Mobile Menu

Deep learning framework accurately predicts and diagnoses Alzheimer’s Disease

Early and accurate diagnosis of Alzheimer’s remains the most effective way of reducing the burden of the disease, which is the primary cause of dementia worldwide. As current methods lack in both sensitivity and specificity, a team at Boston University School of Medicine have applied a deep learning strategy to delineate Alzheimer’s disease signatures from multi-modal inputs for both risk prediction and diagnosis.

Published in Brain last week the team report a framework that out-performed practising neurologists in diagnosis that’s also clinically adaptable with the use of routine imaging techniques, including MRI.

In constructing the novel framework, MRI results, gender, age, and mini-mental state examination score, were inputted in a fully convolutional network that linked to a traditional multilayer perceptron to create high-resolution maps of disease probability.

The team used four datasets to train and validate the model from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset, Australian Imaging, Biomarker and Lifestyle Flagship Study of Ageing (AIBL), Framingham Heart Study (FHS), and National Alzheimer’s Coordinating Center (NACC).

Model predictions were associated with neuropathological data and put head-to-head with a multi-institutional team of neurologists to assess the validity of prediction performance.

In their work, the team sought to overcome the failings of previous efforts to develop similar models for classifying cognitive status. Other works that lacked demonstration of the underlying diagnostic decision-making, despite promising results, have yet to be incorporated in a clinical setting. The probabilities of this model, the authors say, are intuitive and readily interpretable, a step closer to explainable AI in medicine.

Original article: Development and validation of an interpretable deep learning framework for Alzheimer’s disease classification

Shangran Qiu, Prajakta S Joshi, Matthew I Miller, Chonghua Xue, Xiao Zhou, Cody Karjadi, Gary H Chang, Anant S Joshi, Brigid Dwyer, Shuhan Zhu, Michelle Kaku, Yan Zhou, Yazan J Alderazi, Arun Swaminathan, Sachin Kedar, Marie-Helene Saint-Hilaire, Sanford H Auerbach, Jing Yuan, E Alton Sartor, Rhoda Au, Vijaya B Kolachalama.

More on these topics

AI / Deep Learning / Machine Learning

Share this article