Ledged Beam Walking Test Automatic Tracker: Artificial intelligence-based functional evaluation in a stroke model

Area of Science:

• Neuroscience
• Biomedical Engineering
• Artificial Intelligence

Background:

• Quantitative motor function evaluation is crucial for preclinical stroke research.
• Current methods, like manual counting in beam walking tests, suffer from subjectivity and low sensitivity.
• There is a need for objective and reproducible tools to assess motor deficits.

Purpose of the Study:

• To develop and validate an artificial intelligence-based system for objective motor function analysis.
• To improve the sensitivity and reproducibility of motor deficit assessment in experimental stroke models.
• To provide a reliable tool for preclinical evaluation of therapeutic strategies.

Main Methods:

• An artificial intelligence system utilizing a residual deep network model was developed.
• The system was trained using DeepLabCut (DLC) to extract paretic hindlimb coordinates.
• The extracted coordinates were categorized to calculate a ratio measuring neurological deficit.

Main Results:

• The AI system provided automatic, accurate, and objective analysis of motor function parameters.
• Results demonstrated higher sensitivity and greater reproducibility compared to manual assessment.
• The system's measurements showed strong correlation with those of professional observers.

Conclusions:

• The developed AI system offers a reliable and objective method for assessing motor deficits in stroke models.
• This technology can enhance the preclinical evaluation of novel therapeutic strategies.
• The tool has potential applications in other conditions characterized by motor impairments.