Tracking Non-Stationary Spectral Peaks in EEG Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing EEG data fail to accurately track non-stationary spectral peaks, which are essential for monitoring and controlling physiological states such as sedation, anesthesia, and sleep, as they collapse broadband physiological oscillations into single frequencies, losing important information.
Innovation Solution
A system and method that utilize statistical sampling to decompose time-varying spectra into multiple concurrent spectral peaks using parametric or semi-parametric models, allowing for the estimation and tracking of instantaneous peak frequency, amplitude, and bandwidth, thereby retaining the full spectral structure and information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If sinusoidal models are used to track spectral content, then the tracking of single frequency components is simplified, but broadband physiological oscillations are collapsed to a single frequency, losing information content in the peak bandwidth and structure
Solution Approach 1:
The patent segments the broadband spectral peak into multiple frequency components by fitting a sum of sinusoids to the spectral data. Each sinusoid represents a specific frequency component within the broadband peak, allowing the system to track individual frequency trajectories while preserving the overall spectral structure. This segmentation approach transforms the single-frequency limitation into a multi-frequency analysis capability.
Solution Approach 2:
The patent extends the analysis from a single frequency dimension to a time-frequency dimension by tracking the evolution of multiple sinusoidal components over time. The spectral peaks are represented as time-varying functions with parameters including instantaneous frequency, amplitude, and bandwidth, effectively adding temporal dynamics to the spectral analysis and recovering information that would be lost in static single-frequency models.
2Device complexity
If discrete time period analysis is used to characterize neural rhythms, then computational complexity is reduced, but the non-stationary nature of physiological signals is not captured, reducing measurement precision
Solution Approach 1:
The patent implements dynamic tracking of spectral peaks by allowing the sinusoidal parameters (frequency, amplitude, phase) to vary continuously over time rather than being fixed for discrete time periods. The model uses time-varying parameters that adapt to the non-stationary nature of physiological signals, enabling precise tracking of spectral evolution while maintaining computational efficiency through parametric modeling.
Solution Approach 2:
The patent changes the parameters of the sinusoidal model from fixed values to time-varying functions. The instantaneous frequency, amplitude, and bandwidth are allowed to change continuously, capturing the non-stationary characteristics of physiological signals. This parameter adaptation enables the model to follow spectral peaks as they evolve over time without requiring complex computational methods.
3Ease of operation
If broadband physiological oscillations are collapsed to single frequencies, then the analysis becomes simpler, but the full spectral structure and important information are lost
Solution Approach 1:
The patent segments the broadband oscillation into multiple sinusoidal components, each representing a specific frequency band within the overall spectral peak. By fitting a sum of sinusoids with different frequencies, amplitudes, and phases, the system preserves the spectral structure while maintaining analytical simplicity through parametric representation.
Solution Approach 2:
The patent uses a composite model consisting of multiple sinusoidal components combined in a sum to represent the broadband physiological oscillation. Each sinusoid acts as a building block with specific parameters, and their superposition creates a composite signal that captures the full spectral structure. This composite approach maintains simplicity while preserving information that would be lost in single-frequency models.
Data Source
AI summary
Systems and methods for tracking dynamic structure in physiological data are provided. In some aspects, the method includes providing physiological data, including electroencephalogram (“EEG”) data, acquired from a subject and assembling a time-frequency representation of signals from the physiological data. The method also includes generating a dynamic model of at least one non-stationary spectral feature, such as at least one non-stationary spectral peak, using the time-frequency representation and a user indication, and applying a dynamic model of at least one non-stationary spectral feature in a parameter estimation algorithm to compute concurrent estimates of spectral parameters describing the at least one non-stationary spectral feature. The method also includes tracking the spectral parameters of the at least one spectral feature over time.