Cognitive Impairment Prediction by Normal Cognitive Brain MRI Scans Using Deep Learning

Justin Bardwell, Ghulam Mubashar Hassan, Farzaneh Salami, Naveed Akhtar

Research output: Chapter in Book/Conference paperConference paperpeer-review

1 Citation (Scopus)

Abstract

Alzheimer’s disease is a neurodegenerative disease without a cure and is one of the leading causes of death across the world. The early detection of cognitive impairment could prove crucial for reducing the occurrence of Alzheimer’s disease in the future. Significant research into detecting the disease from MRI images has already been performed and has produced encouraging results. However, there has been very limited work on predicting conversion from normal cognition to cognitive impairment. This study is aimed at producing a deep learning model to predict whether a subject will remain cognitively normal or progress to a state of cognitive impairment in the future. We found that the use of a patch-based approach combined with pre-trained ResNet-50 model using 3D MRI scans provide better results as compared to equivalent whole brain voxel-based approach and other state-of-the-art CNN models. Our proposed model achieved an accuracy of 90% and an area under the receiver operating characteristic curve of 0.99, which are better than the existing state-of-the-art results.

Original languageEnglish
Title of host publicationAI 2022
Subtitle of host publicationAdvances in Artificial Intelligence - 35th Australasian Joint Conference, AI 2022, Proceedings
EditorsHaris Aziz, Débora Corrêa, Tim French
PublisherSpringer Science + Business Media
Pages571-584
Number of pages14
ISBN (Print)9783031226946
DOIs
Publication statusPublished - 2022
Event35th Australasian Joint Conference on Artificial Intelligence, AI 2022 - Perth, Australia
Duration: 5 Dec 20229 Dec 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13728 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference35th Australasian Joint Conference on Artificial Intelligence, AI 2022
Country/TerritoryAustralia
CityPerth
Period5/12/229/12/22

Fingerprint

Dive into the research topics of 'Cognitive Impairment Prediction by Normal Cognitive Brain MRI Scans Using Deep Learning'. Together they form a unique fingerprint.

Cite this