Title: Automated computer aided system for early diagnosis of Alzheimer's disease by regional atrophy analysis in functional magnetic resonance imaging

Authors: R. Sampath; J. Indumathi

Addresses: Department of Information Science and Technology, Anna University, Chennai, India ' Department of Information Science and Technology, Anna University, Chennai, India

Abstract: A growing body of researchers suggest that a preanalytic step of neurogenetic Alzheimer's disease (AD) characterised by precise neuropsychological and brain changes, may exist several years earlier to the overt manifestation of clinical symptoms. FMRI is a one of the effective neuroimage which helps to predict the brain change in AD. This paper introduces the automatic AD disease detection system (ACAS) for early diagnosis of AD using FMRI. The system consists of four stages: preprocessing, feature extraction, segmentation and regional atrophy analyses. The preprocessing removes the noise in the FMRI image; multiscale analysis (MSA) is used to analyse FMRI to obtain its fractals at six different scales, which produce different feature vectors to discriminate between healthy and pathological patients; self-organising map (SOM) is applied to divide the affected region which is one of the effective unsupervised network that utilises the obtained feature vectors for competitive learning; the regional atrophy analyses are used to differentiate AD from other neurodegenerative diseases. Compared to MRI, the proposed system gives more satisfactory results for early diagnosis and differentiation of AD from other neurodegenerative diseases.

Keywords: neurogenetic Alzheimer's disease; AD; automated computer aided system; ACAS; multiscale analysis; MSA; self-organising map network; SOMN; regional atrophy analysis.

DOI: 10.1504/IJBET.2020.107205

International Journal of Biomedical Engineering and Technology, 2020 Vol.32 No.4, pp.305 - 316

Received: 14 Sep 2017
Accepted: 12 Jan 2018
Published online: 11 May 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article