caa a22 4500 606199276 CHVBK 20210128100937.0 cr unu---uuuuu 210128e20150101xx s 000 0 eng 10.1007/s11045-013-0254-3 doi (NATIONALLICENCE)springer-10.1007/s11045-013-0254-3 Voxel selection and neural decoding of fMRI data based on robust sparse programming with multi-dimensional derivative constraints [Elektronische Daten] [Zhuliang Yu, Bao Feng, Zhenghui Gu, Zhenxia Xue, Yuanqing Li, Cong Wang] Recently, sparse representation has received a great deal of attention in voxel selection and decoding analysis of functional magnetic resonance imaging (fMRI) data. Due to contamination of large noise in fMRI data, the conventional sparse representation methods may not get stable results. Moreover, the selected activated brain regions may lose clustering effects and are less biologically interpretable. In order to overcome the above mentioned problems, we exploit the error-tolerant formulation of sparse representation and introduce multi-dimensional derivative constraints (smoothness constraints) in optimization. Two new methods are proposed in this paper. One is robust voxel selection with multi-dimensional constraint (RVSMDC). With the error-tolerant formulation and smoothness constraints on regression coefficients, RVSMDC is robust against noise/error and achieves clustering effects. To directly decode neural activities from fMRI data, we also proposed robust sparse decoding with multi-dimensional constraints (RSDMDC), which minimize the regression error of fMRI data to the task function with sparsity and smoothness constraints on regression coefficients. Numerical results validate the effectiveness of the two proposed methods. Springer Science+Business Media New York, 2013 Voxel selection nationallicence Neural decoding nationallicence fMRI nationallicence Sparse programming nationallicence Multi-dimensional derivative constraints nationallicence Yu Zhuliang Center for Brain Computer Interfaces and Brain Information Processing, South China University of Technology, 510641, Guangzhou, China aut Feng Bao South China University of Technology, 510641, Guangzhou, China aut Gu Zhenghui South China University of Technology, 510641, Guangzhou, China aut Xue Zhenxia Henan University of Science and Technology, 471023, Luoyang, China aut Li Yuanqing South China University of Technology, 510641, Guangzhou, China aut Wang Cong South China University of Technology, 510641, Guangzhou, China aut Multidimensional Systems and Signal Processing Springer US; http://www.springer-ny.com 26/1(2015-01-01), 225-241 0923-6082 26:1<225 2015 26 11045 https://doi.org/10.1007/s11045-013-0254-3 text/html Onlinezugriff via DOI BK010053 XK010053 XK010000 Metadata rights reserved Springer special CC-BY-NC licence nationallicence 1 research-article jats NATIONALLICENCE NATIONALLICENCE NL-springer NATIONALLICENCE 856 40 https://doi.org/10.1007/s11045-013-0254-3 text/html Onlinezugriff via DOI NATIONALLICENCE 700 1- Yu Zhuliang Center for Brain Computer Interfaces and Brain Information Processing, South China University of Technology, 510641, Guangzhou, China aut NATIONALLICENCE 700 1- Feng Bao South China University of Technology, 510641, Guangzhou, China aut NATIONALLICENCE 700 1- Gu Zhenghui South China University of Technology, 510641, Guangzhou, China aut NATIONALLICENCE 700 1- Xue Zhenxia Henan University of Science and Technology, 471023, Luoyang, China aut NATIONALLICENCE 700 1- Li Yuanqing South China University of Technology, 510641, Guangzhou, China aut NATIONALLICENCE 700 1- Wang Cong South China University of Technology, 510641, Guangzhou, China aut NATIONALLICENCE 773 0- Multidimensional Systems and Signal Processing Springer US; http://www.springer-ny.com 26/1(2015-01-01), 225-241 0923-6082 26:1<225 2015 26 11045