Assessment of Geriatric-Specific Changes in Brain Texture Complexity Using a Backpropagation Neural Network Classifier

A method to assess the aging of a human subject by modeling the devolution of the textural features in brain images using a backpropagation neural network (BPNN) is described in this paper. Normally, the brain white matter (BWM) undergoes degenerative changes in its physical and functional stochastics during the aging process. Relevant structural morphology observed in the brain complex can be measured via diffusion tensor magnetic resonance imaging (DTMRI). Using the underlying statistical details of the pixels in the brain image captured, BPNN is used to classify the distinct BWM parameters, which are then correlated to the subject’s age. The brain complex invariably shows an evolutionary changing trend (in the negative direction) in its textural features during the aging process. Clinical DTMRI datasets from subjects of different age groups are used to study the efficacy of the proposed method of correlating brain-textural degeneration versus age.