PT - JOURNAL ARTICLE AU - Daria La Rocca AU - Nicolas Zilber AU - Patrice Abry AU - Virginie van Wassenhove AU - Philippe Ciuciu TI - Self-similarity and multifractality in human brain activity: a wavelet-based analysis of scale-free brain dynamics<sup>☆</sup> AID - 10.1101/315853 DP - 2018 Jan 01 TA - bioRxiv PG - 315853 4099 - http://biorxiv.org/content/early/2018/05/07/315853.short 4100 - http://biorxiv.org/content/early/2018/05/07/315853.full AB - Background The temporal structure of macroscopic brain activity displays both oscillatory and scale-free dynamics. While the functional relevance of neural oscillations has been largely investigated, both the nature and the role of scale-free dynamics in brain processing have been disputed.New Method Here, we offer a novel method to rigorously enrich the characterization of scale-free brain activity using a robust wavelet-based assessment of self-similarity and multifractality. For this, we analyzed human brain activity recorded with magnetoencephalography (MEG) while participants were at rest or performing a task.Results First, we report consistent infraslow (from 0.1 to 1.5 Hz) scalefree dynamics (i.e., self-similarity and multifractality) in resting-state and task data. Second, we observed a fronto-occipital gradient of self-similarity reminiscent of the known hierarchy of temporal scales from sensory to higherorder cortices; the anatomical gradient was more pronounced in task than in rest. Third, we observed a significant increase of multifractality during task as compared to rest. Additionally, the decrease in self-similarity and the increase in multifractality from rest to task were negatively correlated in regions involved in the task, suggesting a shift from structured global temporal dynamics in resting-state to locally bursty and non Gaussian scalefree structures during task.Comparison with Existing Method(s) We showed that the wavelet leader based multifractal approach extends power spectrum estimation methods in the way of characterizing finely scale-free brain dynamics.Conclusions Altogether, our approach provides novel fine-grained characterizations of scale-free dynamics in human brain activity.HighlightsWe estimated scale-free human brain dynamics using wavelet-leader formalism.High-to-low self-similarity defined a fronto-occipital gradient.The gradient was enhanced in task compared to resting-state.Scale-free brain dynamics showed multifractal properties.Self-similarity decreased whereas multifractality increased from rest to task.