Processing strategies for time-course data sets in functional mri of the human brain

摘要: Image processing strategies for functional magnetic resonance imaging (FMRI) data sets acquired using a gradient-recalled echo-planar sequence are considered. The analysis is carried out the mathematics of vector spaces. Data consisting N sequential images same slice brain tissue analyzed in time-domain and also, after Fourier transformation, frequency domain. A technique thresholding introduced that uses shape response pixel compared with reference waveform as decision criterion. method presented to eliminate drifts arise from subject movement. methods applied experimental FMRI motor—cortex more conventional image—subtraction methods. Several finger motion paradigms considered context various image strategies. most effective involves by characterized correlation coefficient respect followed formation cross-correlation image. Emphasis placed not only on formation, but also use signal techniques characterize temporal paradigm.