Indication of Electromagnetic Field Exposure via RBF-SVM Using Time-Series Features of Zebrafish Locomotion

Area of Science:

• Neuroscience
• Biophysics
• Animal Behavior

Background:

• Electromagnetic fields (EMFs) are ubiquitous, and their biological effects require further investigation.
• Zebrafish (Danio rerio) are a valuable model organism for behavioral studies due to their sensitivity to environmental changes.
• Quantifying EMF exposure effects necessitates reliable and sensitive detection methods.

Purpose of the Study:

• To introduce a novel machine learning model for indicating electromagnetic field (EMF) exposure using zebrafish locomotion.
• To assess the efficacy of Support Vector Machine with Radial Basis Function kernel (RBF-SVM) in classifying EMF exposure levels.
• To establish a new bio-indicator model for EMF exposure based on animal behavior.

Main Methods:

• Collected time-series locomotion data from zebrafish (Danio rerio) exposed to sham or real 6.78 MHz EMF (1 A/m).
• Extracted time-series features using Highly Comparative Time-series Analysis (HCTSA) and selected relevant features with Minimum Redundancy Maximum Relevance (mRMR).
• Trained and validated a RBF-SVM classification model to differentiate between EMF exposure conditions.

Main Results:

• The developed RBF-SVM model achieved 100% classification accuracy in indicating different EMF exposures.
• Classification precision exceeded 95% for several classifiers based on specific feature sets.
• Ambient environmental parameters showed minimal impact on zebrafish locomotion, as verified by regression analysis.

Conclusions:

• The proposed model effectively indicates EMF exposure levels through zebrafish locomotion analysis.
• The findings suggest that specific EMFs may influence zebrafish neural activity, manifesting in behavioral changes.
• This study provides a novel bio-indicator for EMF exposure and a reference for future research on EMF impacts.