Integrated neuronavigation and acoustic-thermal modeling for simultaneous transcranial ultrasound stimulation (TUS) and functional MRI
Moon Jeong, Ali K Zadeh, Alan Coreas, Fady Girgis, Samuel Pichardo, G Bruce Pike
Journal of Neural Engineering 2026 · 10.1088/1741-2552/ae9347
Abstract
Objective . We present and validate an integrated workflow for accurate and stable targeting of a TUS transducer inside the MRI scanner using subject-specific imaging, neuronavigation, acoustic-thermal modeling, transducer positioning and fixation, and in-scanner verification. Approach . A stimulation target was defined in the left primary motor cortex (M1) hand knob using structural imaging and task-based functional magnetic resonance imaging (fMRI) collected during a planning session, then transferred to neuronavigation for trajectory planning and scalp marking. Acoustic and thermal simulations characterized the acoustic beam geometry and adjusted ultrasound power to achieve the desired acoustic field at the target within safety constraints. During the experiment scanning session, low-resolution T 1 -weighted localizer images visualized transducer fiducials and the planned target directly on the scanner console, enabling iterative in-bore refinement. Main results . Coupled with a custom MRI-compatible transducer fixation cap enabling controlled multi-axis adjustments, this workflow achieved a mean post-refinement off-target distance of 1.41.0 mm across five participants over 19 sessions, falling within one fMRI voxel (2.5 mm) of the planned target in 17 of 19 sessions (89%). Repeat transducer position verification approximately 40 min after initial setup showed minimal positional drift (0.2 mm). In an expanded cohort of 12 participants with balanced sex distribution (six males, six females), mean post-adjustment off-target distance was 2.180.78 mm in males and 1.930.69 mm in females, with comparable performance between sexes (Mann-Whitney U = 22; descriptive), and an overall success rate of nine of 12 sessions (75%). fMRI tSNR in M1 was preserved with the passive transducer (left-to-right M1 ratio 0.900.12) and during active sonication (0.900.09) across the same 12 participants, indicating that the 20 mm coupling gel pad effectively mitigated susceptibility-related signal dropout and that active sonication introduced no additional signal degradation. Significance . This workflow provides a practical experimental foundation for in-scanner targeting and verification for simultaneous human TUS-fMRI studies, with MRI signal quality preserved during active sonication and consistent targeting performance across male and female participants.
Abstract via europepmc.
Exposures
Exposure 1: Left M1 hand knob, BabelBrain-planned in-scanner TUS
Target: primary motor cortex — “hand knob region of the left primary motor cortex”
Device: BrainSonix · BrainSonix Inc. · BX Pulsar 1002 ✓
| Waveform | pulsed | |
|---|---|---|
| Fundamental frequency (kHz) | 650 | ✓✓✓ |
| Pulse duration (ms) | 1 | ✓✓✓ |
| Pulse repetition frequency (Hz) | 100 | ✓✓✓ |
| Duty cycle (%) | 10pulse duration × PRF gives 10% | ✓✓✓ |
| Sonication duration (s) | 90 | ✓✓✓⚑ |
| Free-field pressure (kPa) | not reported | |
|---|---|---|
| Free-field Isppa (W/cm²) | not reported | |
| Free-field Ispta (W/cm²) | not reported | |
| In-situ estimate | simulationmean or range across subjects | |
| In-situ pressure (kPa) | not reported | |
| In-situ Isppa (W/cm²) | 1.73, 4.08swept | ✓✓✓⚑ |
| In-situ Ispta (W/cm²) | not reported |
Sonication parameters applied across all participants: fundamental frequency 650 kHz, PD 1 ms, PRF 100 Hz, DC 10%, pulse train duration 90 s, 2 sonication repetitions (concurrent TUS-fMRI).
Flags from extraction
n_subjects— 17 unique participants overall, but active sonication (tSNR analysis) stated for 12 participants (Subjects 6-17); the others only had positioning validation. Recorded 12.exposures[0].in_situ.isppa_w_cm2— Simulated Isppa after skull attenuation across participants (mean 3.50 +/- 0.87 W/cm2; range given); per-subject values only in supplementary table S3. No free-field values stated in the main text.exposures[0].timing.sonication_duration_s— Pulse train duration stated for simulation and for the concurrent TUS-fMRI acquisition; study is a methods paper with no stimulation outcomes.