The Dynamics Of Brain Entropy And Its Relations With Functional Connectivity And Network Topology

As an important measure for irregularity and complexity of the temporal fluctuations of brain activities,brain entropy(BEN)has attracted great attention in rf MRI studies during the last decade.Previous studies have shown its associations with general cognitive abilities and neuropsychiatric disorders.Conventional research assumes BEN is approximately stationary during scan sessions.However,the brain,even at its resting state,is a highly dynamic system,which could be characterized by non-stationary neural activity and rapidly-changing neural interaction.Thus,the one major aim of the current study is to explore the dynamics of brain entropy,and its potential links with general cognitive abilities and neuropsychiatric disorders.The brain also organizes as an adaptive network in which the evolution of the topology would entangle with the dynamics of the local brain regions.Both time-varying nature of brain activity and brain interaction are found to be associated with abilities and neuropsychiatric disorders.It is reasonable to assume the covarying relations between BEN and functional connectivity(or even network topology),and such covarying relations could be associated with cognitive abilities and neuropsychiatric disorders.That serves as the other major aims of our study.In Study 1,we adopted a sliding window approach to calculate the dynamical brain entropy(d BEN)of the whole-brain functional networks from the HCP(Human Connectome Project)rf MRI dataset that includes 812 young adults.The d BEN was further clustered into 4 reoccurring BEN states.The fraction window(FW)and mean dwell time(MDT)of one BEN state,characterized by the lowest overall BEN value and the lower within-state BEN located in SMN and VN,were found to be negatively correlated with general cognitive abilities(i.e.,cognitive flexibility,inhibitory control and processing speed).Another BEN state,characterized by the intermediate overall BEN value and the lower within-state BEN located in DMN,ECN and part of SAN,with its FW and MDT positively correlated with the above cognitive abilities.In Study 2,we adopted the same sliding window approach to calculate whole-brain d BEN from the ADHD-200 rf MRI dataset that includes 239 adolescents(97 ADHD patients).The d BEN was further clustered into 3 reoccurring BEN states.The FW and MDT of one BEN state,characterized by the lowest overall BEN value and the lower within-state BEN located in SMN,were higher in ADHD patients and were positively correlated with symptom severity of ADHD patients.Another BEN state,characterized by the high overall BEN value and the lower within-state BEN was located in DMN and ECN,with its FW and MDT found to be lower in ADHD patients and negatively correlated with symptom severity of ADHD patients.In Study 3,the relation between BEN and functional connectivity(FC),and its association with general cognitive abilities were investigated.FC was found to be varied across different states.In the overall level,the state with the lowest BEN value corresponded to the highest FC strength,and vice versa.At the specific level,the state that is negatively associated with general cognitive abilities,was characterized by the strong connections densely located within SMN and VN.The state positively associated with general cognitive abilities,was characterized by the strong connections densely located between DMN,ECN,and SAN.We also found the d BEN and dynamical functional connectivity(d FC)were negatively correlated in the whole-brain network,and the d BEN-d FC correlations also varied across BEN states.In the overall level,the state with the lowest BEN value corresponded to the highest d BEN-d FC correlation strength,and vice versa.In the specific level,compared with the state which is negatively associated with general cognitive abilities,the locations of strong d BEN-d FC correlations were more concentrated in the connections related to DMN,ECN,and SAN.In Study 4,abnormal covarying relation between BEN and FC in ADHD patients was investigated.The results showed that,compared with the state negatively associated with ADHD symptom severity,strong FCs in the state positively associated with ADHD symptom severity were more densely distributed within SMN,and more sparsely distributed between DMN and ECN.We also found disruption of the whole-brain d BEN-d FC correlations in ADHD patients,especially in the connections related to DMN and ECN.In Study 5,the covarying relation between BEN and brain network topological properties,and its association with general cognitive abilities were investigated.We also identified variations of brain network topological properties across different states.In the overall level,the state with the lowest BEN value corresponded to the lowest global efficiency and highest local efficiency,and vice versa.At the specific level,the state that is negatively associated with general cognitive abilities was characterized by higher nodal local efficiency in SMN and VN,while the state that is positively associated with general cognitive abilities was characterized by both higher nodal efficiency and local efficiency in DMN,ECN,and SAN.We also found negative correlations between d BEN and nodal dynamical efficiency/local efficiency in the whole-brain network.The overall strength of such correlations varied across different states: the state with the lowest BEN value corresponded to the highest d BEN-d Efficiency/Local Efficiency correlation strength,and vice versa.In Study 6,an abnormal covarying relation between BEN and FC in ADHD patients was investigated.The results showed that compared with the state negatively associated with ADHD symptom severity,the state positively associated with ADHD symptom severity corresponded with lower global efficiency and higher local efficiency,lower nodal efficiency and local efficiency in DMN and ECN,and higher nodal local efficiency in SMN.We also identified disruption of d BEN-d Local Efficiency correlation in ADHD patients.In summary,this thesis systematically investigated the dynamics of BEN and the covarying pattern among BEN,FC and network topology.The current findings not only advances our understanding about the brain activity-network topology co-evolution,but also provided new insights into the neuropathology of neuropsychiatric disorders.