Background: Tai Chi has been shown to improve motor symptoms in Parkinson’s disease (PD), but its long-term effects and the related mechanisms remain to be elucidated. In this study, we investigated the effects of long-term Tai Chi training on motor symptoms in PD and the underlying mechanisms. Methods: Ninety-five early-stage PD patients were enrolled and randomly divided into Tai Chi (n = 32), brisk walking (n = 31) and no-exercise (n = 32) groups. At baseline, 6 months and 12 months during one-year intervention, all participants underwent motor symptom evaluation by Berg balance scale (BBS), Unified PD rating-scale (UPDRS), Timed Up and Go test (TUG) and 3D gait analysis, functional magnetic resonance imaging (fMRI), plasma cytokine and metabolomics analysis, and blood Huntingtin interaction protein 2 (HIP2) mRNA level analysis. Longitudinal self-changes were calculated using repeated measures ANOVA. GEE (generalized estimating equations) was used to assess factors associated with the longitudinal data of rating scales. Switch rates were used for fMRI analysis. False discovery rate correction was used for multiple correction. Results: Participants in the Tai Chi group had better performance in BBS, UPDRS, TUG and step width. Besides, Tai Chi was advantageous over brisk walking in improving BBS and step width. The improved BBS was correlated with enhanced visual network function and downregulation of interleukin-1β. The improvements in UPDRS were associated with enhanced default mode network function, decreased L-malic acid and 3-phosphoglyceric acid, and increased adenosine and HIP2 mRNA levels. In addition, arginine biosynthesis, urea cycle, tricarboxylic acid cycle and beta oxidation of very-long-chain fatty acids were also improved by Tai Chi training. Conclusions: Long-term Tai Chi training improves motor function, especially gait and balance, in PD. The underlying mechanisms may include enhanced brain network function, reduced inflammation, improved amino acid metabolism, energy metabolism and neurotransmitter metabolism, and decreased vulnerability to dopaminergic degeneration. Trial registration This study has been registered at Chinese Clinical Trial Registry (Registration number: ChiCTR2000036036; Registration date: August 22, 2020).

Extensive evidence suggests that frontoparietal regions can dynamically update their pattern of functional connectivity, supporting cognitive control and adaptive implementation of task demands. However, it is largely unknown whether this flexibly functional reconfiguration is intrinsic and occurs even in the absence of overt tasks. Based on recent advances in dynamics of resting-state functional resonance imaging (fMRI), we propose a probabilistic framework in which dynamic reconfiguration of intrinsic functional connectivity between each brain region and others can be represented as a probability distribution. A complexity measurement (i.e., entropy) was used to quantify functional flexibility, which characterizes heterogeneous connectivity between a particular region and others over time. Following this framework, we identified both functionally flexible and specialized regions over the human life span (112 healthy subjects from 13 to 76 years old). Across brainwide regions, we found regions showing high flexibility mainly in the higher-order association cortex, such as the lateral prefrontal cortex (LPFC), lateral parietal cortex, and lateral temporal lobules. In contrast, visual, auditory, and sensory areas exhibited low flexibility. Furthermore, we observed that flexibility of the right LPFC improved during maturation and reduced due to normal aging, with the opposite occurring for the left lateral parietal cortex. Our findings reveal dissociable changes of frontal and parietal cortices over the life span in terms of inherent functional flexibility. This study not only provides a new framework to quantify the spatiotemporal behavior of spontaneous brain activity, but also sheds light on the organizational principle behind changes in brain function across the human life span.SIGNIFICANCE STATEMENTRecent neuroscientific research has demonstrated that the human capability of adaptive task control is primarily the result of the flexible operation of frontal brain networks. However, it remains unclear whether this flexibly functional reconfiguration is intrinsic and occurs in the absence of an overt task. In this study, we propose a probabilistic framework to quantify the functional flexibility of each brain region using resting-state fMRI. We identify regions showing high flexibility mainly in the higher-order association cortex. In contrast, primary and unimodal visual and sensory areas show low flexibility. On the other hand, our findings reveal dissociable changes of frontal and parietal cortices in terms of inherent functional flexibility over the life span.