Abstract #3263

Determination of Alzheimer's disease based on morphology and atrophy using machine learning combined with automated segmentation

Natsuki Ikemitsu¹, Yuki Kanazawa², Akihiro Haga², Hiroaki Hayashi³, Yuki Matsumoto², and Masafumi Harada²

¹Graduate school of Health Science, Tokushima University, Tokushima, Japan, ²Institute of Biomedical Sciences, Tokushima University Graduate School, Tokushima, Japan, ³Faculty of Health Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa, Japan

Synopsis

To classify healthy subjects or patients with Alzheimer's disease (AD) using three-dimensional T₁w data, we developed a machine learning system which can capture morphology features and determine atrophy of brain tissue in early-stage AD. Deep learning, a support vector machine (SVM), and 3D convolutional neural networks (3DCNN) were performed. The accuracies of SVM and deep learning based on volume values were comparable and greatly exceeded the accuracy of 3DCNN. It was found that atrophic features were more considerable than morphological features in early-stage AD.

This abstract and the presentation materials are available to members only; a login is required.

Join Here