Association between Scale-Free Brain Dynamics and Behavioral Performance: Functional MRI Study in Resting State and Face Processing Task

Abstract

The scale-free dynamics of human brain activity, characterized by an elaborate temporal structure with scale-free properties, can be quantified using the power-law exponent (PLE) as an index. Power laws are well documented in nature in general, particularly in the brain. Some previous fMRI studies have demonstrated a lower PLE during cognitive-task-evoked activity than during resting state activity. However, PLE modulation during cognitive-task-evoked activity and its relationship with an associated behavior remain unclear. In this functional fMRI study in the resting state and face processing + control task, we investigated PLE during both the resting state and task-evoked activities, as well as its relationship with behavior measured using mean reaction time (mRT) during the task. We found that (1) face discrimination-induced BOLD signal changes in the medial prefrontal cortex (mPFC), posterior cingulate cortex (PCC), amygdala, and fusiform face area; (2) PLE significantly decreased during task-evoked activity specifically in mPFC compared with resting state activity; (3) most importantly, in mPFC, mRT significantly negatively correlated with both resting state PLE and the resting-task PLE difference. These results may lead to a better understanding of the associations between task performance parameters (e.g., mRT) and the scale-free dynamics of spontaneous and task-evoked brain activities.

Scale‐free dynamics of core‐periphery topography

“…The mid‐frequency oscillatory sine wave of 0.18 Hz caused a minor impact on the model. Finally, the high‐frequency sine wave of 0.45 Hz decreased the slope of the PLE and increased the MF, consistent with the shorter ITI of other studies (He et al, 2010 ; Kasagi et al, 2017 ; Lin et al, 2016 ). We used Pearson's correlation for the PLE (r‐values are shown in Figure 7 ).…”

“…The PLE increase in task states represents a surprising finding since previous EEG, MEG, ECoG (He et al, 2010 ; Lin et al, 2016 ), and fMRI (He et al, 2010 ; Kasagi et al, 2017 ) often observed PLE decreases in relatively fast event‐related and block task designs. Conversely, our fMRI analysis with atypically long ITI found substantial PLE increases in task states across the cerebral cortex in two core‐periphery topographies (SCP and JCP) in the primary and replication datasets.…”

“…Modulation of scale-free dynamics by the task's temporal structure Aim two uncovered wherever substantial changes of scale-free dynamics occur in task states relative to rest states. Aim three focused to what extent the task's atypically slow event rates themselves, provided by the task's infra-slow periodicity in the primary dataset (ITI = 52-60 s; 0.016-0.019 Hz) and the replication dataset (ITI = 15.5-25.5 s; 0.039-0.064 Hz), modulated the significant PLE increase from rest to task states.The PLE increase in task states represents a surprising finding since previous EEG, MEG, ECoG(He et al, 2010;Lin et al, 2016), and fMRI(He et al, 2010;Kasagi et al, 2017) often observed PLE decreases in relatively fast event-related and block task designs. Conversely, our fMRI analysis with atypically long ITI found substantial PLE increases in task states across the cerebral cortex in two coreperiphery topographies (SCP and JCP) in the primary and replication datasets.…”

“…Other studies observed a reduction of both the PLE and Hurst exponent in event‐related task designs compared to the resting‐state (Barnes et al, 2009 ; Churchill et al, 2016 ; He, 2011 ). Kasagi et al ( 2017 ) showed that the task's temporal structure or periodicity significantly impacts LRTCs including the PLE in fMRI. In contrast, it remains an open question if and to what extend the BOLD's infra‐slow frequency band aligns with the corresponding infra‐slow band of environmental stimuli or tasks in the core‐periphery topography.…”

A Copernican Approach to Brain Advancement: The Paradigm of Allostatic Orchestration

“…The potential relationship between cortical criticality and optimal anticipatory oscillation is presented in Figure 4. The theorized value of allostatic state criticality is further highlighted through depiction that the relationship between criticality and optimal anticipatory oscillation may be non-linear, given that as with other complex systems, many functions of the brain are likely to operate on the basis of power laws (e.g., Thiel, 2014; Bettinger, 2017; Kasagi et al, 2017).…”