On quantifying MRS metabolites using a constrained Genetic Algorithm

The MRS metabolites quantification procedure has attracted the scientific interest of the engineering community, regarding the development of noninvasive and computationally efficient methodologies. Significant contributions based on Artificial Intelligence (AI) tools, such as Neural Networks (NNs), with good results have been presented lately but showing several drawbacks already discussed by the authors. Also, preliminary application of Genetic Algorithms (GA) has already been reported in the literature by the authors regarding the peak detection problem encountered in MRS spectroscopy. A novel constrained genetic algorithm is investigated in this paper aiming at extending the simple genetic algorithm methodology in case of noisy signals as well as at addressing the issue of quantifying MRS metabolites in artificial MRS signals. Although additional experiments with real MRS data are needed, the herein presented results illustrate the method's potential in MRS spectroscopy to be established as a generic metabolite quantification procedure.