An imbalance between functional segregation and integration in patients with pontine stroke: A dynamic functional network connectivity study

https://doi.org/10.1016/j.nicl.2020.102507Get rights and content
Under a Creative Commons license
open access

Highlights

  • Pontine stroke patients show abnormal time-varying brain activity.

  • Pontine stroke patients exist aberrant functional segregation and integration.

  • The alterations of dFNC may lead to a poor functional recovery after stroke.

Abstract

Background

Previous studies on brain functional connectivity have revealed the neural physiopathology in patients with pontine stroke (PS). However, those studies focused only on the static features of intrinsic fluctuations, rather than on the time-varying effects throughout the entire scan. In the present study, we sought to explore the underlying mechanism of PS using the dynamic functional network connectivity (dFNC) method.

Methods

Resting-state functional magnetic resonance imaging (fMRI) data were collected from 58 patients with PS and 52 healthy controls (HC). Independent component analysis (ICA), the sliding window method, and k-means clustering analysis were performed to extract different functional networks, to calculate dFNC matrices, and to estimate distinct dynamic connectivity states. Additionally, temporal features were compared between the two groups in each state to explore the brain’s preference for different dynamic connectivity states in PS, and global and local efficiency were compared among states to explore the differences of topologic organization across different dFNC states. The correlations between clinical scales and the temporal features that differed between the two groups also were calculated.

Results

The dFNC analyses suggested four recurring states; in two of these states, the PS group showed a different duration from that of the HC group. Patients with PS spent significantly more time in a sparsely connected state (State 1), which was characterized by relatively low levels of connectivity within and between all brain networks. In contrast, patients with PS spent significantly less time in a highly segregated state (State 2), which was characterized by high levels of positive connectivities within primary perceptional domains and within higher cognitive control domains, and by high levels of negative inter-functional connectivities (inter-FCs) among primary perceptional and higher cognitive control domains. Additionally, the dwell time in State 2 was positively correlated with HC group’s long-term memory scores in the Rey Auditory Verbal Learning Test (RAVLT-L), whereas there was no correlation between the State-2 dwell time and RAVLT-L scores in the PS group. Furthermore, the sparsely connected state and the highly segregated state mentioned above had the highest global efficiency and the highest local efficiency among the four states, respectively.

Conclusions

In summary, we observed a preference in the aberrant brain for dynamic connectivity states with different network topologic organization in patients with PS, indicating abnormal functional segregation and integration of the whole brain and confirming the imperfection of functional network connectivity in patients with PS. These findings provide new evidence for the dynamic neural mechanisms underlying clinical symptoms in patients with PS.

Abbreviations

PS
pontine stroke
dFNC
dynamic functional network connectivity
HC
healthy controls
ICA
independent component analysis
fMRI
functional magnetic resonance imaging
RAVLT
Rey Auditory Verbal Learning Test
FC
functional connectivity
BOLD
blood oxygen level-dependent
SBA
seed-based analysis
VMHC
voxel-mirrored homotopic connectivity
IC
independent component
RSN
resting-state network
FLAIR
fluid-attenuated inversion recovery
TR
repetition time
TE
echo time
FOV
field of view
3D-T1WI
sagittal three-dimensional T1-weighted images
BRAVO
brain volume sequence
MNI
Montreal Neurological Institute
DARTEL
Diffeomorphic Anatomical Registration Through Exponentiated Lie algebra
PCA
principal component analysis
FDR
false discovery rate
AUC
area under the curve
FMT
Fugl-Meyer Test
ACC
accuracy
RT
reaction time
SD
standard deviation
AUN
auditory network
VIS
visual network
SMN
sensorimotor network
ATN
attention network
DMN
default mode network
ECN
executive control network
SAN
salience network
CB
cerebellar network
RF
reoccurrence fraction
DT
mean dwell time
TTN
total transition number
CBF
cerebral blood flow
GMV
gray matter volume
rCEN
right central executive network

Keywords

Pontine stroke
Resting-state functional magnetic resonance imaging
Dynamic functional network connectivity
Graph theoretic analysis
Functional integration and segregation

Cited by (0)

1

These authors contributed equally.