Sparse Representation of Higher-Order Functional Interaction Patterns in Task-Based FMRI Data

详细信息    查看全文

• 作者：Shu Zhang ; Xiang Li ; Jinglei Lv ; Xi Jiang ; Dajiang Zhu ; Hanbo Chen ; Tuo Zhang ; Lei Guo
• 关键词：task ; based fMRI ; GES ; sparse coding ; dictionary learning ; higherorder interaction
• 刊名：Lecture Notes in Computer Science
• 出版年：2013
• 出版时间：2013
• 年：2013
• 卷：8151
• 期：1
• 页码：635-642
• 全文大小：1469KB
• 参考文献：1. Friston, K.J., et al.: Statistical parametric maps in functional imaging: a general linear approach. Human Brain Mapping?2(4), 189-10 (1994) CrossRef
  2. Lv, J., Guo, L., Li, K., Hu, X., Zhu, D., Han, J., Liu, T.: Activated Fibers: Fiber-centered Activation Detection in Task-based FMRI. In: Székely, G., Hahn, H.K. (eds.) IPMI 2011. LNCS, vol.?6801, pp. 574-87. Springer, Heidelberg (2011) CrossRef
  3. Li, X., Lim, C., Li, K., Guo, L., Liu, T.: Detecting Brain State Changes via Fiber-Centered Functional Connectivity Analysis. Neuroinformatics (2012)
  4. Sun, J., et al.: Inferring Consistent Functional Interaction Patterns from Natural Stimulus FMRI Data. NeuroImage (2012)
  5. Ramsey, J., et al.: Six problems for causal inference from fMRI. Neuroimage?49(2), 1545-558 (2010) CrossRef
  6. Calhoun, V.D., et al.: A method for making group inferences from functional MRI data using independent component analysis. Human Brain Mapping?14(3), 140-51 (2001) CrossRef
  7. Roebroeck, A., Formisano, E., Goebel, R.: Mapping directed influence over the brain using Granger Causality and fMRI. NeuroImage?25, 230-42 (2005) CrossRef
  8. Friston, K.J., et al.: Dynamic causal modelling. NeuroImage?19(3), 1273-302 (2003) CrossRef
  9. Meek, C.: Graphical Models: Selecting Causal and Statistical Models. Carnegie Mellon University, Pittsburgh (1997)
  10. Mairal, J., Bach, F., Ponce, J., Sapiro, G.: Online dictionary learning for sparse coding. In: Proceedings of the International Conference on Machine Learning, ICML (2009)
  11. Faraco, C.C., et al.: Complex span tasks and hippocampal recruitment during working memory. NeuroImage?55(2), 773-87 (2011) CrossRef
  12. Zhu, D., Li, K., Guo, L., Jiang, et al.: DICCCOL: Dense Individualized and Common Connectivity-Based Cortical Landmarks. Cerebral Cortex (2012)
  13. Susanne, G., et al.: DEAL: A Package for Learning Bayesian Networks (2003)
  14. Eberhardt, F., Hoyer, P.O., Scheines, R.: Combining Experiments to Discover Linear Cyclic Models with Latent Variables. Journal of Machine Learning, Workshop and Conference Proceedings (AISTATS 2010)?9, 185-92 (2010)
  15. Mairal, J., Bach, F., Ponce, J., Sapiro, G.: Online learning for matrix factorization and sparse coding. Journal of Machine Learning Research?11, 19-0 (2010)
  16. Shu, Z., et al.: Activated Cliques: Network-based Activation Detection in Task-based FMRI. In: ISBI 2013 (2013)
  
• 作者单位：Shu Zhang (21) (22)
  Xiang Li (21)
  Jinglei Lv (22)
  Xi Jiang (21)
  Dajiang Zhu (21)
  Hanbo Chen (21)
  Tuo Zhang (22)
  Lei Guo (22)
  Tianming Liu (21)
  
  21. Department of Computer Science and Bioimaging Research Center, The University of Georgia, Athens, GA, USA
  22. School of Automation, Northwestern Polytechnical University, Xi’an, China
  
• ISSN：1611-3349

文摘

Traditional task-based fMRI activation detection methods, e.g., the widely used general linear model (GLM), assume that the brain’s hemodynamic responses follow the block-based or event-related stimulus paradigm. Typically, these activation detections are performed voxel-wise independently, and then are usually followed by statistical corrections. Despite remarkable successes and wide adoption of these methods, it remains largely unknown how functional brain regions interact with each other within specific networks during task performance blocks and in the baseline. In this paper, we present a novel algorithmic pipeline to statistically infer and sparsely represent higher-order functional interaction patterns within the working memory network during task performance and in the baseline. Specifically, a collection of higher-order interactions are inferred via the greedy equivalence search (GES) algorithm for both task and baseline blocks. In the next stage, an effective online dictionary learning algorithm is utilized for sparse representation of the inferred higher-order interaction patterns. Application of this framework on a working memory task-based fMRI data reveals interesting and meaningful distributions of the learned sparse dictionary atoms in task and baseline blocks. In comparison with traditional voxel-wise activation detection and recent pair-wise functional connectivity analysis, our framework offers a new methodology for representation and exploration of higher-order functional activities in the brain.