← Explore

Ultrasound transcutaneous auricular vagus nerve stimulation enhances semantic processing

Youbin Kang, Marcus Kaiser, JeYoung Jung

2026 · 10.64898/2026.02.20.707135

human healthyhealthybehaviour

Abstract

Efficient access to conceptual knowledge is central to everyday cognition, yet the effects of non-invasive neuromodulation on semantic retrieval remain insufficiently characterised. Transcutaneous auricular vagus nerve stimulation (taVNS) offers a mechanistically distinct route to modulate distributed cognitive systems via ascending vagal afferent pathways, but its effects on semantic processing has not been explored. Here, we tested whether ultrasound-based taVNS delivered to the right cymba conchae can acutely enhance semantic retrieval efficiency in healthy adults. Twenty-seven participants completed a single blind, sham-controlled study comprising two counterbalanced sessions (active vs. sham) in a within-subject design, separated by at least five days. In each session, participants performed a semantic association task and a number-judgement control task before and immediately after 30 minutes of stimulation. Active taVNS produced selective post-stimulation improvements in semantic retrieval performance relative to sham and baseline, while performance on the control task remained unchanged. Adverse effects were minimal and did not differ between active and sham conditions. These findings provide causal evidence that ultrasound-based taVNS can acutely improve the efficiency of semantic processing, expanding the cognitive profile of taVNS beyond predominantly episodic and emotional memory domains and supporting its feasibility as a scalable approach for memory enhancement.

Abstract via europepmc.

Specieshuman
Subjects27 participants
Sessions per subject1
Randomisednot reported
Blindingsingle
Sham / controlundescribed
Auditory controlnot reported
Readout timingoffline
Anaesthesianot applicable
ReadoutsbehaviourSemantic association task (accuracy, reaction time, inverse efficiency) and a number-judgement control task
Direction of effectmixed or unclearActive taVNS increased post-stimulation accuracy relative to sham (one-tailed t(26)=2.29, p=.015) and produced faster reaction times relative to pre-stimulation (t(26)=2.56, p=.017); the omnibus ANOVA interaction was not significant, and effects relied on planned, partly one-tailed post hoc comparisons. Performance on the non-semantic control task was unchanged.
Adverse eventsobservedA side-effects questionnaire (skin irritation, dizziness, ear pain, etc.) was completed after each session; a small number of mild events were reported (e.g. 3 instances of ear pain in both active and sham), with no significant difference between active and sham stimulation (all chi-square <1.02, p>.313).

Exposures

Exposure 1: Active U-taVNS (ultrasound taVNS, right cymba conchae)

Target: vagus nerve — “auricular branch of the vagus nerve (right cymba conchae)
Device: other named manufacturer · NeurGear Inc. · ZenBud

Pulse timing
Waveformpulsed
Fundamental frequency (kHz)5,300✓✓
Pulse duration (ms)not reportedimplied by duty cycle ÷ PRF: 12.1951 ms (not stated by the paper)
Pulse repetition frequency (Hz)41✓✓
Duty cycle (%)50✓✓
Sonication duration (s)1,800✓✓
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 estimatenot reported
In-situ pressure (kPa)not reported
In-situ Isppa (W/cm²)not reported
In-situ Ispta (W/cm²)not reported
Pressure, domain unspecified (kPa)1,030✓✓
Isppa, domain unspecified (W/cm²)0.081✓✓
Protocol, in the paper’s words

Participants received 30 minutes of taVNS (active or sham) after the pre-stimulation session; the semantic association task and a number-judgement control task were repeated immediately after stimulation (post-stimulation session).

Consistency checks: intensity pressure inconsistent unspecified domain.

Flags from extraction

  • sham_typeThe paper states there was an 'active stimulation session and a sham session' but does not describe the sham device/mechanism (unlike the companion Kang emotional-bias paper), so sham mechanism is unspecified.
  • exposures[0].unspecified_domain.pressure_kpaSentence gives 'an average pressure amplitude of 1.03 MPa' alongside a separately labelled 'peak mechanical output of 0.081 W/cm2'; domain (free-field vs in-situ) is not stated for either value.
  • exposures[0].timing.pulse_duration_msOnly duty cycle (50%) and PRF (41 Hz) are stated; per-pulse duration is not explicitly reported and was not computed from DC/PRF.
  • readout_timingTasks were performed before and immediately after the 30-minute stimulation period; no assessment during stimulation is described, so timing is classified as offline.