This is a free and open source Matlab program of 2CFastICA, a deflationary high density surface EMG decomposition method. 2CFastICA incorporates two key algorithms, namely kernel constrained FastICA and correlation constrained FastICA. The former searches new motor units in the kernel space constructed by using the identified results, therefore overcoming the local convergence of FastICA. The latter further refines the output of kernel constrained FastICA to warrant the decomposition accuracy.
Details of 2CFastICA can be found in the following paper:
M. Chen and P. Zhou, "2CFastICA: A Novel Method for High Density Surface EMG Decomposition Based on Kernel Constrained FastICA and Correlation Constrained FastICA," in IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 32, pp. 2177-2186, 2024, doi: 10.1109/TNSRE.2024.3398822.
The program is for research and non-commercial purposes only. Please provide appropriate citation when using the program. For any questions regarding the program, please contact Maoqi Chen, PhD, at maoqi.chen@uor.edu.cn or hiei@mail.ustc.edu.cn.