Two-Stage Evolutionary Quantification of In Vivo MRS Metabolites

The main subject of this work is the in vivo quantification of the metabolites concentrations revealed in the magnetic resonance spectroscopy (MRS) spectra. For this purpose, a novel two-stage processing methodology, consisting of the denoising of the MRS signal and the quantification of the metabolites’ peaks using a genetic algorithm (GA), is proposed. The denoising stage tends to improve the quality of the acquired MRS signal in a way that makes the fitting procedure performed by the genetic algorithm (GA) more successful. Two different approaches for improving the MRS signal quality, the denoising via wavelet analysis and signal separation by singular value decomposition (SVD), under possible combinations are examined. The introduced quantification technique deals with metabolites’ peaks overlapping, a considerably difficult situation occurred in real conditions. Extensive experiments have proved the efficiency of the introduced methodology in artificial MRS data by establishing it as a generic metabolite quantification procedure.

[1]  Hojjat Adeli,et al.  A parallel genetic/neural network learning algorithm for MIMD shared memory machines , 1994, IEEE Trans. Neural Networks.

[2]  Taghi M. Khoshgoftaar,et al.  Hybrid sampling for imbalanced data , 2008, 2008 IEEE International Conference on Information Reuse and Integration.

[3]  A Heerschap,et al.  Common processing of in vivo MR spectra , 2001, NMR in biomedicine.

[4]  D. A. Karras,et al.  In vivo MRS metabolite quantification using genetic optimization , 2011 .

[5]  David Coley,et al.  Introduction to Genetic Algorithms for Scientists and Engineers , 1999 .

[6]  D.A. Karras,et al.  On quantifying MRS metabolites using a constrained Genetic Algorithm , 2010, 2010 IEEE International Conference on Imaging Systems and Techniques.

[7]  N. Schuff,et al.  Magnetic resonance spectroscopic imaging reconstruction with deformable shape‐intensity models , 2003, Magnetic resonance in medicine.

[8]  P Boesiger,et al.  Heuristic optimization algorithms applied to the quantification of spectroscopic data , 1998, Magnetic resonance in medicine.

[9]  H. Adeli,et al.  Concurrent genetic algorithms for optimization of large structures , 1994 .

[10]  Truong Q. Nguyen,et al.  Wavelets and filter banks , 1996 .

[11]  S. Huffel,et al.  MR spectroscopy quantitation: a review of time‐domain methods , 2001, NMR in biomedicine.

[12]  Alicia Troncoso Lora,et al.  Mining quantitative association rules based on evolutionary computation and its application to atmospheric pollution , 2010, Integr. Comput. Aided Eng..

[13]  David L. Donoho,et al.  De-noising by soft-thresholding , 1995, IEEE Trans. Inf. Theory.

[14]  D. van Ormondt,et al.  SVD-based quantification of magnetic resonance signals , 1992 .

[15]  Subhash C. Bagui,et al.  Combining Pattern Classifiers: Methods and Algorithms , 2005, Technometrics.

[16]  Hojjat Adeli,et al.  Distributed Genetic Algorithm for Structural Optimization , 1995 .

[17]  Dimitris A. Karras,et al.  A Constrained Genetic Algorithm with Adaptively Defined Fitness Function in MRS Quantification , 2010, FGIT-GDC/CA.

[19]  S Mierisová,et al.  MR spectroscopy quantitation: a review of frequency domain methods , 2001, NMR in biomedicine.

[20]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .