Characterisation of Physiological Tremor using Multivariate Empirical Mode Decomposition and Hilbert Transform.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Human Factors Engineering

Background:

• Fatigue-induced physiological tremor (FIPT) negatively impacts precision tasks like micromanipulation.
• Traditional methods (sEMG, MMG with Fourier transforms) assume signal linearity and stationarity, limiting their analysis of tremor.
• Developing robust methods to characterize FIPT is crucial for improving performance in high-precision activities.

Purpose of the Study:

• To characterize fatigue-induced physiological tremor (FIPT) using advanced signal processing techniques.
• To introduce a novel indicator, Energy Ratio (ER), for quantifying FIPT.
• To assess the efficacy of MEMD and Hilbert spectral analysis for tremor characterization.

Main Methods:

• Utilized multivariate empirical mode decomposition (MEMD) to extract relevant frequency bands from physiological tremor signals.
• Applied Hilbert spectral analysis to estimate signal energy within the extracted bands.
• Proposed and calculated the Energy Ratio (ER) as a parameter to quantify FIPT.

Main Results:

• The Energy Ratio (ER) demonstrated a significant increasing trend with task epoch, indicating rising fatigue levels.
• Linear regression showed high correlation coefficients (R²≈0.7 for sEMG, R²≈0.9 for accelerometer data) for ER predicting fatigue.
• MEMD and Hilbert spectral analysis proved effective for analyzing non-linear, non-stationary tremor signals.

Conclusions:

• The proposed Energy Ratio (ER) is an effective and versatile indicator for quantifying fatigue-induced physiological tremor.
• This characterization method can be integrated into control strategies to mitigate tremor effects in prolonged micromanipulation, such as surgery.
• The findings support the use of MEMD and Hilbert spectral analysis for analyzing physiological tremor and have implications for surgical training and performance enhancement.