A Structural Parametrization of the Brain Using Hidden Markov Models Based Paths in Alzheimer's Disease

Abstract

The usage of biomedical imaging in the diagnosis of dementia is increasingly widespread. A number of works explore the possibilities of computational techniques and algorithms in what is called Computed Aided Diagnosis. Our work presents an automatic parametrization of the brain structure by means of a path generation algorithm based on Hidden Markov Models. The path is traced using information of intensity and spatial orientation in each node, adapting to the structural changes of the brain. Each path is itself a useful way to extract features from the MRI image, being the intensity levels at each node the most straightforward. However, a further processing consisting of a modification of the Gray Level Co-occurrence Matrix can be used to characterize the textural changes that occur throughout the path, yielding more meaningful values that could be associated to the structural changes in Alzheimer's Disease, as well as providing a significant feature reduction. This methodology achieves high performance, up to 80.3\% of accuracy using a single path in differential diagnosis involving Alzheimer-affected subjects versus controls belonging to the Alzheimer's Disease Neuroimaging Initiative (ADNI).