Development and evaluation of the piecewise Prony method for evoked potential analysis.

Area of Science:

• Signal Processing
• Neuroscience
• Biomedical Engineering

Background:

• Classical Prony method is limited to analyzing signals with fixed or monotonic parameters.
• Nonstationary signals, common in biological systems, require advanced analysis techniques.
• Understanding temporal dynamics of evoked potentials (EPs) is crucial for neuroscience.

Purpose of the Study:

• To introduce a novel piecewise Prony method (PPM) for decomposing nonstationary signals.
• To enable the study of time-varying frequency, amplitude, and phase in oscillatory components.
• To investigate post-stimulus signal changes in single-trial evoked potentials (EPs).

Main Methods:

• Developed the piecewise Prony method (PPM) to handle time-varying signal characteristics.
• Evaluated PPM using simulated data to determine its detection limits for amplitude, rate, and frequency changes.
• Applied PPM to analyze single-trial electroencephalogram (EEG)/EP data from flash visual evoked potentials.

Main Results:

• PPM successfully detected small changes: 10% in amplitude, 10% in rate, and 0.15 Hz in frequency in simulations.
• Analysis of flash visual EPs showed significant attenuation of alpha activity post-stimulus, with recovery around 0.5s.
• PPM provided evidence of phase alignment in delta activity following stimulus presentation in EEG trials.

Conclusions:

• The piecewise Prony method (PPM) is a capable tool for analyzing nonstationary signals with time-varying parameters.
• PPM offers new insights into the dynamic generation of evoked potentials (EPs) by analyzing single trials.
• The method demonstrates potential for advancing the understanding of neural signal processing and responses.