Just out todayAugmented reality: The AR trainees were faster on the model they trained on, and no better anywhere elseVirtual reality: They matched the exercise dose exactly, and the VR advantage fell below the line that mattersVirtual reality: Both groups used the same headset. One watched a video first and picked up 10.8 more blocks

Biotech & health tech/USA

452 Stroke Patients Used A Brain Device At Home. Read Week 12 Carefully

The largest real-world look at home brain-computer interface therapy so far: 452 stroke patients. At week 12 arm scores rose 4.3 points. The study's own threshold for a noticeable difference is 5.25.

What the paper found

Among 452 stroke patients using at-home BCI therapy in routine care, mean Upper Extremity Fugl-Meyer scores rose 4.3 points by week 12 (95% CI 3.7-4.9, P<0.001), short of the 5.25-point minimal clinically important difference; 31.7% of patients reached it individually. The mean crossed that threshold at week 16 and kept rising to 8.1 points at week 52. There was no control group.

Grade 6 reading level5 min readPreprint · not yet peer reviewed

What happened

A stroke can leave an arm weak for years. For severe, long-standing cases there is very little that works.

One idea is a brain-computer interface. Sensors read the electrical activity that happens when you try to move. A device decodes that intention and moves a brace on your hand, so the attempt produces a real movement you can feel. Doing that over and over appears to help the brain rebuild the link.

Randomised trials have shown this helps, including for severe weakness. But a trial is a careful setting with selected patients and paid staff. This team asked a different question: what happens when real people use it at home, prescribed by their own clinician?

The test

The team looked back at adults prescribed the IpsiHand System through a commercial home therapy programme in the United States. Looking back at records that already exist is called a retrospective observational study.

They kept patients who had a baseline measurement and who averaged more than one session a week. Occupational therapists scored the arm by video every four weeks, using a standard scale called the Upper Extremity Fugl-Meyer, or UEFM.

452 patients ended up in the analysis. On average their stroke was 5.4 years in the past, so this is long-standing weakness, not fresh recovery. Average starting score was 20.1. By the study’s own bands, 333 of them — 73.7 percent — were severely impaired.

The result

At week 12, the main measure, average scores had risen 4.3 points. The confidence interval ran from 3.7 to 4.9, and the p value was below 0.001. So the rise is real and not a fluke.

Here is the sentence that matters. The paper states that the smallest change a patient would actually notice — researchers call it the minimal clinically important difference — is 5.25 points on this scale. The week 12 average of 4.3 did not reach it. And 31.7 percent of patients individually did.

So at the three month mark, which was the study’s primary outcome, the average patient had not gained enough to feel a difference, and roughly two in three had not either.

Then it kept going. The average crossed that threshold at week 16, at 5.6 points. At week 24 it was 6.1. At week 52 it was 8.1 points.

Why does the ordering matter so much? This is because the headline most people would write from this paper is “brain device helps 452 stroke patients”. That is true at one year. At three months, the number everyone quotes, it is not yet true. THE CALENDAR IS PART OF THE CLAIM.

What it means

Two honest readings sit side by side, and both belong in your head.

The encouraging one: the gains did not plateau. They grew all the way to a year, in patients whose strokes were years old and who were mostly severely affected. That is unusual and genuinely hopeful.

The cautious one: this is a study with no comparison group. Nobody was given a fake device, and nobody was simply watched for a year. So you cannot tell how much of the 8.1 points came from the device and how much from doing an hour of structured arm work several times a week for twelve months. The authors say so themselves, and end by saying it merits testing in randomised trials with longer follow-up. Therefore the fair summary is: promising, uncontrolled, and the authors agree.

Business ideas from this paper

1. Adherence support for home neuro-rehab. Every patient here averaged more than one session a week — that was the entry condition. In real life most people stop. Sell the phone calls, reminders and scheduling that keep people going for a year. Buyer: device makers and home therapy programmes paid on outcomes. First test: offer one programme a 90-day adherence pilot and report the drop-out rate against their own baseline.

2. Video assessment as a service. Therapists scored these patients by video every four weeks. That is a repeatable, trainable job that clinics struggle to staff. Build the trained panel and sell scored assessments. Buyer: rehab programmes and trial sponsors. First test: score 20 videos for one clinic and compare your raters against theirs.

3. Plain-language outcome reports for patients and families. A family told “scores improved” cannot tell whether that means anything. Sell the one-page explanation of what a score change means and when to expect it. Buyer: rehab clinics and patient groups. First test: write one for a single clinic and ask whether it reduced repeat questions.

How sure can you be?

Be interested, not convinced.

The strengths are real: 452 patients is far larger than most brain-interface studies, the scale used is standard, the assessments were repeated every four weeks for a year, and the statistical handling of missing data was spelled out.

The weaknesses are equally real. It is a preprint, so no peer review yet. There is no control group, which is the central limit. Everyone included was already using the device regularly, so people who gave up are not represented, and that alone makes results look better. Measurement was by video, not in person.

And there is a declared conflict you should know about: the first author is employed by Kandu, Inc., which makes the device, and another author lists Kandu among many consulting clients. That does not make the numbers wrong. It does mean the result should be confirmed by people with nothing riding on it.

Do this today

Next time you read that a treatment “significantly improved” something, look for two numbers: the size of the change, and the size a patient would notice. If the second one is missing, the first one is decoration.

Source: A Retrospective Real-World Assessment of Motor Status After Stroke Using a Home-Based Brain-Computer Interface, medRxiv, October 2026. 10.64898/2026.10.30.26366945 · doi.org (preprint · not yet peer reviewed).

Just Out Tech explains new research in plain language. This article was drafted with AI assistance and checked by a human against the original source.

What to remember
  • In 452 patients prescribed at-home BCI therapy (medRxiv, DOI 10.64898/2026.10.30.26366945), mean time since stroke was 5.4 years and 333 patients (73.7%) had severe impairment.
  • At week 12, the primary endpoint, mean UEFM change was 4.3 points (95% CI 3.7-4.9, P<0.001), below the 5.25-point minimal clinically important difference the paper states.
  • 31.7% of patients individually reached that 5.25-point threshold by week 12.
  • The mean change crossed the threshold at week 16 (5.6 points) and continued rising to 6.1 at week 24 and 8.1 at week 52.
  • The study has no control group, included only patients already averaging more than one session a week, and the authors conclude the approach merits randomised trials.
Where this really is

The IpsiHand System is a commercial product prescribed through a US home therapy programme. This study looks back at records from routine use, not at an experiment.

Who paid for the research
The paper's competing-interest statement names the first author as an employee of Kandu, Inc., which manufactures the IpsiHand System studied here. Another author lists Kandu among many consulting clients. A positive result benefits the manufacturer.

Questions people ask

Did the therapy work?

Scores improved and kept improving to one year. Without a control group you cannot tell how much came from the device and how much from doing structured arm exercise several times a week for a year.

What is the 5.25-point threshold?

It is the minimal clinically important difference, the smallest change on this scale a patient would actually notice. The paper states it.

Why does week 12 matter more than week 52?

Week 12 was the study's primary outcome, the measure it was designed to test. The later figures come from a statistical model of continued follow-up.

About the author

Mark Alex

Mark Alex is the founder and Managing Director of Real Biz Digital, a technology company operating out of Nairobi since 2018. He works in agentic AI and the Model Context Protocol, AI governance, enterprise software architecture and cybersecurity. He holds an MSc in Mechatronical Engineering from Obuda University in Budapest and a BSc in IT, Forensic Technology and Cybercrime, from USIU-Africa in Nairobi, and has published IEEE conference research on an AI-powered digital twin for greenhouse systems. He is the author of seven books. Between 2020 and 2024 he mentored more than 200 university students and interns in Nairobi. He writes every Just Out Tech article from the original research paper.