Fractal complexity of spontaneous brain activity and the effect of scanning parameters

• Published in: Advances in continuous and discrete models Vol. 2025; no. 1; p. 42
• Main Authors: Guan, Sihai, Meng, Chun, Bharat, Biswal
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.12.2025; Springer Nature B.V
• Subjects: Analysis; Biomarkers; Biomathematical Modelling and Stochastic Analysis; Brain research; Correlation coefficients; Difference and Functional Equations; Dynamic characteristics; Fractals; Functional Analysis; Magnetic resonance imaging; Mathematics; Mathematics and Statistics; Neurosciences; Nonlinear dynamics; Ordinary Differential Equations; Parameters; Partial Differential Equations; Time series; Wavelet transforms; Multi-sessions; ICC; Different TRs; Rs-fMRI; Multifractal
• ISSN: 1687-1839; 2731-4235; 1687-1847
• DOI: 10.1186/s13662-025-03910-2

The relationship between the multifractal of the resting-state functional magnetic resonance imaging (rs-fMRI) and its parameters still needs to be discovered. So, the primary purpose of this study is to evaluate the effects of repetition time (TR) on rs-fMRI signals, and the relationship between multifractal and session. Our findings revealed that the larger the TR, the more scattered the analysis results of multifractal, and the multifractal increase with the TR between Youths and Adults except for the degree of multifractal. However, it is worth noting that when TR = 1.4 s, the average degree of multifractal is generally greater than TR = 0.645 s and TR = 2.5 s. Furthermore, the reliability decreases with the increase in session numbers, increases with rs-fMRI data length increase, and is higher for the short intervals than long ones. Besides, with the rise in the number of sessions, when the data length is ≥400, the intraclass correlation coefficient (ICC) > 0.5 of the Hurst exponent (HE); when the data length is equal to 808, the ICC > 0.4 of the maximum gradient, and the spectrum width of ICC < 0.3. With the increase of the session interval, when the data length is ≥200, the ICC > 0.4 of the HE; when the rs-fMRI data length is equal to 808, the ICC ≈ 0.4 of the maximum gradient, and the spectrum width and degree of multifractal of ICC < 0.3. The results are of great significance for understanding the relationship between nonlinear dynamic characteristics and parameters of rs-fMRI.