TY - JOUR
T1 - Influence of epoch length on measurement of dynamic functional connectivity in wakefulness and behavioural validation in sleep
AU - Wilson, Rebecca S.
AU - Mayhew, Stephen D.
AU - Rollings, David T.
AU - Goldstone, Aimee
AU - Przezdzik, Izabela
AU - Arvanitis, Theodoros N.
AU - Bagshaw, Andrew P.
PY - 2015/5/5
Y1 - 2015/5/5
N2 - Conventional functional connectivity (FC) analysis of fMRI data derives a single measurement from the entire scan, generally several minutes in duration, which neglects the brain's dynamic behaviour and potentially loses important temporal information. Short-interval dynamic FC is an attractive proposition if methodological issues can be resolved and the approach validated. This was addressed in two ways; firstly we assessed FC of the posterior cingulate cortex (PCC) node of the default mode network (DMN) using differing temporal intervals (8. s to 5. min) in the waking-resting state. We found that 30-second intervals and longer produce spatially similar correlation topography compared to 15-minute static FC measurements, while providing increased temporal information about changes in FC that were consistent across interval lengths. Secondly, we used NREM sleep as a behavioural validation for the use of 30-second temporal intervals due to the known fMRI FC changes with sleep stage that have been observed in previous studies using intervals of several minutes. We found significant decreases in DMN FC with sleep depth which were most pronounced during stage N2 and N3. Additionally, both the proportion of time with strong PCC-DMN connectivity and the variability in dynamic FC decreased with sleep. We therefore show that dynamic FC with epochs as short as tens of seconds is a viable method for characterising intrinsic brain activity.
AB - Conventional functional connectivity (FC) analysis of fMRI data derives a single measurement from the entire scan, generally several minutes in duration, which neglects the brain's dynamic behaviour and potentially loses important temporal information. Short-interval dynamic FC is an attractive proposition if methodological issues can be resolved and the approach validated. This was addressed in two ways; firstly we assessed FC of the posterior cingulate cortex (PCC) node of the default mode network (DMN) using differing temporal intervals (8. s to 5. min) in the waking-resting state. We found that 30-second intervals and longer produce spatially similar correlation topography compared to 15-minute static FC measurements, while providing increased temporal information about changes in FC that were consistent across interval lengths. Secondly, we used NREM sleep as a behavioural validation for the use of 30-second temporal intervals due to the known fMRI FC changes with sleep stage that have been observed in previous studies using intervals of several minutes. We found significant decreases in DMN FC with sleep depth which were most pronounced during stage N2 and N3. Additionally, both the proportion of time with strong PCC-DMN connectivity and the variability in dynamic FC decreased with sleep. We therefore show that dynamic FC with epochs as short as tens of seconds is a viable method for characterising intrinsic brain activity.
KW - Dynamic functional connectivity
KW - NREM sleep
KW - Seed-based analysis
UR - https://www.sciencedirect.com/science/article/pii/S1053811915001718
UR - http://www.scopus.com/inward/record.url?scp=84925131205&partnerID=8YFLogxK
U2 - 10.1016/j.neuroimage.2015.02.061
DO - 10.1016/j.neuroimage.2015.02.061
M3 - Article
C2 - 25765256
AN - SCOPUS:84925131205
SN - 1053-8119
VL - 112
SP - 169
EP - 179
JO - NeuroImage
JF - NeuroImage
ER -