Hello!
Welcome back to AIMedily.
This week I’m in Colorado, and about to go off-grid to do some backcountry camping.
I won’t have internet access for a few days.
That means I won’t know if you read (and enjoyed) my email until I’m back from the mountains. I hope you like it!
There are 2 things I want to talk about today:
1️⃣ I’m sure you notice that I add links in blue to the original articles I summarize (titles and references).
I also add links to other things like:
Medical assesstments
Equipment used (motion analysis, apps, software, etc)
Medical or Technical terms
Feel free to click on those links to learn more about that specific topic.
2️⃣ In the 🦾Techtools section, I’ll be sharing an AI app or software that could be useful for you (let me know if you enjoy this!).
Now, let’s take a look at what you’ll find in today’s issue:
.🤖AIBytes: An app that accelerates hand rehabilitation, a lightweight prosthetic hand that follows voice commands and Smartwear that tracks muscle activation with textile sensors.
🦾TechTools: A free AI medical scribe and an AI secretary for meetings.
🧬AIMedily Snaps: Can we trust AI in medicine? Medical digital twins, Health Tech in 2040, Machine Learning in cerebral palsy, and Augmented and Virtual Reality in disability.
🧩TriviaRX: Find the new question and the answer to our last triviaRX.
🤖 AIBytes
🔬 Methods
Study design: Prospective randomized controlled trial
Participants: 112 postoperative patients (ages 18–65, mean 36.9 ± 15.5 years) with either:
Metacarpal/finger fractures
Flexor/extensor tendon injuries
Intervention:
The Novio Hand smartphone app: Uses AI and the phone camera for motion analysis:
Tracks finger movements and measures range of motion (ROM) in real time.
AI adjusts exercise type and intensity based on the patient’s performance and symptoms.
Novio Therapy app at home + standard outpatient hand therapy (18 sessions).
Assessments:
ROM was recorded at baseline and at 2, 6, and 12 weeks.
Achievement of the minimal clinically important difference defined as ≥ 40° ROM improvement.
📊 Results
At 2 and 6 weeks:
MCID (≥40° improvement):
2 weeks: Combined group 50% vs Control group 26% reached MCID (Minimally clinical important difference)
6 weeks: Combined group 79% vs Control group 54%
At 6 weeks, combined therapy group patients with fractures achieved near full ROM on average.
🔑 Key Takeaways
Novio Hand AI app + Standard Therapy significantly accelerates early postoperative ROM recovery.
The app serves as a virtual physiotherapist, enabling adaptive home exercises.
A higher proportion of patients hit clinically meaningful recovery in 2 to 6 weeks.
🔗 Bauknecht S, Moeller R, Mentzel M, et al. Physiotherapy in Your Pocket: Effectiveness of Home Exercises Using an AI-Based Smartphone App for the Postoperative Follow-Up of Hand Injuries. Handchir Mikrochir Plast Chir. 2025;57(3):186–194. doi:10.1055/a-2599-8250.
This hand prosthetic control mechanism is a closed-loop system where voice commands are translated to the SMA actuating system.
It has the structural and kinematic characteristics of human hands.
🔬 Methods
The prothesis has 9 active joints (with micro motors) and 10 passive joints (fingers and wrist).
Voice recognition tech: iFliteck is a voice recognition module that interfaces with an LLM. It responds to 60 different languages 🗣️.
Allows precise, independent movements across all fingers.
3D printed.
Weights only 370 grams.
Has 2.5 kg maximum load capacity.
Participant: 1 transradial amputee.
Assesments:
33 standard grasping modes (Feix Grasp Taxonomy) + 6 complex grasp.
Activities of Daily Living: Combing hair, shaking hands, and writing with a pen.
📊 Results
The ability to control finger and wrist movements through voice commands allows users to improve functionality.
Movements were precise and responsive (91% success rate at following the verbal instruction).
Patient could perform daily activities with high accuracy.
🔑 Key Takeaways
This prosthetic hand approaches natural hand movement while remaining lightweight and low-cost.
Future research should prioritize long-term user trials in a wider patient demographic and usesfulness in real-world scenarios.
🔗 Yang H, Tao Z, Yang J, et al. A lightweight prosthetic hand with 19-DOF dexterity and human-level functions. Nat Commun. 2025;16:955. doi:10.1038/s41467-025-56352-5ht
A wearable, AI-integrated smart clothing system using ultrasensitive textile-based strain sensors to track muscle activation and breathing patterns.
🔬 Methods
Validation study.
Participants: 8 healthy subjects (5 males and 3 females)performing bodyweight exercises (bench press)
Intervention:
Compression clothing with ultrasensitive strain sensors at the abdomen and left and right chest muscle.
Data was collected via WiFi. It was then streamed to a computer or mobile device connected to a Printer Circuit Board.
The data was segmented and classified through a Convolutional Neural Network (CNN) as correct or incorrect.
The Transformer-based model for high-fidelity pose estimation.
Environment: Indoor controlled lab.
📊 Results
The system classified movements with 92.1% accuracy.
The model achieved precise pose reconstruction using only the sensor data.
Low latency and high fidelity confirmed the potential for real-time feedback.
🔑 Key Takeaways
The system is accurate at detecting breathing irregularities and asymmetric muscle exertion.
AI-integrated textile sensors enable accurate, real-time monitoring.
No cameras - all data collected from fabric strain sensors.
Foundation for next-generation AI-powered clothing for rehabilitation, home exercise monitoring, and sports injury prevention.
🔗 Tang C, Yi W, Zhang Z, Occhipinti E, Occhipinti LG. AI-Driven Smart Sportswear for Real-Time Fitness Monitoring Using Textile Strain Sensors. arXiv, April 2025. doi:10.48550/arXiv.2504.08500
🦾TechTool
Heidi Health: A free AI medical scribe. Heidi Health transcribes, writes notes and gives you summaries while you give medical consultation.
Otter: Your private AI secretary for meetings with your team. You don’t need someone taking notes, Otter will do it for you.
🧬AIMedily Snaps
🧬Can we trust AI in Medicine? Medical AI can transform medicine - but only if we carefully track data.
🧬A paper on medical digital twins.
🧬Transformative Personal Health Tech in 2040. The Politecnico di Milano identify disruptive tech in personal health.
🧬Can Machine Learning tell us if children with Cerebral Palsy will walk independently?
🧬A paper on Augmented Reality and Virtual Reality for people with disabilities.
🧩TriviaRX
Which device pioneered motion capture in early physiotherapy research in the 1970s?
A) Infrared camera arrays
B) Force plates
C) Selspot system
D) Inertial sensors
(The answer will be in the next issue)
👉Now, the answer from last Trivia!
✅ B – World War I . It led to a dramatic increase in amputations and birth of modern PM&R. 📚 Ref: Krusen FH, Phys Med Rehabil. 1947.
That’s all for today! Thank you for taking the time to read.
Please help me spread the word about AIMedily.
💌You can forward this email to your friends
📲Share this link on your phone.
Thank you!
See you next week,
Itzel Fer
MD PM&R
Forwarded this email? Sign up here
