Stroke is a leading cause of global disability, and motor impairment of the upper limb remains a primary target for rehabilitation. The literature reports clinically meaningful motor improvements associated with placebo responses in post-stroke rehabilitation, but the neurophysiological mechanisms—especially differences between open-label placebo and covert/sham approaches—are not well defined. Neuroplasticity-based frameworks implicate prefrontal networks and dopaminergic corticostriatal circuitry in placebo-driven modulation of motor recovery. The protocol aims to link these anatomical and psychological models to measurable electrophysiological signatures using high-density EEG and transcranial magnetic stimulation.
The principal aim is to characterise neurophysiological signatures of hidden and open placebo during motor-cortex stimulation in chronic stroke survivors and to determine how these signatures correlate with motor recovery. The study hypothesises that placebo mechanisms mediated by expectation and reward circuits will modulate prefrontal activity, reflected as changes in frontal alpha asymmetry (FAA), and increase beta-band premotor-to-M1 connectivity, potentially enhancing motor outcomes.
This is a single-centre, controlled, parallel-group, randomised clinical trial enrolling 56 adults ≥6 months post-stroke with upper-limb hemiparesis. Participants are block-randomised using a web-based programme in a 2:2:2:1 allocation to four groups: OLP (16), sham rTMS (16), no intervention (16), and active rTMS (8). Randomisation sequence will be concealed in sequentially numbered, opaque, sealed envelopes. Outcome assessors remain blinded; stimulation personnel and randomisation staff are unblinded. Subjects in rTMS arms are blinded to stimulation condition, while OLP and no-intervention arms are unblinded by design.
All procedures occur at the Neuromodulation Center, Spaulding Rehabilitation Hospital, Cambridge, MA. Inclusion criteria include adults with chronic hemiparesis (≥6 months), Fugl–Meyer upper-extremity score >11 and ≤56, pre-stroke Modified Rankin Scale <3, and age ≥18. Exclusion criteria include anterior cerebral artery territory strokes with prefrontal lesions or marked reduction in prefrontal EEG power (because FAA depends on reliable prefrontal signals), inability to follow instructions, TMS contraindications (e.g., cochlear implant), unstable medical conditions, HDRS ≥24, interfering pain, and pregnancy.
Active rTMS: Low-frequency (1 Hz) stimulation over the contralesional primary motor cortex at 80% motor threshold, 1200 pulses delivered in a single 20-minute continuous train. This protocol has been used in chronic stroke populations and is positioned as a safer modulation for chronic stroke.
Sham rTMS: Same coil placement and parameters but with a sham coil to prevent cortical stimulation while preserving procedural parity.
Open-label placebo (OLP): Participants receive placebo pills (microcrystalline cellulose) in a labelled prescription bottle instructing two pills twice daily. Investigators provide a brief semi-scripted, transparent rationale emphasising uncertainty of benefit to standardise communication and manage expectations.
No intervention (treatment as usual): Participants continue standard care and rehabilitation without additional experimental intervention. Semi-scripted communication explains the critical role of this control group.
After the main participation, participants in any non-active-intervention groups (sham rTMS, OLP, no intervention) are offered a task-specific, upper-limb, home-based rehabilitation programme consistent with NIH/NINDS and American Stroke Association recommendations. The programme includes three supervised lab sessions then 21 unsupervised home sessions (3×/week for 7 weeks), stratified by impairment level to reduce disappointment and potential nocebo effects.
Primary outcome: change in FAA measured by high-density EEG from baseline to two weeks.
Secondary outcomes: β-band premotor-to-M1 coherence and Fugl–Meyer Upper Extremity (FM) motor scores.
Exploratory outcomes: additional motor tests (Jebsen–Taylor Hand Function Test, Nine-Hole Peg Test), corticospinal excitability via single- and paired-pulse TMS, and patient-reported measures including mood, expectancy, and behavioural activation–inhibition scales. All outcomes are assessed from baseline to two weeks.
EEG uses a 64-channel actiCHamp system. Recording includes setup, 5 minutes eyes-open, 5 minutes eyes-closed, and an 8-minute task-related conditioning composed of movement observation (MO), movement imagery (MI) and movement execution (ME) trials. Event-related desynchronisation is assessed to index motor cortex activation. Data are preprocessed in EEGLAB and MATLAB with downsampling to 250 Hz, re-referencing, filtering (1–60 Hz), ICA-based artefact correction, visual inspection and FFT-based power extraction. FAA is computed as the difference between natural log-transformed alpha power in right frontal leads (F4, F8, Fp2) minus left frontal leads (F3, F7, Fp1).
Motor function is measured with the validated Fugl–Meyer motor scale for the upper extremity, with additional timed and dexterity assessments. Corticospinal excitability is assessed with resting motor threshold and paired-pulse measures using TMS. Patient-reported measures of mood and expectancy are collected to characterise psychological mediators.
Fifty-six participants will be analysed according to intention-to-treat principles. Multivariable regression models will adjust for age, sex, stroke severity, lesion location and time since stroke. The protocol indicates data availability upon reasonable request. Specific sample size justification and statistical power details are reported in the full protocol (details not repeated here).
The study follows the Declaration of Helsinki and received Mass General Brigham IRB approval (protocol #2023P000884). All participants provide informed consent. The trial is registered (NCT05832567). Results will be published in peer-reviewed journals and presented at scientific meetings. Data are available on reasonable request.
Strengths include a rigorous parallel-group, block-randomised design with four arms, and combined use of high-density EEG and TMS to capture cortical activation and corticospinal excitability. Limitations include potential partial unblinding due to sensory differences between real and sham low-frequency rTMS and potential expectancy spillover across arms that may dilute the psychological purity of hidden versus open placebo conditions.