Multifractal spin-glass chaos projection and interrelation of multicultural music and brain signals

E. Can Artun, Ibrahim Keçoğlu, Alpar Türkoğlu, A. Nihat Berker

TL;DR

This paper introduces a novel complexity classification scheme based on the fractal spectra of spin-glass chaos. By applying this framework to multicultural music and brain EEG signals, the authors reveal surprising systematic patterns and interrelations, suggesting a universal language of complexity across seemingly disparate domains.

What Problem Does This Paper Address?

Understanding and classifying complex systems remains a significant challenge. While diverse systems like spin glasses, music, and brain activity exhibit intricate behaviors, finding common ground and robust metrics for comparison is difficult. The authors propose that the "chaos under scale change" characteristic of spin-glass ordering, quantified by multifractal spectra, can serve as a universal benchmark to classify and connect these complex systems. This approach aims to bridge the gap between abstract theoretical models and real-world data, offering a new lens to analyze cultural trends, geographical influences, and even physiological states.

Methods

The core of the paper lies in developing and applying a complexity classification scheme derived from spin-glass chaos. The authors leverage the exact renormalization-group solution of frustrated Ising models on hierarchical lattices. This process generates chaotic renormalization-group trajectories, which are characteristic of the spin-glass phase. The multifractal spectra of these chaotic trajectories are then calculated.

The key innovation is using these spin-glass multifractal spectra as a "standard metric." For other complex systems (music and EEG), their own multifractal spectra are computed. The authors then fit the experimental multifractal spectrum points to a continuous quartic function derived from the spin-glass model. This fitting process involves optimizing parameters of the spin-glass model (frustration number $p_b$ and repressed strand length $m_1$) to minimize the root-mean-square separation between the experimental and spin-glass spectra. The goodness of this fit is quantified by correlation coefficients ($R_s$ and $R_m$ or $R_e$), and the strength of the fitting chaos is measured by the Lyapunov exponent ($\lambda$).

Datasets

The paper utilizes two primary types of datasets:

  • Multicultural Music: 80 music tracks from 8 genres (Turkish: Arabesque, Rap, Pop, Classical; Western: Blues, Jazz, Pop, Classical). These tracks were converted into time series data for analysis.
  • Brain EEG Signals: Electroencephalogram (EEG) data collected from participants in different states (music listening, resting with eyes open, resting with eyes closed) and at various cranular probe locations.

The paper does not introduce new benchmark datasets but rather applies its novel classification scheme to existing types of complex data.

Key Insights and Ideas

  • Spin-Glass Chaos as a Universal Metric: The central idea is that the "chaos under scale change" inherent in spin-glass ordering can provide a universal quantitative measure for complexity across diverse systems.
  • Interconnectedness of Seemingly Disparate Systems: The study reveals that music and brain signals exhibit patterns that can be mapped onto the spin-glass chaos framework, suggesting a deeper underlying connection in their complexity.
  • Surprising Group Defection in Music: A notable finding is the emergence of distinct groupings within the music data. Western classical, blues, and jazz music tend to stand apart, while Turkish music genres (Arabesque, Rap, Pop, Classical) cluster together, surprisingly including Western Pop.
  • Clearer Patterns in Brain Signals: EEG data shows even more pronounced systematic patterns, with distinct streaks and near-orthogonal branches observed in different cranular regions and resting states.
  • Sensitivity to Brain States: The EEG analysis highlights differences between resting states (eyes open vs. closed) and music listening, with music listening and resting with eyes closed showing significant overlap.

Key Results

  • The spin-glass multifractal spectrum fitting procedure successfully maps music and EEG data onto the spin-glass chaos framework, with high correlation coefficients ($R_s$, $R_m$, $R_e$) and calculated Lyapunov exponents ($\lambda$) indicating the strength of the fitting chaos.
  • Figure 3 shows examples of excellent fits for both music and EEG data, with high $R$ values and non-zero $\lambda$.
  • Figure 4 (music) and Figure 5 (EEG) visually demonstrate the systematic patterns identified, with distinct streaks and clusters emerging based on genre, origin, and brain state.
  • The paper quantifies the "goodness of fit" between the experimental data and the spin-glass model, allowing for a comparative analysis of complexity across different data types.

Limitations and Future Work

  • Data Representation: The paper mentions converting audio files to time series and selecting the left-ear channel for music, and using specific filtering and data selection for EEG. The impact of these choices on the results could be further explored.
  • Generalizability: While the scheme is presented as widely usable, its application to a broader range of complex systems (e.g., financial markets, biological networks) would further validate its universality.
  • Interpretation of Parameters: The physical meaning of the optimized spin-glass parameters ($p_b$, $m_1$) in relation to the specific characteristics of music and brain signals could be explored in more depth.
  • Future Work: The authors suggest that this classification scheme can be applied to "any complex system data set," implying a broad scope for future research. Further investigation into the specific mechanisms underlying the observed interrelations between music, brain activity, and spin-glass chaos is also a natural next step.

Summary updated

Abstract

A complexity classification scheme is developed from the fractal spectra of spin-glass chaos and demonstrated with multigeographic multicultural music and brain electroencephalogram signals. Systematic patterns are found to emerge. Chaos under scale change is the essence of spin-glass ordering and can be obtained, continuously tailor-made, from the exact renormalization-group solution of Ising models on frustrated hierarchical lattices. The music pieces are from Turkish music, namely Arabesque, Rap, Pop, Classical, and Western music, namely Blues, Jazz, Pop, Classical. A surprising group defection occurs.
BibTeX
@inproceedings{artun2022multifractal,
  author = {E. Can Artun and Ibrahim Keçoğlu and Alpar Türkoğlu and A. Nihat Berker},
  title = {{Multifractal spin-glass chaos projection and interrelation of multicultural music and brain signals}},
  booktitle = {Chaos Solitons \& Fractals},
  year = {2022},
  doi = {10.1016/j.chaos.2022.113005},
}

Loading related content...

Figure