← Explore

Low-intensity focused ultrasound pulsation along the anterior–posterior thalamic axis differentially modulates the latency of reporting conscious visual experience

Hyunwoo Jang, Jiyang Liu, Anthony G. Hudetz, Zirui Huang

2026 · 10.64898/2026.08.21.746115

human healthyhealthybehaviour

Abstract

Background Transcranial low-intensity focused ultrasound (LIFU) neuromodulation can alter human task performance depending on target and acoustic configurations. However, single behavioral endpoints cannot locate effects within multistep tasks, and predefined target labels ignore acoustic variations. Objective To determine whether thalamic LIFU affects visual categorization or subsequent subjective report latency and whether the effects vary with target, acoustic parameter, and beam location. Methods Sixty healthy adults were randomized to 70% or 5% duty cycle (DC70 or DC5) and received sonication on four left thalamic targets with matched pulse repetition frequency (10 Hz) and temporal-average intensity (0.72 W/cm 2 ). Behavioral models tested target-by-DC interactions in categorization (RT1) and subjective report (RT2) latencies. Spatial analyses correlated focal spot coordinates and voxel-wise intensity from 179 acoustic simulations to baseline-adjusted RT2. Results The target-by-DC interaction was detected for RT2 but not RT1. At the ventroposterior thalamic target, adjusted RT2 was 55.9 ms longer under DC70 than DC5. More anterior focal spots shortened RT2 under DC70 but increased RT2 under DC5. Correlation between intensity and adjusted RT2 significantly differed between DC70 and DC5 in 18.8% of thalamic voxels. These voxels formed an anterior mediodorsal–motor set and a posterior pulvinar-dominant set. Conclusions The latency of reporting conscious visual experience, but not categorization latency, was affected by thalamic LIFU. This effect varied jointly with anterior–posterior target engagement and acoustic configuration. Analyzing sequential reaction times separately and treating field variation as an anatomical variable revealed associations not fully captured by a single endpoint or predefined target labels. HIGHLIGHTS Ultrasound changed conscious visual report latency, not categorization latency. Reports were slower at 70% than 5% duty cycle in ventroposterior thalamus. Anterior sonication made reports faster at 70% and slower at 5% duty cycle. Posterior sonication made reports slower at 70% and faster at 5% duty cycle.

Abstract via europepmc.

Specieshuman
Subjects60 participants
Sessions per subject1
Randomisedyes
Blindingnot reported
Sham / controlnone
Auditory controlnot reported
Readout timingboth
Anaesthesianot applicable
ReadoutsbehaviourReaction times to a near-threshold visual task: RT1 (four-alternative forced-choice object categorization) and RT2 (subjective report of visual experience, yes/no)
Direction of effectbidirectionalRT2 (subjective report latency) was 55.9 ms longer under DC70 than DC5 at the ventroposterior thalamic target; the relationship between focal-spot position/simulated intensity and RT2 reversed sign between DC70 and DC5 along the anterior-posterior axis. RT1 (categorization latency) showed no effect.
Adverse eventsnot reported

Exposures

Exposure 1: ventroanterior thalamus (VA)

Target: thalamus — “ventroanterior thalamus (VA)
Device: BrainSonix · BrainSonix · BXPulsar 1002

Pulse timing
Waveformpulsed
Fundamental frequency (kHz)650✓✓
Pulse duration (ms)70, 5swept✓✓
Pulse repetition frequency (Hz)10✓✓
Duty cycle (%)70, 5swept✓✓
Sonication duration (s)30✓✓
Pressure and intensity, by domain
Free-field pressure (kPa)not reported
Free-field Isppa (W/cm²)not reported
Free-field Ispta (W/cm²)1.02✓✓
In-situ estimatederatingsingle value
In-situ pressure (kPa)not reported
In-situ Isppa (W/cm²)1.03, 14.4swept✓✓
In-situ Ispta (W/cm²)0.72✓✓
Protocol, in the paper’s words

Each LIFU-ON block contained twelve 30-s sonication epochs alternating with 30-s intervals, producing 23 alternating ON/OFF epochs over approximately 11.5 min; LIFU epochs were not synchronized to task trials. Four thalamic targets were sonicated in pseudo-randomized, counterbalanced order across the four LIFU-ON blocks, each using a fixed duty cycle (DC70 or DC5) for that subject.

Exposure 2: ventroposterior thalamus (VP)

Target: thalamus — “ventroposterior thalamus (VP)
Device: BrainSonix · BrainSonix · BXPulsar 1002

Pulse timing
Waveformpulsed
Fundamental frequency (kHz)650✓✓
Pulse duration (ms)70, 5swept✓✓
Pulse repetition frequency (Hz)10✓✓
Duty cycle (%)70, 5swept✓✓
Sonication duration (s)30✓✓
Pressure and intensity, by domain
Free-field pressure (kPa)not reported
Free-field Isppa (W/cm²)not reported
Free-field Ispta (W/cm²)1.02✓✓
In-situ estimatederatingsingle value
In-situ pressure (kPa)not reported
In-situ Isppa (W/cm²)1.03, 14.4swept✓✓
In-situ Ispta (W/cm²)0.72✓✓
Protocol, in the paper’s words

Each LIFU-ON block contained twelve 30-s sonication epochs alternating with 30-s intervals, producing 23 alternating ON/OFF epochs over approximately 11.5 min; LIFU epochs were not synchronized to task trials. Four thalamic targets were sonicated in pseudo-randomized, counterbalanced order across the four LIFU-ON blocks, each using a fixed duty cycle (DC70 or DC5) for that subject.

Exposure 3: dorsoanterior thalamus (DA)

Target: thalamus — “dorsoanterior thalamus (DA)
Device: BrainSonix · BrainSonix · BXPulsar 1002

Pulse timing
Waveformpulsed
Fundamental frequency (kHz)650✓✓
Pulse duration (ms)70, 5swept✓✓
Pulse repetition frequency (Hz)10✓✓
Duty cycle (%)70, 5swept✓✓
Sonication duration (s)30✓✓
Pressure and intensity, by domain
Free-field pressure (kPa)not reported
Free-field Isppa (W/cm²)not reported
Free-field Ispta (W/cm²)1.02✓✓
In-situ estimatederatingsingle value
In-situ pressure (kPa)not reported
In-situ Isppa (W/cm²)1.03, 14.4swept✓✓
In-situ Ispta (W/cm²)0.72✓✓
Protocol, in the paper’s words

Each LIFU-ON block contained twelve 30-s sonication epochs alternating with 30-s intervals, producing 23 alternating ON/OFF epochs over approximately 11.5 min; LIFU epochs were not synchronized to task trials. Four thalamic targets were sonicated in pseudo-randomized, counterbalanced order across the four LIFU-ON blocks, each using a fixed duty cycle (DC70 or DC5) for that subject.

Exposure 4: dorsoposterior thalamus (DP)

Target: thalamus — “dorsoposterior thalamus (DP)
Device: BrainSonix · BrainSonix · BXPulsar 1002

Pulse timing
Waveformpulsed
Fundamental frequency (kHz)650✓✓
Pulse duration (ms)70, 5swept✓✓
Pulse repetition frequency (Hz)10✓✓
Duty cycle (%)70, 5swept✓✓
Sonication duration (s)30✓✓
Pressure and intensity, by domain
Free-field pressure (kPa)not reported
Free-field Isppa (W/cm²)not reported
Free-field Ispta (W/cm²)1.02✓✓
In-situ estimatederatingsingle value
In-situ pressure (kPa)not reported
In-situ Isppa (W/cm²)1.03, 14.4swept✓✓
In-situ Ispta (W/cm²)0.72✓✓
Protocol, in the paper’s words

Each LIFU-ON block contained twelve 30-s sonication epochs alternating with 30-s intervals, producing 23 alternating ON/OFF epochs over approximately 11.5 min; LIFU epochs were not synchronized to task trials. Four thalamic targets were sonicated in pseudo-randomized, counterbalanced order across the four LIFU-ON blocks, each using a fixed duty cycle (DC70 or DC5) for that subject.

Flags from extraction

  • exposures[*].target.termsPaper explicitly states the Tian Scale-II parcel labels (VA/VP/DA/DP) 'are derived from functional connectivity gradient mapping and are not equated with similarly named anatomical structures'; mapped to generic 'thalamus' rather than a specific nucleus term.
  • n_subjects60 subjects were enrolled and exposed to ultrasound; 6 were subsequently excluded for technical reasons, leaving 54 analysed for the reaction-time results.
  • auditory_controlPaper notes as a limitation that DC5's larger pulse-average intensity created larger transducer noise and that 'nonspecific auditory confounds could have acted differentially', but no auditory control/masking procedure is described.
  • exposures[*].free_field.isppa_w_cm2Only derated (in_situ) ISPPA values are reported for DC70/DC5; no free-field ISPPA value is stated.