Related Experiment Video
Updated: Jan 7, 2026
06:11
Brain-Computer Interface-controlled Upper Limb Robotic System for Enhancing Daily Activities in Stroke Patients
Published on: April 18, 2025
1.5K
Artificial Intelligence in Managing Spasticity with Botulinum Toxin Type A-Insights from an Exploratory Pilot
Mirko Filippetti1,2, Rita Di Censo1,3, Lyria Arcari1
1Section of Physical and Rehabilitation Medicine, Department of Neurosciences, Biomedicine and Movement Sciences, University of Verona, 37134 Verona, Italy.
Toxins
|December 24, 2025
Summary
Artificial intelligence (AI) shows promise for optimizing botulinum toxin type A treatments for spasticity. However, current AI models struggle with complex cases, highlighting the need for enhanced clinical data integration.
Area of Science:
- Neurology
- Medical Informatics
- Rehabilitation Medicine
Background:
- Spasticity management often involves complex treatment decisions.
- Botulinum toxin type A is a key therapeutic agent.
- Optimizing treatment requires integrating clinical expertise and evidence-based guidelines.
Purpose of the Study:
- To evaluate artificial intelligence (AI) for optimizing botulinum toxin type A treatment in spasticity.
- To compare AI-generated recommendations with expert clinical decisions.
- To identify AI's potential and limitations in spasticity management.
Main Methods:
- Comparative analysis of AI recommendations versus physician decisions.
- Utilized thirty hypothetical clinical cases of spasticity.
- Involved five experienced rehabilitation physicians and an AI model trained on literature and guidelines.
Main Results:
- AI demonstrated consistency and guideline adherence but lacked adaptability in complex cases.
- AI generally recommended lower dosages than physicians.
- Significant differences observed in muscle selection and treatment strategies between AI and physicians.
Conclusions:
- AI shows potential as a clinical support tool for standardized spasticity management.
- Current AI limitations in interpreting clinical subtleties restrict practical application.
- Future AI models require multimodal data and clinician feedback for improved decision-making emulation.
