A Gaussian Mixture Model Based Diagnosis of Alzheimer´s Using Diffusion Tensor Imaging

Diffusion Tensor Imaging (DTI) is increasingly being used to study the damage of brain microstructure due to neurodegenerative disorder. In this work an attempt is made to evaluate the Gaussian Mixture Model (GMM) for classification of Alzheimer´s, healthy controls and Mild Cognitive Impairment (MCI) subjects using diffusion tensor indices. GMM´s performance is evaluated against linear discriminant analysis and Parzen window techniques of classification. Close to 90% classification accuracy has been achieved using this approach. The early diagnosis of Alzheimer´s plays a critical role since the effect of drugs reduce drastically as disease becomes more pronounced thus this technique can be a viable tool for mass screening of Alzheimer disease.