| The brain networks undergo functional reorganization across the whole lifespan,but the dynamic patterns behind the reorganization remain largely unclear.Although some studies using dynamic functional connectivity(d FC)based on sliding window approach have discovered the relationship between spontaneous dynamic states and age,their findings are inconsistent and may suffer from a number of methodological limitations.Comparing with SW-based d FC approach,hidden Markov model(HMM)based on single frames assumes that f MRI time courses are generated by a number of hidden states that following a Markov chain and has huge potential for discovering the dynamics of brain activity.Therefore,we employed HMM to discover the changes of hidden state dynamic across the adult lifespan.This study models the dynamics of spontaneous activity of large-scale networks using HMM,and investigates how these “brain states” change with age on two adult lifespan datasets of 176/157 subjects(aged 20-80 years).Results showed that(A)for both datasets,older adults tended to spend less time on a state where default mode network(DMN)and attentional networks(ATNs)show antagonistic activity,and(B)older adults spent more time on and had less transitions from and more transitions to a ‘baseline’ state with moderate-level activation of all networks,which validated the prediction of Naik et al.(C)HMM exhibited higher sensitivity,robustness and reproducibility in uncovering the age effects compared with temporal clustering method,and was proved to be the unique dynamical approach in discovering “baseline” state.Our results suggest that the aging brain is characterized by the shortening of the antagonistic instances between DMN and attention systems,as well as the prolongation of the inactive period of all networks,which might reflect the shift of the dynamical working point near criticality in older adults. |
