← Explore

Transcranial ultrasound stimulation in humans is associated with an auditory confound that can be effectively masked

Verena Braun, Joseph Blackmore, Robin O. Cleveland, Christopher R. Butler

Brain Stimulation 2020, 13, 1527-1534 · 10.1016/j.brs.2020.08.014

human healthyhealthyeeg megbehaviour

Abstract

Background Transcranial ultrasound stimulation (TUS) is emerging as a potentially powerful, non-invasive technique for focal brain stimulation. Recent animal work suggests, however, that TUS effects may be confounded by indirect stimulation of early auditory pathways. Objective We aimed to investigate in human participants whether TUS elicits audible sounds and if these can be masked by an audio signal. Methods In 18 healthy participants, T1-weighted magnetic resonance brain imaging was acquired for 3D ultrasound simulations to determine optimal transducer placements and source amplitudes. Thermal simulations ensured that temperature rises were Results We found that participants can hear sound during TUS and can distinguish between stimulation and non-stimulation trials. This was corroborated by EEG recordings indicating auditory activation associated with TUS. Delivering an audio waveform to participants through earphones while TUS was applied reduced detection rates to chance level and abolished the TUS-induced auditory EEG signal. Ex vivo skull experiments demonstrated that sound is conducted through the skull at the pulse repetition frequency of the ultrasound. Conclusion Future studies using TUS in humans need to take this auditory confound into account and mask stimulation appropriately.

Abstract via europepmc.

Specieshuman
Subjects18 participants
Sessions per subject1
Randomisedyes
Blindingnot reported
Sham / controlundescribed
Auditory controlmasking sound
Readout timingonline
Anaesthesianot applicable
Readoutseeg meg, behaviourdetection task (d' from hit/false-alarm rates distinguishing stimulation from sham trials); auditory event-related potentials (US-N1, US-P2) over electrode Cz
Direction of effectnot assessedThe study assessed an auditory confound rather than the intended neuromodulatory effect at V1: TUS elicited an audible tone and an auditory-evoked-potential-like EEG response (pronounced N1) that were abolished by an audio mask delivered through earphones.
Adverse eventsnot reportedAcoustic and thermal participant-specific modelling was carried out to satisfy safety constraints (maximum peak modelled pressure in CSF or brain tissue 0.6 MPa, maximum temperature rise in skull bone 3C, maximum temperature rise in brain tissue 1C).

Exposures

Exposure 1: TUS to right visual cortex (unmasked and masked blocks)

Target: primary visual cortex — “right visual cortex
Device: Sonic Concepts · Sonic Concepts Inc · H-107

Pulse timing
Waveformpulsed
Fundamental frequency (kHz)500✓✓
Pulse duration (ms)0.5✓✓
Pulse repetition frequency (Hz)1,000✓✓
Duty cycle (%)50pulse duration × PRF gives 50%✓✓
Sonication duration (s)0.3✓✓
Pressure and intensity, by domain
Free-field pressure (kPa)not reported
Free-field Isppa (W/cm²)not reported
Free-field Ispta (W/cm²)not reported
In-situ estimatesimulationmean or range across subjects
In-situ pressure (kPa)500✓✓
In-situ Isppa (W/cm²)not reported
In-situ Ispta (W/cm²)not reported
Protocol, in the paper’s words

Each stimulation trial consisted of a 300 ms burst; 100 unmasked trials (50 stimulation, 50 sham) followed by 100 masked trials (50 stimulation, 50 sham) with a 1 kHz square-wave audio mask starting ~112 ms before TUS onset and lasting 700 ms, played through earphones.

Flags from extraction

  • sham_typeSham/no-stimulation trials are described only as trials in which 'no TUS was applied'; the physical mechanism (transducer powered off vs. simply not triggered) is not described.
  • direction_of_effectThis paper's purpose is to characterise an auditory confound rather than an intended neuromodulatory effect at the visual-cortex target, so direction_of_effect is recorded as not_assessed with the confound described in direction_notes.
  • exposures[0].in_situ.pressure_kpa0.50 MPa is the group mean modelled (simulated) in-situ pressure at the V1 target reported for a subset of analysis (one participant's context); individual/group range was 0.30-0.60 MPa.
  • n_sessions_per_subjectThe paper does not state whether the unmasked and masked blocks (which were always run in that order) occurred within one session or across multiple visits.