← Explore

Short-Term Cognitive and Behavioral Changes After Low-Intensity Transcranial Ultrasound Stimulation in Early Alzheimer’s Disease: An IPTW-Adjusted Retrospective Exploratory Comparative Study

YoungSoon Yang, Jaeho Kim, Youngsuep Kang, Seungho Choi, Jeonghyeon Seo, SangYun Kim

Brain Sciences 2026, 16, 913 · 10.3390/brainsci16090913

human patientalzheimers diseaseclinical scale

Abstract

Background: Low-intensity ultrasound (LIUS) has been proposed as a non-invasive approach to modulate cerebrospinal fluid movement, glymphatic–lymphatic clearance, and amyloid-beta (Aβ) handling in Alzheimer’s disease (AD). However, clinical comparisons in amyloid biomarker-confirmed patients receiving standard pharmacotherapy remain limited, and the short-term cognitive domains most likely to change after LIUS remain uncertain. Objective: The objective of this study was to compare 4-week changes in cognitive function, executive function, quality of life, and neuropsychiatric symptoms between amyloid PET-positive patients receiving add-on LIUS with donepezil-based standard treatment and those receiving donepezil-based standard treatment alone. Methods: This retrospective exploratory comparative cohort study included amyloid PET-positive patients with mild cognitive impairment or early AD who received donepezil-based standard treatment. The treatment cohort received 4 weeks of add-on LIUS at Dongtan or Bundang centers, whereas the comparator cohort received donepezil-based standard treatment alone at Soonchunhyang. Outcomes were changes from baseline to 4 weeks in Trail Making Test A (TMT-A), Trail Making Test B (TMT-B), Mini-Mental State Examination, Second Edition (MMSE-II), Quality of Life in Alzheimer’s Disease (QoL-AD), and Neuropsychiatric Inventory (NPI). Group differences were examined using unadjusted change-score comparisons and stabilized inverse probability of treatment weighting (IPTW). Results: The full analysis set included 51 amyloid PET-positive participants receiving donepezil-based standard treatment: 20 who additionally received LIUS and 31 who received standard treatment alone. Compared with controls, the add-on LIUS group showed greater improvement in TMT-A completion time in unadjusted analysis (−20.0 ± 31.3 s vs. −0.1 ± 5.6 s; p = 0.011), and this association remained significant after IPTW adjustment. TMT-B showed a favorable but less robust pattern, with borderline findings in the primary Fail-excluded analyses and a significant result in the Fail = 300 sensitivity analysis. MMSE-II change did not differ between groups. NPI improved more in the add-on LIUS group in unadjusted analysis, although this finding was exploratory. Conclusions: In amyloid PET-positive patients receiving donepezil-based standard treatment, 4 weeks of add-on LIUS was associated with short-term improvement in TMT-A performance compared with standard treatment alone, while MMSE-II remained unchanged. The findings suggest a preliminary signal for processing-speed or visual-search improvement and support prospective randomized sham-controlled studies with harmonized cognitive testing and biomarker endpoints.

Abstract via crossref.

Specieshuman
Subjects23 participants
Sessions per subject12
Randomisedno
Blindingnot reported
Sham / controlnone
Auditory controlnot reported
Readout timingoffline
Anaesthesianot applicable
Readoutsclinical scaleTrail Making Test Part A and B (TMT-A, TMT-B); Mini-Mental State Examination, Second Edition (MMSE-II); Quality of Life in Alzheimer's Disease (QoL-AD); Neuropsychiatric Inventory (NPI)
Direction of effectmixed or unclearAdd-on LIUS was associated with significantly greater improvement in TMT-A completion time versus standard treatment alone; TMT-B showed a favourable but statistically borderline pattern; MMSE-II showed no between-group difference; NPI improved more with LIUS in unadjusted analysis (exploratory).
Adverse eventsnone observedNo serious adverse events or treatment discontinuations attributed to LIUS were documented during the 4-week treatment period. However, adverse events were identified retrospectively from available clinical records rather than through standardized prospective monitoring; therefore, minor or transient symptoms may have been under-recorded.

Exposures

Exposure 1: Add-on low-intensity ultrasound stimulation (LIUS), bilateral frontal/temporal scalp placement, in amyloid PET-positive MCI/early AD patients

Target: whole brain or unfocused — “broad bilateral frontal and temporal exposure (unfocused, no stereotactic target)
Device: other named manufacturer · DeepSon Bio · Neuclare B-01

Pulse timing
Waveformpulsed
Fundamental frequency (kHz)250✓✓
Pulse duration (ms)100✓✓
Pulse repetition frequency (Hz)1✓✓
Duty cycle (%)10pulse duration × PRF gives 10%✓✓
Sonication duration (s)1,800✓✓
Pressure and intensity, by domain
Free-field pressure (kPa)390✓✓
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)280✓✓
In-situ Isppa (W/cm²)2.6✓✓
In-situ Ispta (W/cm²)0.26✓✓
Protocol, in the paper’s words

Each session comprised 30 min of sonication. LIUS was administered three times weekly for 4 weeks (12 sessions). Two transducers were positioned over the frontal scalp and one over each temporal region; the frontal transducers were oriented toward the posterior fontanelle, and the temporal transducers were oriented toward the contralateral side. This standardized transducer arrangement was intended to provide broad bilateral frontal and temporal exposure rather than individualized stimulation of a stereotactically defined brain target.

Flags from extraction

  • n_subjects23 records initially comprised the LIUS-treated dataset (enrolled/treated), but 3 were excluded from the final analysis set due to insufficient outcome data, leaving 20 analysed; enrolled/treated number (23) is used here.
  • exposures[0].in_situIn-situ pressure/intensity values are taken from a previously published acoustic simulation of the same device platform (Jo et al.), not from patient-specific measurements in this cohort, per the paper's own caveat.
  • sham_typeThere was no sham procedure; the comparator group received standard treatment alone rather than a sham/placebo ultrasound device, which does not map cleanly onto any listed sham_type other than 'none'.