“School of Cognitive Sciences”

Back to Papers Home
Back to Papers of School of Cognitive Sciences

Paper   IPM / Cognitive Sciences / 14228
School of Cognitive Sciences
  Title:   Sparse Reconstruction of Localized Correlated Spectroscopy: From Sub-Sampled Priors to Fast Acquisition
  Author(s): 
1.  M. Abdi-Shektaei
2.  A. Nasiraei Moghaddam
3.  R. Nagarajan
4.  M. Albert Thomas
  Status:   In Proceedings
  Proceeding: Proc. Intl. Soc. Mag. Reson. Med. 23 (2015)
  Year:  2015
  Supported by:  IPM
  Abstract:
INTRODUCTION: Despite its successful application to detect many pathological disorders in brain [1, 2], breast [3] and prostate [4], Localized Correlated Spectroscopy (L-COSY) still suffers from long acquisition that prevents this technique to enter clinical studies [5]. Inherent sparse nature of L-COSY can be exploited to accelerate the long acquisition times ensued from iterative sampling along indirect dimension (t1). In this study, a new reconstruction strategy is presented that takes the advantage of sparsity to reconstruct the whole L-COSY spectrum with less number of acquisitions along the t1 dimension. METHOD: The proposed reconstruction strategy consists of two main steps: first finding the location of diagonal and cross peaks from subsampling priors, and then quantifying the spectrum at each peak. Using sparse nature of spectrum, the location of diagonal and cross peaks could be determined from deterministic subsampling along t1. The amplitude of peaks is calculated through iterative least-square (LS) on those deterministic subsampled t1 data. The (LS) problem is formulated as follows: The relationship between time domain matrix (T) and spectrum domain matrix (S) is 2D Fourier Transform which is presented is Eq 1.

Download TeX format
back to top
scroll left or right