the most commonly reported adverse event, occurring during both FUS and sham treatment and follow-up; other transient symptoms included tremor (n=1), dental vibration sensation (n=1), and abdominal discomfort (n=1). All adverse events were self-limited and resolved spontaneously without medical intervention.

Exposures

Exposure 1: LIFU to individualized seizure onset zone (SOZ)

Target: hippocampus, frontal lobe, parietal lobe, temporal lobe — “seizure onset zone (SOZ); individualized targets predominantly hippocampus, with frontocentral, parietal, and temporal operculum targets in a subset of patients”
Device: other named manufacturer · NaviFUS Corporation · 256-element phased array LIFU ✓

Pulse timing
Waveform	pulsed
Fundamental frequency (kHz)	500	✓✓✓
Pulse duration (ms)	3	✓✓✓
Pulse repetition frequency (Hz)	100	✓✓✓
Duty cycle (%)	30pulse duration × PRF gives 30%	✓✓✓
Sonication duration (s)	300	✓✓✓

Pressure and intensity, by domain
Free-field pressure (kPa)	not reported
Free-field Isppa (W/cm²)	not reported
Free-field Ispta (W/cm²)	1.8, 7.2swept	✓✓✓⚑
In-situ estimate	simulationsingle value
In-situ pressure (kPa)	180	✓✓✓
In-situ Isppa (W/cm²)	not reported
In-situ Ispta (W/cm²)	0.3	✓?⚑

Protocol, in the paper’s words

Each session consisted of three consecutive 5-min sonications separated by two 5-min intermissions. Acoustic parameters were defined at both a safety ceiling (estimated intracranial ISPTA of 2.8 W/cm2, MI=.75, peak pressure .53 MPa, RMS pressure .37 MPa) and a planned therapeutic level (intracranial ISPTA of .3 W/cm2, MI=.25, peak pressure .18 MPa, RMS pressure .12 MPa); treatment trajectories were selected to approach the therapeutic level as closely as possible.

Flags from extraction

exposures[0].target — Target (seizure onset zone) was individualized per patient across hippocampus, frontal, parietal, and temporal sites (Table 1); reported as a single exposure with combined target terms rather than per-target breakdown, since outcomes are pooled across the cohort.
exposures[0].free_field.ispta_w_cm2 — Paper reports 'system output ISPTA' per patient without explicitly stating this was measured in water/free-field; domain assignment (free_field vs in_situ pre-skull) is inferred from context (it precedes the skull-corrected 'estimated transcranial' value).
exposures[0].in_situ.ispta_w_cm2 — Per-patient post hoc simulation gave an estimated transcranial ISPTA ranging from .24 to .36 W/cm2 (mean .3±.04 W/cm2), closely matching the planned therapeutic single value of .3 W/cm2 used here; the range is not separately recorded.
n_sessions_per_subject — Reflects the two sham-controlled crossover sessions; the open-label extension phase may have involved additional LIFU treatments, but the number of extension sessions is not clearly stated.