Spatial Independent Component Analysis for Multi-task Functional MRI Data Processing

Multi-task in one epoch usually exists in our life, but the brain functional activation response is few analyzed. One key problem is that the multi-task response data processing has not been addressed. In this paper, spatial independent component analysis (sICA) is presented to separate the different response of the complex visual-movement task by analyzing the unmixing matrices temporal components corresponding to spatial separated pattern. An invivo fMRI experiment, with ten subjects (five male, five female) with the visual stimulation and hands’ movement synchronously, is performed. Separated component patterns of the spatial ICA are chosen by computing the relation between the unmixing matrices temporal component corresponding to spatial separated pattern and experiment pattern. The result shown that the spatial ICA can separate the two response activation patterns, and the response of the unmixing matrices temporal information corresponding to spatial component patterns between two tasks are obvious different. The visual response is obvious faster than the movement response by analyzing relation between the neuron dynamic response and unmixing matrices temporal response.