Identifying stage of Alzheimer disease using multiclass particle swarm optimisation technique

• Published in: Journal of experimental & theoretical artificial intelligence Vol. 30; no. 6; pp. 911 - 925
• Main Author: Jiji, G. Wiselin
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 02.11.2018; Taylor & Francis Ltd
• Subjects: Biomarkers; Brain; Change detection; Classification; Diagnosis; Empirical analysis; grey matter; Image detection; Magnetic resonance; Medical imaging; particle swarm optimisation technique; Particle swarm optimization; segmentation; Volumetric analysis
• ISSN: 0952-813X; 1362-3079
• DOI: 10.1080/0952813X.2018.1509380

Detecting brain structural changes from magnetic resonance (MR) images can facilitate early diagnosis and treatment of neurological and psychiatric diseases. Alzheimer Disease (AD) is a progressive neurodegenerative disorder that causes structural changes in patient's brain. As such, it is essential to develop an algorithm for identifying the biomarkers of this disease stage. We developed a novel volumetric analysis of anatomical components of brain with multiclass particle swam optimisation technique (MPSO) approach to detect the stages of AD as potential biomarkers. To avoid image distortion bias correction is applied. We have used anatomical structures i.e. tissue and ventricle volume are used as criteria to categorise image features into four classes such as Alzheimer Mild cognitive decline, Alzheimer Moderate Cognitive decline and Alzheimer Severe Cognitive decline and healthy subject. This work was experimented with 30 AD and 10 normal cases. We observed that grey matter content was reduced from 4 to 20% of normal brain and volume of ventricle is increasing gradually from mild to severe cognitive decline. The statistical performance measures are calculated for proposed and existing work. The value shows that our empirical evaluation has superior diagnosis performance. We found that AD patient's brain has reduced volume in grey matter and subsequently shrunk the volume of brain. The size of ventricle is also the major concern to predict the severity of AD disease. Therefore, the volumes of grey matter and ventricle size more discriminately classify the AD patient with severity from normal subject.