2868 Reduced Local BOLD Correlation in PreFrontal Cortex
# 2868 Reduced Local BOLD Correlation in Pre-Frontal Cortex during Deep Sleep Xiao Liu 1, Silvina G. Horovitz 1, 2, Allen R. Braun 3, Walter S. Carr 4, Dante Picchionie 5, Masaki Fukunaga 1, and Jeff H. Duyn 1 1 AMRI, LFMI, NINDS, NIH 2 HMCS, NINDS, NIH 3 VSLB, NIDCD, NIH, 4 Naval Medical Research Center, 5 Department of Behavioral Biology, Walter Reed Army Institute of Research Melbourne, Australia May 5 th-11 th, 2012
q Prefrontal part of default mode network (DMN) reduces its BOLD correlations with other DMN components during deep sleep. Background Methods Results Discussions Decoupled Prefrontal DMN Component During Deep Sleep Further question: other than the long-range BOLD correlation, does the local BOLD correlations also change during deep sleep? Horovitz, S. G. , PNAS, 2009
Discussions Reanalyzing Previous Dataset REPROCESSING NOTES q Reprocess dataset published previously (Horovitz, S. G. , PNAS, 2009) Results q Only include 7 subjects with separate wake data acquisition q Divide data into 300 -sec segments q Carefully avoid periods with serious head motion q Separate pure Stage 2 from those mixed with Stage 3~4 sleep Methods q Regress out global signal and physiological noise q Align to Harvard-Oxford-Atlas Cortical Background ROI # 1 Subcortical 48 55
Discussions Local BOLD Correlation: Regional Homogeneity (Re. Ho*) Re. Ho Map Results Averaged pairwise BOLD correlations within a 3 x 3 x 3 voxel cube Wake Deep Sleep Log (p-value) Methods Using atlas ROIs Statistical Inference for each ROI Background PCC Example An Example of ROIs: Posterior Cingulate Cortex ROI # * Zhang, Y. , Neuro. Image, 2004
Results Re. Ho Discussions ROI Analysis Result 0. 8 0. 6 0. 4 Methods 0. 2 Wake Deep Sleep Background p-value < 0. 05 * < 0. 01 ** <0. 001 *** Decreased Local correlation Increased Local correlation
q ROIs showing significant Re. Ho modulation during deep sleep Methods Results Discussions ROI Analysis Result Background p-value (Bonf. Corrected) 10 -3 Anterior and posterior brain show 10 Wake > Deep Sleep Wake < Deep Sleep opposite modulation ! 10 -10
Discussions Group Re. Ho Map Deep Sleep (n=58) Results Wake (n=27) 0 0 1 1 Deep Sleep − Wake Background Methods Deep Sleep − Wake -0. 2 p (Bonf. Corrected) 0. 05 0. 2 10 -10 < Wake 10 -10 > Wake
Discussions Group Re. Ho Map Deep Sleep (n=58) Results Wake (n=27) 0 Amplitude of f. MRI BOLD fluctuation 0 1 ? Deep Sleep − Wake Local f. MRI BOLD correlation change Methods Background 1 -0. 2 p (Uncorrected) 10 -3 0. 2 10 -10 < Wake 10 -10 > Wake
Discussions Group Map of BOLD Fluctuation Amplitude (Standard Deviation) Deep Sleep (n=58) Results Wake (n=27) 0% 0% 2% 2% Deep Sleep − Wake Background Methods Deep Sleep − Wake -1% p (Bonf. Corrected) 0. 05 1% 10 -10 < Wake 10 -10 > Wake
Discussions Group Map of BOLD Fluctuation Amplitude (Standard Deviation) Deep Sleep (n=58) Results Wake (n=27) 0% No correspondent modulation in BOLD 0% 2% fluctuation amplitude 2% Deep Sleep − Wake Background Methods Deep Sleep − Wake -1% p (Uncorrected) 10 -3 1% 10 -10 < Wake 10 -10 > Wake
Results Discussion q During deep sleep, local f. MRI signal correlations alter in a distinct, spatially consistent manner, with prefrontal regions showing a decrease and posterior regions showing an increase. Background Methods q This observation is consistent with the notion that higher level cognitive function, commonly associated with frontal cortex, is reduced during sleep. q Analysis of local correlations patterns may provide information about the brain’s cognitive state that is supplemental to information derived from long-range correlation analysis
Discussions Acknowledgements Dr. Thomas J. Balkin, Results NINDS, WRAIR, and NIDCD Background Methods References [1] Biswal, B. et al. MRM 1995 [2] Horovitz, SG. et al. PNAS 2009 [3] Spoormaker, VI. et al. J. Neurosci. 2010. [4] Boly, M. et al. Ann N Y Acad Sci. 2008. [5] Vincent, JL. et al. Nature. 2007. [6] Desikan, RS. et al. Neuro. Image. 2006. Thank you!
- Slides: 12