Transcranial ultrasound stimulation in the treatment of Alzheimer's disease: a systematic review and Bayesian meta-analysis of randomized controlled trials
Abstract
Background: Alzheimer's Disease (AD) is a leading cause of irreversible dementia worldwide. Consequently, various therapies have emerged aiming to slow the progression of cognitive decline, among which Transcranial Ultrasound Stimulation stands out as a recent field of analysis in the literature. Objective: To compare, using a Bayesian statistical model, the potential clinical and cognitive improvement that Transcranial Ultrasound Stimulation may offer to patients with AD. Methods: A Bayesian meta-analysis of randomized controlled trials (RCTs) comparing Ultrasound Stimulation with sham or placebo techniques. Searches were conducted in PubMed, Embase, and Cochrane databases without language or date restrictions in February 2026. The primary outcome was the difference in ADAS-Cog scale scores between groups during the follow-up period, using Standardized Mean Difference (SMD) as the effect measure. A random-effects model was implemented in R software (brms/Stan packages) with weakly informative priors: Normal (0, 1) for the mean effect and Half-Normal (0, 0.5) for heterogeneity (tau). Convergence was assessed via R-hat, and prior sensitivity analysis was performed with wider distributions to verify the robustness of the conclusions. Risk of bias was assessed using the RoB 2.0 tool. Results: Three studies were included, totaling 84 patients with AD. The Bayesian model showed a trend toward lower ADAS-Cog increases (reflecting greater cognitive preservation) in the stimulation group, although this difference was not significant compared to the control group (SMD = -0.336; 95% CrI: [-1.075, 0.24]), with moderate heterogeneity (tau = 0.366). Bayesian posterior probabilities indicated an 87.2% chance of Transcranial Ultrasound Stimulation providing some benefit over the control (SMD < 0). The consistency of the results was reinforced by the similarity of the effect using a wider prior (SMD = -0.415; 95% CrI: [-1.458, 0.45]). Regarding the overall risk of bias (RoB 2.0), one RCT was classified as “low risk" and two as “high risk". Conclusion: Although a promising technique, it is not yet possible to state with precision whether Transcranial Ultrasound Stimulation is effective in slowing cognitive decline in AD. The low volume of available RCTs, methodological heterogeneity, and risk of bias constitute significant limitations to establishing solid conclusions.