Power-law scaling of brain wave activity associated with mental fatigue

Vo Anh, Hung T. Nguyen, Ashley Craig, Yvonne Tran, Yu Guang WangbioRxiv 2020 · DOI

TL;DR

This paper introduces a novel mathematical model for brain activity using a stochastic delay-differential equation (SDDE) to explain power-law scaling in EEG signals. They demonstrate that this model, incorporating brain cortex heterogeneity and initial state (alert vs. fatigue), can effectively capture and differentiate between these states, offering a new perspective on analyzing brain dynamics.

What Problem Does This Paper Address?

Mental fatigue, particularly in tasks like driving, is a significant safety concern. While electroencephalography (EEG) is widely used to study brain activity changes associated with fatigue, the underlying mechanisms and how to robustly detect these states remain areas of active research. Specifically, the paper addresses the phenomenon of "power-law scaling" observed in EEG signals, which suggests long-range temporal dependence and scale invariance, characteristic of complex systems near critical points. The authors aim to understand the cause of this scaling, linking it to the brain's inherent heterogeneity and the initial state of alertness or fatigue, and to develop a model that can quantify and differentiate these states.

Methods

The core of the paper is the development and analysis of a stochastic delay-differential equation (SDDE) to model the global brain wave activity as a random field on the cortical surface (modeled as a unit sphere S²). Key components of their model include:

  • Fractional Diffusion Operator: This operator, denoted by $\Psi(-\Delta_{S^2})$, captures the effects of the heterogeneous brain cortex on signal diffusion. It's a generalization of standard diffusion and is designed to inherently produce power-law scaling.
  • Delay Operator: This component accounts for the influence of the initial condition, specifically the brain's alert or fatigue state, on the ongoing activity.
  • Karhunen-Loève Expansion: The authors derive an analytical solution to their SDDE using this expansion, which is a standard technique for representing random fields. This allows for a closed-form expression of the covariance function.
  • Parameter Estimation: They develop numerical schemes to estimate the key parameters of their model (fractional diffusion exponents $\alpha, \gamma$ and the delay parameter $\tau$) from real EEG data.

Datasets

The study utilizes real EEG data from 50 participants performing a driving fatigue task. The data consists of 32 channels of EEG measurements, recorded for 20 seconds each, with a sampling rate of 5120 time points per second. This dataset is used to estimate the model parameters and validate the findings.

Key Insights and Ideas

  • Brain Heterogeneity as a Source of Power-Law Scaling: The paper proposes that the inherent structural and functional heterogeneity of the brain cortex is a primary driver of the observed power-law scaling in EEG signals.
  • Dual Role of Model Parameters: The fractional diffusion parameters ($\alpha, \gamma$) are linked to the medium's heterogeneity, while the delay parameter ($\tau$) captures the influence of the initial state (alert/fatigue). Crucially, the same exponent ($\alpha$) governs both the power-law scaling and the asymptotic temporal correlation.
  • Non-Gaussianity Differentiates States: The diffusion process is found to be non-Gaussian, and this non-Gaussianity is more pronounced in the alert state than in the fatigue state. This suggests that the brain's response to stimuli is more complex and less predictable when alert.
  • Initial Condition's Impact Varies with State: The influence of the initial brain state (captured by the delay parameter) is stronger in the alert state compared to the fatigue state. This implies that the brain's transition to fatigue might involve a dampening of the initial state's influence.
  • Global Model Complementarity: The study highlights the value of a global SDDE model for understanding brain activity, complementing traditional time-series analysis of individual EEG channels.

Key Results

  • EEG time series exhibit power-law scaling (long memory) in the low-frequency range.
  • The extent of this long memory is significantly reduced in the fatigue state compared to the alert state.
  • The EEG random field diffusion is non-Gaussian, with this non-Gaussianity being more pronounced in the alert state.
  • The system's response to the initial condition is stronger in the alert state than in the fatigue state.
  • Parameter estimation from real EEG data showed that the averaged value of $\alpha$ is higher in the fatigue state (1.116) than in the alert state (0.921), indicating a change in the nature of brain activity.
  • The delay parameter $\tau$ was found to be larger in the alert state (1.676) than in the fatigue state (0.183), suggesting a longer memory or stronger influence of the initial state when alert.

Limitations and Future Work

  • Approximation of Model: The paper uses approximations for certain operators and calculations, particularly when dealing with large $l$ values in the Fourier coefficients.
  • Interpretation of Parameters: While the authors provide interpretations for $\alpha, \gamma,$ and $\tau$, the precise physiological meaning of some combinations (e.g., $\alpha + \gamma$) and the exponent $\theta$ (related to initial condition variances) could be further explored.
  • Generalizability: The study focuses on driver fatigue. Future work could investigate whether these findings generalize to other types of mental fatigue or cognitive states.
  • Model Complexity: The SDDE model is complex. Future research could explore simpler models that capture similar phenomena or investigate the computational efficiency of the current model.
  • Further Validation: While validated on EEG data, further validation with other neuroimaging techniques or different experimental paradigms could strengthen the findings.

Summary updated

Abstract

Abstract This paper investigates the cause and detection of power-law scaling of brain wave activity due to the heterogeneity of the brain cortex, considered as a complex system, and the initial condition such as the alert or fatigue state of the brain. Our starting point is the construction of a mathematical model of global brain wave activity based on EEG measurements on the cortical surface. The model takes the form of a stochastic delay-differential equation (SDDE). Its fractional diffusion operator and delay operator capture the responses due to the heterogeneous medium and the initial condition. The analytical solution of the model is obtained in the form of a Karhunen-Loève expansion. A method to estimate the key parameters of the model and the corresponding numerical schemes are given. Real EEG data on driver fatigue at 32 channels measured on 50 participants are used to estimate these parameters. Interpretation of the results is given by comparing and contrasting the alert and fatigue states of the brain. The EEG time series at each electrode on the scalp display power-law scaling, as indicated by their spectral slopes in the low-frequency range. The diffusion of the EEG random field is non-Gaussian, reflecting the heterogeneity of the brain cortex. This non-Gaussianity is more pronounced for the alert state than the fatigue state. The response of the system to the initial condition is also more significant for the alert state than the fatigue state. These results demonstrate the usefulness of global SDDE modelling complementing the time series approach for EEG analysis.
BibTeX
@misc{anh2020power,
  author = {Vo Anh and Hung T. Nguyen and Ashley Craig and Yvonne Tran and Yu Guang Wang},
  title = {{Power-law scaling of brain wave activity associated with mental fatigue}},
  year = {2020},
  doi = {10.1101/2020.08.03.234120},
  howpublished = {bioRxiv},
}

Loading related content...

Figure