Study analysis · Journal of neural engineering · 2018
A 77% improvement with fake brain stimulation: why a stroke rehab study with only 3 patients exposes the messy reality of neuroscience.
In a tiny study of three stroke patients, combining magnetic brain stimulation with a brain-computer interface changed brain activity in the right direction, but the patient who got fake stimulation actually improved more on some hand movement tests.
Overview
What the study found
The study in plain English — the bottom line, every takeaway we extracted, and what to do with them.
In simple terms
This study is like trying a new recipe on just 3 people. We can see if they liked it, but we can't say it will work for everyone. It's a first step to see if the idea is worth testing on more people, not proof that it works.
What’s the bottom line?
This study tested a new therapy for stroke patients: using a brain-computer interface (a device that reads brain waves) together with a gentle magnetic pulse to the brain. Three people with stroke did this therapy for six weeks. The therapy helped patients use the damaged side of their brain more, but because there were only three people, we can't be sure it works for everyone.
How strong is this study?
The study was well-planned with a blindfold test and random assignment, but only 3 people took part. That's like flipping a coin three times and thinking you know the pattern. We need many more people to be sure the results aren't just luck.
0 / 100
- COI disclosureconflicts of interest not disclosed
- Data availabilitydata not shared
- Code availabilitycode not shared
75 / 100
- Randomization+20/20
- Blinding+15/15
- Control group+15/15
- Sample size (n=3)+0.3/20
- Follow-up+10/10
100 / 100
54 / 100
- P-values+15/15
- Effect size+20/20
- Confidence intervalsno confidence intervals
- Pre-registrationnot pre-registered
Each component is scored out of 100 and then capped by the study design — a case series cannot reach the ceiling a randomised trial can, however well it is reported.
Where it sits
RCT reviewsReviews of RCTs (Meta-analyses)
Max 100Randomized TrialsRandomized Trials
Max 90Reviews of Cohort StudiesReviews of Cohort Studies
Max 85Cohort StudiesCohort Studies
Max 72Reviews of Case-Control StudiesReviews of Case-Control Studies
Max 63Case-Control StudiesCase-Control Studies
Max 58Cross-Sectional & Case SeriesCross-Sectional & Case Series
Max 50Expert OpinionExpert Opinion
Max 560 / 100
Probability of being correct
Participants are randomly assigned to treatment or control groups, minimizing bias. The gold standard for testing whether an intervention causes an effect.
This design cannot establish causation — the findings describe an association, not a cause. The study has only 3 subjects, making it essentially a case series. Randomization and blinding are present but with such a small sample, the results cannot establish causation. There is high risk of confounding and chance findings.
No Conflicts
No conflicts of interest identified
No conflicts of interest identified; study appears to be independently conducted at an academic institution with no disclosed industry funding or author affiliations.
The provided text does not include a Conflict of Interest or Funding section. The study is small (n=3) and conducted at the University of Minnesota. Equipment from Magstim and National Instruments is used, but these are standard commercial tools and not necessarily indicative of a conflict. No author names or affiliations are given, so no specific author-industry ties can be assessed.
Key takeaways
- 01
Only the two patients who got the real magnetic pulse showed changes in how their brain hemispheres talk to each other (from the healthy side to the damaged side) and more brain activity on the damaged side.
- 02
Their brain-computer control accuracy improved by 7.5%.
- 03
All three patients got better at finger tracking, but the two who got the real pulse improved 51%, and one who got the fake pulse improved 77%.
- 04
The results are promising but not definitive because there were only three people.
- 05
Both groups got better, so it's unclear if the magnetic pulse adds much benefit.
- 06
The patient who got the fake pulse had milder problems to begin with, which might explain their bigger improvement.
Surprising findings
- The sham patient showed a 77% improvement in finger tracking accuracy, while the real rTMS group improved only 51%.Common sense says the real treatment should outperform the fake one. But the sham patient had milder deficits, making the real group's smaller absolute gains actually more impressive in context.
- The sham patient actually increased inhibition from the healthy hemisphere to the damaged one over the course of the study (65.4% increase), opposite to what the real rTMS group achieved.This suggests that BCI training alone, without brain stimulation, might actually worsen the brain's maladaptive compensation pattern in some patients, potentially making recovery harder.
Practical takeaways
If you're recovering from a stroke, consider asking your therapist about motor imagery or BCI training – even without brain stimulation, it may help.
This study had only 3 participants, so the evidence is extremely weak. Do not abandon standard rehab without consulting a doctor.
low confidenceFor clinicians: measuring inter-hemispheric inhibition (via paired-pulse TMS) could help identify which patients might benefit from rTMS.
This requires expensive equipment and expertise, and the predictive value is not yet proven in larger studies.
low confidenceFor researchers: future studies must match groups on baseline severity and include a no-treatment control. The sham patient's 77% improvement is a red flag for confounding.
Recruiting matched stroke patients is notoriously difficult, as the authors note.
high confidenceWhy this study matters
The Brain's Balancing Act After Stroke
The study tested whether suppressing the healthy side of the brain with magnetic pulses (rTMS) before training with a brain-computer interface (BCI) could help stroke patients recover hand function. The two patients who got real rTMS showed the desired shift: reduced inhibition from the healthy hemisphere to the damaged one (measured by paired-pulse TMS, a 35.8% change) and increased activity in the damaged side on fMRI.
This gives a clear physiological target for treatment – it's not just 'exercising the hand,' but retraining the brain's hemispheres to talk to each other properly.
The Sham Patient Did Better?
The patient who received sham (fake) rTMS plus BCI training improved a whopping 77% on the finger tracking test, compared to an average 51% improvement in the real rTMS group. However, the sham patient had much milder hand impairment at the start, making the comparison unfair.
This highlights a huge problem in stroke rehab research: patients with less severe deficits often show the biggest gains, making it hard to tell if the treatment is actually working.
BCI Training Alone Might Be Enough
The sham patient, who only did BCI training (imagining moving their hand to control a virtual reality cursor), improved significantly on the Box and Block test (24% increase) and finger tracking. This suggests that just the mental practice of using a brain-computer interface, without any brain stimulation, can drive recovery.
BCIs are cheaper, safer, and more portable than TMS machines. If they work alone, it could democratize stroke rehab.
Want the whole report?
Detailed mode opens the full scientific breakdown — every score component, the methodology, conflicts of interest, the evidence analysis behind each claim, and the raw study data.
Overview
What the study found
The study in plain English — the bottom line, every takeaway we extracted, and what to do with them.
Not medical advice. For informational purposes only. Always consult a healthcare professional. Terms
This study tested a new therapy for stroke patients: using a brain-computer interface (a device that reads brain waves) together with a gentle magnetic pulse to the brain. Three people with stroke did this therapy for six weeks. The therapy helped patients use the damaged side of their brain more, but because there were only three people, we can't be sure it works for everyone.
Research results
Only the two patients who got the real magnetic pulse showed changes in how their brain hemispheres talk to each other (from the healthy side to the damaged side) and more brain activity on the damaged side. Their brain-computer control accuracy improved by 7.5%. All three patients got better at finger tracking, but the two who got the real pulse improved 51%, and one who got the fake pulse improved 77%.
What this means - more context
The results are promising but not definitive because there were only three people. Both groups got better, so it's unclear if the magnetic pulse adds much benefit. The patient who got the fake pulse had milder problems to begin with, which might explain their bigger improvement.
To evaluate whether combining low-frequency rTMS over the contralesional motor cortex with motor imagery-based BCI training improves motor recovery after chronic stroke compared to sham rTMS with BCI training.
Three chronic stroke patients (two real rTMS+BCI, one sham+BCI) completed a 6-week protocol. The real rTMS+BCI group showed desired changes in inter-hemispheric inhibition (reduced contralesional-to-ipsilesional inhibition, increased ipsilesional-to-contralesional inhibition), increased ipsilesional fMRI activation, and improved BCI performance (7.5% increase, p<0.001). Both groups showed behavioral improvements, but the sham subject also improved. Findings are preliminary due to very small sample size.
Methods Used
Three chronic stroke patients (~1 year post-stroke) were randomized to real or sham 1Hz rTMS (90% rMT, 10 min) over contralesional M1 followed by motor imagery-based BCI training (8-10 runs of 20 trials). Treatment: 3 weeks of combined rTMS+BCI then 3 weeks of BCI-only. Assessments: Box and Block test, finger tracking test, inter-hemispheric inhibition (paired-pulse TMS), and fMRI laterality index at baseline, after 3 weeks, and after 6 weeks.
Main Finding
Only the real rTMS+BCI group showed significant alterations in inter-hemispheric inhibition in the desired direction (35.8% increased ipsilesional-to-contralesional inhibition, p<0.01) and increased ipsilesional cortical activation (fMRI laterality index). BCI performance improved only in the real rTMS+BCI group (+7.5%, p<0.001). Both groups improved on finger tracking (real: +51%, p<0.01; sham: +77%, p<0.001), but direct comparison is confounded by baseline differences.
Confidence Level
Very low: sample size of 3 (unbalanced: 2 real, 1 sham) precludes generalizability; baseline motor function differed substantially between groups; statistical tests on very few data points inflate uncertainty.
Study Flags
Red Flags
- •Extremely small sample size (n=3, unbalanced groups)
- •Substantial between-group differences in baseline motor function complicates comparison
- •No control condition for spontaneous recovery or placebo effects
Surprising Findings
The sham patient showed a 77% improvement in finger tracking accuracy, while the real rTMS group improved only 51%.
Common sense says the real treatment should outperform the fake one. But the sham patient had milder deficits, making the real group's smaller absolute gains actually more impressive in context.
Practical Takeaways
If you're recovering from a stroke, consider asking your therapist about motor imagery or BCI training – even without brain stimulation, it may help.
RCT reviewsReviews of RCTs (Meta-analyses)
Max 100Randomized TrialsRandomized Trials
Max 90Reviews of Cohort StudiesReviews of Cohort Studies
Max 85Cohort StudiesCohort Studies
Max 72Reviews of Case-Control StudiesReviews of Case-Control Studies
Max 63Case-Control StudiesCase-Control Studies
Max 58Cross-Sectional & Case SeriesCross-Sectional & Case Series
Max 50Expert OpinionExpert Opinion
Max 560 / 100
Probability of being correct
Participants are randomly assigned to treatment or control groups, minimizing bias. The gold standard for testing whether an intervention causes an effect.
Human RCT
Subject
Moderate probability
on the GRADE evidence scale
This study is like trying a new recipe on just 3 people. We can see if they liked it, but we can't say it will work for everyone. It's a first step to see if the idea is worth testing on more people, not proof that it works.
The study has a COI section but no disclosure was found. A small penalty has been applied.
Strengths
- Randomized design
- Double-blind (participants, testers, treaters)
- Use of objective outcome measures (fMRI, IHI, BCI performance)
Weaknesses
- Very small sample size (n=3)
- Unbalanced groups (2 vs 1)
- Baseline differences between groups
Methodology
Evidence Keywords
Statistical Reporting
Not medical advice. For informational purposes only. Always consult a healthcare professional. Terms
This study tested a new therapy for stroke patients: using a brain-computer interface (a device that reads brain waves) together with a gentle magnetic pulse to the brain. Three people with stroke did this therapy for six weeks. The therapy helped patients use the damaged side of their brain more, but because there were only three people, we can't be sure it works for everyone.
Research results
Only the two patients who got the real magnetic pulse showed changes in how their brain hemispheres talk to each other (from the healthy side to the damaged side) and more brain activity on the damaged side. Their brain-computer control accuracy improved by 7.5%. All three patients got better at finger tracking, but the two who got the real pulse improved 51%, and one who got the fake pulse improved 77%.
What this means - more context
The results are promising but not definitive because there were only three people. Both groups got better, so it's unclear if the magnetic pulse adds much benefit. The patient who got the fake pulse had milder problems to begin with, which might explain their bigger improvement.
To evaluate whether combining low-frequency rTMS over the contralesional motor cortex with motor imagery-based BCI training improves motor recovery after chronic stroke compared to sham rTMS with BCI training.
Three chronic stroke patients (two real rTMS+BCI, one sham+BCI) completed a 6-week protocol. The real rTMS+BCI group showed desired changes in inter-hemispheric inhibition (reduced contralesional-to-ipsilesional inhibition, increased ipsilesional-to-contralesional inhibition), increased ipsilesional fMRI activation, and improved BCI performance (7.5% increase, p<0.001). Both groups showed behavioral improvements, but the sham subject also improved. Findings are preliminary due to very small sample size.
Methods Used
Three chronic stroke patients (~1 year post-stroke) were randomized to real or sham 1Hz rTMS (90% rMT, 10 min) over contralesional M1 followed by motor imagery-based BCI training (8-10 runs of 20 trials). Treatment: 3 weeks of combined rTMS+BCI then 3 weeks of BCI-only. Assessments: Box and Block test, finger tracking test, inter-hemispheric inhibition (paired-pulse TMS), and fMRI laterality index at baseline, after 3 weeks, and after 6 weeks.
Main Finding
Only the real rTMS+BCI group showed significant alterations in inter-hemispheric inhibition in the desired direction (35.8% increased ipsilesional-to-contralesional inhibition, p<0.01) and increased ipsilesional cortical activation (fMRI laterality index). BCI performance improved only in the real rTMS+BCI group (+7.5%, p<0.001). Both groups improved on finger tracking (real: +51%, p<0.01; sham: +77%, p<0.001), but direct comparison is confounded by baseline differences.
Confidence Level
Very low: sample size of 3 (unbalanced: 2 real, 1 sham) precludes generalizability; baseline motor function differed substantially between groups; statistical tests on very few data points inflate uncertainty.
Study Flags
Red Flags
- •Extremely small sample size (n=3, unbalanced groups)
- •Substantial between-group differences in baseline motor function complicates comparison
- •No control condition for spontaneous recovery or placebo effects
Surprising Findings
The sham patient showed a 77% improvement in finger tracking accuracy, while the real rTMS group improved only 51%.
Common sense says the real treatment should outperform the fake one. But the sham patient had milder deficits, making the real group's smaller absolute gains actually more impressive in context.
Practical Takeaways
If you're recovering from a stroke, consider asking your therapist about motor imagery or BCI training – even without brain stimulation, it may help.
RCT reviewsReviews of RCTs (Meta-analyses)
Max 100Randomized TrialsRandomized Trials
Max 90Reviews of Cohort StudiesReviews of Cohort Studies
Max 85Cohort StudiesCohort Studies
Max 72Reviews of Case-Control StudiesReviews of Case-Control Studies
Max 63Case-Control StudiesCase-Control Studies
Max 58Cross-Sectional & Case SeriesCross-Sectional & Case Series
Max 50Expert OpinionExpert Opinion
Max 560 / 100
Probability of being correct
Participants are randomly assigned to treatment or control groups, minimizing bias. The gold standard for testing whether an intervention causes an effect.
Human RCT
Subject
Moderate probability
on the GRADE evidence scale
This study is like trying a new recipe on just 3 people. We can see if they liked it, but we can't say it will work for everyone. It's a first step to see if the idea is worth testing on more people, not proof that it works.
The study has a COI section but no disclosure was found. A small penalty has been applied.
Strengths
- Randomized design
- Double-blind (participants, testers, treaters)
- Use of objective outcome measures (fMRI, IHI, BCI performance)
Weaknesses
- Very small sample size (n=3)
- Unbalanced groups (2 vs 1)
- Baseline differences between groups
Methodology
Evidence Keywords
Statistical Reporting
Scoring
How strong is this study?
The study was well-planned with a blindfold test and random assignment, but only 3 people took part. That's like flipping a coin three times and thinking you know the pattern. We need many more people to be sure the results aren't just luck.
0 / 100
- COI disclosureconflicts of interest not disclosed
- Data availabilitydata not shared
- Code availabilitycode not shared
75 / 100
- Randomization+20/20
- Blinding+15/15
- Control group+15/15
- Sample size (n=3)+0.3/20
- Follow-up+10/10
100 / 100
54 / 100
- P-values+15/15
- Effect size+20/20
- Confidence intervalsno confidence intervals
- Pre-registrationnot pre-registered
Each component is scored out of 100 and then capped by the study design — a case series cannot reach the ceiling a randomised trial can, however well it is reported.
Where it sits
RCT reviewsReviews of RCTs (Meta-analyses)
Max 100Randomized TrialsRandomized Trials
Max 90Reviews of Cohort StudiesReviews of Cohort Studies
Max 85Cohort StudiesCohort Studies
Max 72Reviews of Case-Control StudiesReviews of Case-Control Studies
Max 63Case-Control StudiesCase-Control Studies
Max 58Cross-Sectional & Case SeriesCross-Sectional & Case Series
Max 50Expert OpinionExpert Opinion
Max 560 / 100
Probability of being correct
Participants are randomly assigned to treatment or control groups, minimizing bias. The gold standard for testing whether an intervention causes an effect.
This design cannot establish causation — the findings describe an association, not a cause. The study has only 3 subjects, making it essentially a case series. Randomization and blinding are present but with such a small sample, the results cannot establish causation. There is high risk of confounding and chance findings.
No Conflicts
No conflicts of interest identified
No conflicts of interest identified; study appears to be independently conducted at an academic institution with no disclosed industry funding or author affiliations.
The provided text does not include a Conflict of Interest or Funding section. The study is small (n=3) and conducted at the University of Minnesota. Equipment from Magstim and National Instruments is used, but these are standard commercial tools and not necessarily indicative of a conflict. No author names or affiliations are given, so no specific author-industry ties can be assessed.