Impact of functional electrical stimulation on nerve-damaged muscles by quantifying fat infiltration using deep...

Area of Science:

• Life Sciences
• Biomedical Engineering
• Histology

Background:

• Quantitative imaging and automated analysis are crucial for life science research.
• Functional electrical stimulation (FES) is explored to aid respiration in patients with bilateral vocal fold paresis.
• Previous studies found no atrophy or fibrosis in posterior cricoarytenoid muscle (PCA) due to FES.

Purpose of the Study:

• To evaluate the impact of long-term FES on intramuscular fat in the PCA using quantitative imaging.
• To test the hypothesis that FES does not significantly alter fat content in the PCA.
• To demonstrate the efficacy of automated deep learning for analyzing histological data.

Main Methods:

• Quantitative image analysis of H&E stained PCA muscle cross-sections from 22 sheep.
• Application of the deep neural network IMFSegNet for automated intramuscular fat segmentation.
• Comparison of fat content between FES-treated and control groups, and between different FES duty cycles.

Main Results:

• No significant differences in intramuscular fat were observed in the PCA between sheep with and without long-term FES.
• No significant differences in intramuscular fat were found between low and high duty cycle FES groups.
• IMFSegNet provided accurate and efficient analysis compared to manual methods.

Conclusions:

• Long-term FES, under the tested parameters, does not negatively impact the PCA in a human-like animal model.
• Automated deep learning-based quantitative imaging is a powerful and objective tool for complex histological analyses.
• FES shows potential as a therapeutic intervention for conditions like bilateral vocal fold paresis without causing muscle damage.