Jie Hao1, M P. Wilson2, N P. Davies2,3, Y Sun2,4, K Natarajan3,4, L MacPherson4, A C. Peet2,4, T N. Arvanitis1,4
1School of Electronic, Electrical & Computer Engineering, University of Birmingham, Birmingham, UK; 2School of Cancer Sciences, University of Birmingham, Birmingham, UK; 3Department of Imaging and Medical Physics, University Hospitals Birmingham NHS Foundation Trust, Birmingham, UK; 4Birmingham Childrens Hospital NHS Foundation Trust, Birmingham, UK
Independent Component Analysis (ICA) has shown the possibility to identify the individual components, and reveal hidden biochemical information about the tissues in MRS. A hybrid ICA approach incorporating the Blind Source Separation (BSS) and Feature Extraction (FE) techniques for automated decomposition of MR spectra is developed and applied to an in vivo paediatric brain tumour MRS dataset. The hybrid method of ICA has the advantages of both BSS and FE, and provides more realistic individual metabolite and MMLip components. It is superior to the well established IC techniques in determining individual metabolite components from brain tumour MRS.