Leveraging AI technology for distinguishing Eucommiae Cortex processing levels and evaluating anti-fatigue potential

Area of Science:

• Pharmacognosy
• Computational Biology
• Traditional Chinese Medicine

Background:

• Eucommiae Cortex (ECO) is a valued traditional Chinese medicinal plant with significant health benefits.
• Processing techniques critically influence ECO's efficacy and consistency.
• Current identification methods for ECO are inefficient and invasive.

Purpose of the Study:

• To develop an accurate, rapid, and non-invasive method for classifying ECO based on processing levels.
• To evaluate the anti-fatigue properties of ECO at different processing stages.
• To correlate image-based classification with functional anti-fatigue efficacy.

Main Methods:

• Deep learning models, specifically ResNet and Vision Transformer (ViT), were utilized for image classification of ECO.
• Anti-fatigue effects were assessed via swimming endurance, pole climbing tests, and biochemical markers (SDH, LDH, ATP, Na+-K+-ATPase, Ca2+-Mg2+-ATPase).
• Image analysis was employed to automatically grade ECO based on processing variations.

Main Results:

• The Vision Transformer model achieved over 95% accuracy in classifying ECO images.
• Image analysis proved effective for automated and accurate ECO classification.
• Mice treated with moderately processed ECO exhibited superior fatigue resistance compared to other groups.

Conclusions:

• Deep learning, particularly ViT, offers a highly accurate and non-invasive approach for quality control of processed ECO.
• Processing methods significantly impact the anti-fatigue efficacy of Eucommiae Cortex.
• Optimizing ECO processing is crucial for maximizing its therapeutic benefits.