Automatic extraction of metabolic maps from /sup 1/H-2D-CSI data through a wavelet packets decomposition method

摘要

The paper presents a wavelet packets (WP) decomposition method for the post-processing of chemical-shift /sup 1/H-MRS data and for an automatic computation of metabolic maps within anatomical brain regions. WP decomposition is used to analyze the recorded FID signals and to isolate the different metabolite contributions in different frequency bands. In each sub-band LPSDV method has been applied to compute the peak parameters. The main metabolic component are then automatically classified by using peak information such as center frequency and damping factor. The estimated peak amplitudes are then used to construct maps of metabolic concentrations in the analyzed region.