Hi!
Welcome back to AIMedily.
Yesterday, OpenAI dropped Sora 2—and it’s making waves across the internet.
This next generation video model can transform simple text prompts into cinematic, realistic video.
Do you think that tools like Sora will become part of medical education—building immersive training simulations or generating patient educational videos?
I’d love to hear your perspective.
Now, let’s dive into this week’s highlights.
🤖 AIBytes
Researchers tested a new hip assisting exoskeleton that uses motion prediction to provide walking support for subacute stroke patients. They compared its effects to conventional gait training.
🔬 Methods
Single blinded randomized trial.
Participants: 40 adults with stroke (subacute). Randomly assigned to either:
RAGT group: Robot-assisted gait training + conventional therapy.
Control group: Therapist-assisted gait training + conventional therapy.
The robot used a motion prediction model that anticipates the user's walking intention and provides assistive force at the hip joint.
Duration: 5x week for 4 weeks.
Assessments:
📊 Results
Robotic therapy led to greater gains in hip and knee flexion and reduced gait asymmetry, outperforming the control group.
Both groups improved gait speed, cadence, and step length, but robotic training produced significantly superior results.
The robotic therapy group achieved higher Fugl-Meyer and Berg Balance scores, and faster TUGT times.
Exoskeleton use reduced paretic limb unloading and enhanced dynamic weight-bearing symmetry.
Step width showed no significant change.
🔑Key Takeaways
Targeting the hip joint with a motor-intention prediction exoskeleton improves gait and balance in patients with subacute stroke.
The robotic therapy improved bilateral hip-knee coordination and gait symmetry.
Real-time motion prediction enabled more adaptive and responsive robotic assistance.
💡This technology could shift gait training to a dynamic, patient-driven interaction, which promotes neurorecovery.
🔗 Li Y, Luo S, Luo R, Liu H. A novel real-time assistive hip-wearable exoskeleton robot based on motion prediction for lower extremity rehabilitation in subacute stroke: a single-blinded, randomized controlled trial. BMC Neurol.2025;25:399. doi:10.1186/s12883-025-04437-5
This study evaluated NeuroSkin®, a textile wearable with an AI-controlled functional electrical stimulation (FES) on subacute stroke patients.
🔬 Methods
Design: Multicenter retrospective study across 7 European rehabilitation centers.
Participants: 15 adults (mean age 60.6 ± 12.6) subacute postroke.
Intervention: 10–20 sessions x 30-45 min. of NeuroSkin® assisted gait therapy + standard physiotherapy (mean 16.1 sessions).
The device delivered multi-muscle, phase-specific FES (gluteus maximus, quadriceps, hamstrings, tibialis anterior, fibularis, gastrocnemius), guided by gait phase detection.
Assessments:
📊 Results
Walking speed (10MWT): +0.35 m/s (+70.3%)
Endurance (6MWT): +123.5 m (+171%)
Mobility (TUG): −36.9 s (−39%)
Ambulation (NFAC): Mean gain of +3 categories.
Usability: 84.6.
Safety: No adverse events reported. Patient satisfaction was 7.3/10.
Watch NeuroSkin® in action.
🔑 Key Takeaways
AI-driven wearable FES improve gait and function in subacute postroke patients.
All gait outcomes improved significantly and were clinically meaningful.
NeuroSkin® was safely implemented across diverse centers with excellent usability.
It can be integrated into stroke rehabilitation with minimal training, which means it can enable real-time, multi-muscle stimulation.
💡 Larger prospective controlled studies are needed to confirm effectiveness.
🔗 Metani A, Popović-Maneski L, Seguin P, Di Marco J. NeuroSkin®: AI-driven wearable functional electrical stimulation for post-stroke gait recovery—A multicenter feasibility study. Sensors. 2025;25:5614. doi:10.3390/s25185614
🦾TechTool
Smart calendar for work and life that automatically schedules tasks and meetings.
It can be used by teams and organizations.
Free on Google and Outlook Calendar.
Motion tracking and feedback for rehabilitation programs and injury prevention (multilingual).
Interactive coaching to keep the patient motivated.
Personalized care with real-time progress tracking so plans adapt as patients improve.
Turns your notes, PDFs, or images into clear summaries.
Can generate quizzes, flashcards, and mind maps from your study material.
Converts notes into audio so you can listen anywhere.
🧬AIMedily Snaps
1️⃣ From breaking point to breakthrough: the $1 trillion opportunity to reinvent healthcare (Link).
2️⃣ An AI telerehabilitation solution to improve mobility on MS (Link).
3️⃣ Physician Vs AI: The Human Side of the Story (Link).
4️⃣ Trump’s recent policy proposals regarding AI (Link).
🧩TriviaRX
In what year did the FDA first clear a robotic exoskeleton for use in stroke and spinal cord injury rehab?
A) 2005
B) 2010
C) 2014
D) 2016
And now, let’s see if you guessed the correct answer from last week’s TriviaRX:
✅ B. In 1917, the National Society for the Promotion of Occupational Therapy (NSPOT) was formed to help recover patients from WWI injuries and improve their mental health.
We’re done for today.
As always, thank you for being here.
Enjoying AIMedily? Don’t keep it to yourself—share it with a colleague and help them stay on top of AI in rehabilitation medicine.
You can either:
💌Forward this email
📲Share this link on your phone.
Thank you!
Until LLM Friday.
Itzel Fer, MD PM&R
Forwarded this email? Sign up here
PS. If you have suggestions, hit reply (your opinion are very important).
