Intensive Rehabilitation With Adjunctive Bilateral Anodal tDCS in Post-Stroke Dysphagia: A Multicenter Randomized Controlled Trial

Clinical Question (PICO)

In adults with ischemic stroke 72 hours to 6 months previously and FEES-confirmed oropharyngeal dysphagia, does bilateral anodal transcranial direct current stimulation (tDCS), added to six weeks of intensive speech-language therapy (SLT), improve swallowing more than sham stimulation with the same SLT? The primary measure was the Dysphagia Outcome and Severity Scale (DOSS), assessed at two and six weeks. Lesion location, treatment timing, sex, and baseline severity were explored as possible moderators, not established treatment-selection rules. This is a look-back review of a July 2026 publication, not a new trial announcement or a statement of current guidelines. Full trial report.

Bottom Line

Intensive SLT was associated with recovery regardless of stimulation condition, while bilateral anodal tDCS did not demonstrate an overall advantage over sham. Among 46 completers, mean DOSS rose from 2.9 to 4.7 with active stimulation and from 3.1 to 4.4 with sham by six weeks; a two-week infratentorial MASA signal (P = 0.04) was exploratory, not confirmation of subgroup efficacy. The absence of superiority does not establish equivalence. Results and Table 2.

Design

Trial type: Multicenter, randomized, double-blind, sham-controlled parallel trial; patients and outcome assessors were blinded.
N: 79 screened; 48 randomized.
- Active bilateral anodal tDCS: 24 randomized; 24 completed.
- Sham tDCS: 24 randomized; 22 completed. Two discontinued for logistical reasons.
Randomization: Centralized 1:1 allocation; the report does not describe stratification or the sequence-generation algorithm in detail.
Setting: Five Italian centers.
Enrollment: June 2021–May 2025.
Follow-up: Scheduled assessments at baseline, two weeks, and six weeks; a separate mean follow-up duration was not reported.
Analysis: Reported outcomes concern the 46 completers, rather than a demonstrated all-randomized intention-to-treat analysis. Friedman tests assessed within-group change; Wilcoxon post-hoc comparisons used a Bonferroni threshold of P < 0.0167. Mann–Whitney U tests compared between-arm change scores.
Primary outcome: Improvement on DOSS, with FEES-based scoring at two and six weeks. Higher scores indicate better swallowing function. The report does not clearly distinguish one of these follow-up times as the single primary time point.
Power: Calculated for within-group pre–post improvement, not the randomized active-versus-sham contrast; modest incremental benefits could therefore be missed.
Registration: NCT07152899, registered retrospectively. The authors state that the protocol, primary outcome, and analysis plan were set before analysis. Methods.

Population

Inclusion Criteria

Age at least 18 years; first or recurrent neuroimaging-confirmed ischemic stroke.
Supratentorial or infratentorial lesions; stroke onset 72 hours to six months before enrollment.
Oropharyngeal dysphagia, defined as DOSS ≤ 6 on fiberoptic endoscopic evaluation of swallowing (FEES).
Ability to participate in SLT and provide written informed consent.

Exclusion Criteria

Pre-existing dysphagia unrelated to stroke.
Severe cognitive impairment preventing therapy participation.
Progressive neurological or oncological disease.
Study-specified tDCS contraindications: metallic cranial implants, pacemakers or other implanted electronic devices, seizure history or active epilepsy, or skin disease at electrode sites.

Baseline Characteristics (46 completers)

Mean age 65.2 years, SD 13.4; 14 women and 32 men.
20 infratentorial and 26 supratentorial strokes.
Mean NIHSS 9, SD 5; mean DOSS 3.0, SD 1.6.
Infratentorial patients had worse swallowing impairment despite lower NIHSS: mean DOSS 2.0 vs 3.7, PAS 5.1 vs 2.4, and NIHSS 6.8 vs 10.6, compared with supratentorial patients.
Enrollment required access to a six-week intensive program, limiting applicability to people unable to continue such rehabilitation. Table 1 and eligibility criteria.

Interventions

Active tDCS plus SLT: Bilateral 5 × 7 cm anodal electrodes over swallowing sensorimotor cortices, centered 3.5 cm lateral and 1 cm anterior to the vertex; 10 × 10 cm supraorbital cathodes contralateral to each anode. 1.5 mA for 20 minutes, once daily, five days per week for two weeks: 10 sessions.
Sham tDCS plus SLT: Same electrode placement and session schedule, with current delivered for 30 seconds at the beginning and 30 seconds at the end to reproduce cutaneous sensations.
Identical intensive SLT in both arms: 40 minutes daily, five days per week for six weeks, concurrent with stimulation during the first two weeks. Individualized swallowing exercises, compensatory/postural strategies, and strengthening continued without stimulation for the remaining four weeks. Intervention protocol.

Outcomes

Primary Outcome — DOSS

Assessment	Active tDCS, mean ± SD	Sham, mean ± SD
Baseline	2.9 ± 1.8	3.1 ± 1.5
Two weeks	4.0 ± 2.0	3.9 ± 1.9
Six weeks	4.7 ± 2.1	4.4 ± 2.0

Both arms improved over time, but between-arm DOSS change comparisons were not statistically significant at either follow-up. Calculated from the rounded table means, the six-week gain was approximately 1.8 vs 1.3 points, an approximately 0.5-point difference in mean change. This descriptive calculation is not the trial's Mann–Whitney treatment-effect estimate and has no reported confidence interval. HR, OR, and RR are not the appropriate reported measures for this score-based analysis; an overall between-arm effect estimate with 95% CI and exact P value is not supplied in the main narrative. Table 2 and Figures 3–4.

Secondary Outcomes

PAS: Mean scores, where lower is better, changed from 3.4 to 2.6 in the active arm and 3.8 to 2.9 with sham by six weeks. No significant overall active-versus-sham advantage was found.
MASA: Rounded mean scores rose from 158 to 176 with active stimulation and 154 to 169 with sham by six weeks, without significant overall superiority of active stimulation.
SWAL-QoL: Improved in both arms at six weeks; between-arm change did not differ significantly. Improvement in quality of life should not be substituted for evidence of fewer pneumonias or less tube feeding.

Exploratory Subgroups — not confirmatory prespecified efficacy claims

Infratentorial lesions: MASA between-arm P = 0.04 at two weeks, 0.08 at six weeks. DOSS P = 0.07 and 0.16, respectively; PAS P = 0.5 and 1, respectively. Numerical subgroup treatment-effect estimates and 95% CIs are not reported in the main narrative. A formal treatment-by-lesion interaction establishing differential efficacy was not reported.
Supratentorial lesions: No significant between-arm differences were reported across outcomes; the narrative does not supply exact subgroup effect estimates, CIs, or all P values.
Early versus late treatment: Therapy started at 72 hours–four weeks versus five weeks–six months. No significant differences in improvement were reported, overall or within either arm. This is not proof that treatment timing is irrelevant.
Sex: Men had larger MASA and SWAL-QoL gains in pooled analyses, against worse baseline dysphagia; this is not evidence of a male-specific randomized tDCS benefit.
Baseline severity: Associations with larger subsequent gains are exploratory and may reflect recovery opportunity or regression to the mean, not a validated biomarker of stimulation response. Subgroup and correlation analyses.

Adverse Events / Safety

No adverse events or side effects were reported; transient sensory experiences were comparable between groups. Blinding success was not formally quantified. Forty-six completers cannot establish rare-event safety, and the trial does not establish reductions in aspiration pneumonia, mortality, or feeding-tube dependence. Hemorrhage was not the treatment question.

Figures

*Figure 1.* Study timeline. Temporal sequence of bilateral anodal transcranial direct current stimulation (tDCS) and intensive speech‐language therapy (SLT) interventions, together with the timing of outcome assessments. DOSS, Dysphagia Outcome and Severity Scale; MASA, Mann Assessment of Swallowing Ability; PAS, Penetration–Aspiration Scale; SWAL‐QoL, Swallowing Quality of Life questionnaire.

*Figure 2.* Participant flow. CONSORT diagram depicting the recruitment, randomization, and analysis of patients with post‐stroke oropharyngeal dysphagia across five Italian centers.

*Figure 3.* Changes in swallowing outcomes. Delta (Δ) values for the Dysphagia Outcome and Severity Scale (DOSS) and the Penetration–Aspiration Scale (PAS) at 2 and 6 weeks in the bilateral anodal tDCS and sham groups. Individual data points are shown for the overall cohort as well as for infratentorial and supratentorial subgroups. Positive Δ values indicate improvement in swallowing function (DOSS), whereas negative Δ values indicate reduced airway penetration and aspiration events, reflecting improved swallowing safety (PAS).

*Figure 4.* Changes in swallowing ability and quality of life. Delta (Δ) values for the Mann Assessment of Swallowing Ability (MASA) at 2 and 6 weeks and for the Swallowing Quality of Life questionnaire (SWAL‐QoL) at 6 weeks in the bilateral anodal tDCS and sham groups. Individual data points are shown for the overall cohort as well as for infratentorial and supratentorial subgroups. Positive Δ values indicate improvement in overall swallowing ability (MASA) and in self‐perceived swallowing‐related quality of life (SWAL‐QoL). The asterisk denotes a statistically significant difference identified in exploratory subgroup analyses.

Criticisms

Wrong contrast for the power calculation: Detecting within-arm improvement does not adequately power the question of adding stimulation to the same therapy.
Within-group improvement is not a causal SLT effect: There was no untreated or lower-intensity therapy arm; spontaneous recovery and therapy contributions cannot be separated.
Completer analysis and retrospective registration: Two sham dropouts were excluded from the reported completed cohort. Retrospective registration weakens independent verification of prospective endpoint and subgroup choices.
Subgroup multiplicity and imprecision: One nominal P = 0.04 on a secondary scale at one time point, without a confirmatory interaction or demonstrated multiplicity control for the subgroup search, is fragile evidence.
Short follow-up and selected population: Six weeks cannot establish durable functional benefit. Hemorrhagic stroke, the hyperacute phase, and patients more than six months from stroke were outside eligibility.
Reporting limitations: Between-arm CIs and clinically anchored responder outcomes would be more informative than within-arm significance. Some baseline and follow-up tabulations are not fully concordant; no unreported precision should be inferred.

Funding

The funding statement names the Italian Ministry of Health, Ricerca Corrente 2025–2027. Acknowledgments also identify collaboration funded through Bando Ricerca Finalizzata 2019, project GR-2019-12369182; open-access publication was facilitated by the University of Pavia's Wiley–CRUI-CARE agreement. A specific funder-role statement is not provided in the retrieved report. Shaheen Hamdy disclosed being Chief Scientific Officer of Phagenesis Ltd. and holding stock/shares; other authors reported no competing financial interests. Registration: NCT07152899, retrospective. Funding and disclosures.

The paper

Authors. Giuseppe Cosentino et al.
Title. Intensive Rehabilitation With Adjunctive Bilateral Anodal tDCS in Post-Stroke Dysphagia: A Multicenter Randomized Controlled Trial.
Journal. European Journal of Neurology, 33(7), e70686.
Year. 2026.
DOI. 10.1111/ene.70686
PMCID. PMC13321463

Deep Dive — click to expand

What this is

A small but well-controlled test of an add-on, not a comparison of rehabilitation with no treatment. The July 2026 trial gave both arms the same intensive swallowing program and found no demonstrated overall incremental benefit from bilateral anodal tDCS. Its value is the separation between a plausible neuroplasticity mechanism, improvement during recovery, and the randomized evidence actually needed to justify an extra intervention. All commentary below is educational interpretation of the trial, not a current-guideline update.

1. Shadow Audit

The headline's rehabilitation success can obscure two different inferential gaps. First, DOSS improved from 2.9 to 4.7 with active stimulation, but it also improved from 3.1 to 4.4 with sham: improvement is not attribution. Second, attributing both arms' gains primarily to SLT goes beyond the randomized comparison because every participant received SLT while spontaneous recovery remained possible.

The subgroup rescue is similarly narrower than it sounds: infratentorial MASA reached P = 0.04 at two weeks, while primary-outcome DOSS did not (P = 0.07), and the MASA difference did not reach nominal significance at six weeks (P = 0.08). This is a proposal for a new trial, not evidence to select a proven responder group. Trial results and discussion.

2. Inversion Engine

For the opposite interpretation—routine incremental efficacy—to become convincing, the active-versus-sham contrast must be both clinically meaningful and precise, not merely positive. The approximately 0.5-point DOSS difference in mean six-week change, calculated from rounded means, could represent a useful effect or sampling variability; the report provides no CI with which to settle that distinction. A future trial should define a clinically meaningful DOSS/responder threshold prospectively and power the randomized contrast to detect it.

For a claim of equivalence, the reverse evidentiary burden applies: a prespecified equivalence margin and a CI excluding important differences would be needed. Neither was supplied. As a multiplicity illustration—not a reanalysis of the study—correcting just four independent tests by Bonferroni would use 0.0125, below the subgroup's 0.04. The actual exploratory family was more complex; this calculation only shows why the nominal result is not robust confirmation.

3. Second-Order Catalyst

The first institution likely to change should be a neurorehabilitation service considering purchasing equipment or turning adjunctive stimulation into routine scheduling. The operational response is an evidence review, not removal of access to therapy: preserve swallowing assessment and structured rehabilitation capacity, and do not count an unproven tDCS increment as established throughput benefit. This is an editorial implementation inference, not an effect tested by the trial.

The research service that should move first is one with enough infratentorial dysphagia cases to run a prospectively registered, lesion-stratified study. Its design should test the treatment-by-lesion interaction and include tube-feeding, pneumonia, and durable patient-centered outcomes—not simply repeat numerous within-subgroup P values.

4. Asymmetric Leverage

The protocol contains 30 planned SLT sessions, or 1,200 therapy minutes, versus 200 stimulation minutes across ten sessions, with stimulation delivered concurrently during the first two weeks. Those calculated quantities describe the tested package, not comparative effectiveness, marginal staffing time, or cost savings. They expose the practical asymmetry: buying a visible device is easier to notice than ensuring six weeks of consistent rehabilitation, yet the device's added contribution was not demonstrated here. Intervention schedule.

The potentially large payoff in a small denominator is severe infratentorial dysphagia, but 20 such completers across both arms do not establish that payoff. The useful leverage now is better trial enrollment and outcome measurement, not a promise of a selectively effective treatment.

5. Paradigm Destroyer

The habit this paper challenges is calling an adjunct effective because patients improve while receiving it. The same improvement can occur with sham, and an underpowered negative comparison does not automatically prove the adjunct useless either.

An educational two-sentence protocol-review note would be: “Do not describe bilateral anodal tDCS as a proven routine add-on on the basis of this trial; preserve access to structured swallowing rehabilitation and individualized swallowing-safety assessment. If stimulation is considered in a research or existing specialist pathway, record uncertainty, lesion location, treatment timing, and prospective outcomes rather than citing the exploratory infratentorial signal as confirmation.” This is an evidence-appraisal implication, not patient-specific treatment advice or an assertion of present guideline policy.

MVP — Minimum Viable Proof

The minimum statement already supported is: 48 randomized, 46 completed, and no demonstrated overall DOSS superiority of active over sham stimulation at two or six weeks despite improvement in both arms. The minimum new proof needed to change the add-on conclusion is a sufficiently powered, prospectively registered randomized estimate with a confidence interval and a clinically interpretable outcome. A subgroup-specific practice claim additionally needs confirmation that lesion location modifies treatment response, not just a nominally significant within-subgroup comparison.

Best Combination

Read this alongside the original 2017 Ahn et al. chronic-stroke trial, not as proof that every tDCS montage and recovery phase has the same effect. That study randomized 26 patients at least six months after stroke and reported no significant between-arm DOSS difference (P = 0.48), despite significant improvement within its active arm. The directly retrieved original report makes the same methodological lesson visible in a different recovery window: “significant in one arm, not in the other” is not a significant treatment contrast.

The present study adds a multicenter sample and a longer intensive SLT program in the 72-hour–six-month window. Together, these trials justify skepticism about claiming established bilateral add-on efficacy, while leaving modest effects, different targeting strategies, and a prospectively defined infratentorial hypothesis unresolved. They do not justify pooling unlike protocols informally or declaring all neurostimulation ineffective.

Overvalue Warning

Do not turn the infratentorial P = 0.04 into a treatment rule. It concerned MASA at two weeks, not confirmed primary-outcome superiority, durable benefit, or a formal interaction.
Do not equate “no significant difference” with equivalence, or both-arm improvement with proof of the SLT dose's causal effect. The study was small, powered around within-group change, and had no untreated comparison. Its most secure conclusion concerns what the randomized add-on comparison failed to demonstrate.

Educational use notice. The Time Is Brain Review is dedicated to medical education. Figures from published articles are embedded for teaching purposes, with attribution to the original publication. We do not claim copyright over any embedded figure; all rights remain with the original publisher and authors under the article's published license. If you are a rights holder and would like a figure removed, please contact the editorial team. See our editorial notice for the full policy.