Boost up the detection sensitivity of ASL perfusion fMRI through support vector machine

Data analysis is challenging in arterial spin labeling (ASL) perfusion fMRI due to the intrinsic low SNR of ASL data. To boost up the detection sensitivity, this paper presented a multivariate method based group analysis approach to analyze ASL perfusion fMRI data. A spatial discriminance map (SDM) was first extracted for each subject by support vector machine learning (SVM) algorithm; a population inference about the discriminance was then given by a random effect analysis (RFX) on these individual SDMs. Evaluations were performed using 7 subjects´ fingertapping ASL perfusion fMRI data, yielding similar activation patterns with enhanced sensitivity compared to the standard GLM based group analysis