Analysis of Brain Functional MRI Time Series Based on Continuous Wavelet Transform and Stimulation-Response Coupling Distance

摘要

Analytical techniques applied to functional magnetic resonance imaging (fMRI) data require restrictive assumptions about the shape of the blood oxygenation level dependent (BOLD) time series observed at each voxel in response to a stimulation paradigm. In this paper, a radically different and fundamentally pattern recognition-oriented fMRI brain activation mapping approach is proposed. Neural activity is assessed at each voxel on a coupling distance principle between the deterministic alternation sequence of the stimulation paradigm and the sequence of BOLD response nonstationarities detected and characterized by a continuous wavelet transform (CWT). The voxel's stimulation-response coupling distance is measured using an adapted version of the string edit distance algorithm. fMRI studies conducted on synthetic and real data demonstrated the superiority of the "coupling distance" brain activation mapping approach over against the Student's t-test.