Abstract
Fluctuation-based dispersion entropy (FDispEn) is a new approach to estimate the dynamical variability of the fluctuations of signals. It is based on Shannon entropy and fluctuation-based dispersion patterns. To quantify the physiological dynamics over multiple time scales, multiscale FDispEn (MFDE) is developed in this article. MFDE is robust to the presence of baseline wanders, or trends, in the data. We evaluate MFDE, compared with popular multiscale sample entropy (MSE), and the recently introduced multiscale dispersion entropy (MDE), on selected synthetic data and five neurological diseases' datasets: 1) focal and non-focal electroencephalograms (EEGs); 2) walking stride interval signals for young, elderly, and Parkinson's subjects; 3) stride interval fluctuations for Huntington's disease and amyotrophic lateral sclerosis; 4) EEGs for controls and Alzheimer's disease patients; and 5) eye movement data for Parkinson's disease and ataxia. MFDE dealt with the problem of undefined MSE values and, compared with MDE, led to more stable entropy values over the scale factors for pink noise. Overall, MFDE was the fastest and most consistent method for the discrimination of different states of neurological data, especially where the mean value of a time series considerably changes along the signal (e.g., eye movement data). This study shows that MFDE is a relevant new metric to gain further insights into the dynamics of neurological diseases recordings.
Introduction
In signal processing, entropy measures the irregularity or unpredictability of a signal. However, biological systems—such as cardiac rhythms or neural activity—are inherently complex, exhibiting meaningful patterns across multiple temporal scales. Traditional measures like Multiscale Sample Entropy (MSE) often struggle with three primary issues: high computational cost, failure to return values for short data segments (becoming "undefined"), and extreme sensitivity to "baseline wander" or slow shifts in the signal's mean caused by movement or sensor drift.
Methodology: Multiscale Fluctuation-based Dispersion Entropy (MFDE)
To address these limitations, researchers developed Multiscale Fluctuation-based Dispersion Entropy (MFDE). The algorithm operates in two primary stages:
- Coarse-Graining: The original signal is divided into segments, which are averaged to create "zoomed-out" versions of the data across various scales.
- Fluctuation-based Dispersion Entropy (FDispEn): Unlike standard methods that look at absolute values, FDispEn analyzes the differences (fluctuations) between adjacent points. These fluctuations are mapped into discrete classes using a Normal Cumulative Distribution Function (NCDF), and Shannon entropy is calculated based on the frequency of resulting patterns.
By focusing on fluctuations rather than absolute amplitudes, the algorithm naturally filters out slow-moving trends or shifts in the signal's mean, making it exceptionally robust for real-world clinical data.
Experimental Validation and Clinical Results
The authors validated MFDE using synthetic noise and five diverse clinical datasets, ranging from EEG recordings to gait analysis.
Computational Efficiency and Stability
MFDE demonstrates superior speed, performing roughly 150 times faster than MSE on large datasets (100,000 samples). Furthermore, while MSE often becomes undefined for signals as short as 100 samples, MFDE remains stable and provides discriminative values, making it suitable for short-term monitoring.
Robustness to Baseline Wander
A critical advantage of MFDE is its performance on "messy" data. In oculomotor (eye movement) recordings, the mean value of the signal shifts constantly as the subject's gaze moves. MFDE was the only method capable of consistently discriminating between Parkinson’s Disease and Ataxia across all scales in these recordings.
Clinical Insights
- Alzheimer’s Disease: MFDE revealed that patients exhibit lower brain complexity at short time scales but higher irregularity at long time scales compared to healthy controls.
- Neurodegenerative Gait: The method successfully distinguished between Amyotrophic Lateral Sclerosis (ALS) and Huntington’s Disease gait patterns, where other modern methods like Multiscale Dispersion Entropy (MDE) failed.
Discussion and Future Directions
MFDE represents a significant advancement in quantifying physiological complexity. Its linear computational growth ($O(N)$) and robustness to trends make it a prime candidate for real-time diagnostic tools. Future work will focus on expanding the method to multivariate recordings to analyze cross-complexity between different brain regions and exploring advanced filtering techniques for the coarse-graining process.
Summary updated