Visual deep learning of unprocessed neuroimaging characterises dementia subtypes and generalises across non-stereotypic samples

Keyphrases

• Alzheime's Disease 33%
• Alzheimer's Disease 33%
• Biological Plausibility 33%
• Biological Specificity 33%
• Classification Results 33%
• Clinical Image 33%
• Computer Vision 33%
• Cultural Origin 33%
• Deep Learning Neural Network 33%
• Dementia Diagnosis 33%
• Dementia Subtypes 100%
• Demographic Variables 33%
• DenseNet 66%
• Diagnostic Protocol 33%
• Disease Classification 33%
• Diverse Sample 33%
• Frontotemporal Dementia 66%
• Fully-automatic 33%
• Global South 66%
• Healthy Controls 33%
• Hippocampus 33%
• Homogeneous Samples 33%
• Insula 33%
• Interpretative Phenomenological Analysis 33%
• Latin America (LATAM) 66%
• Model Interpretability 33%
• MRI Signal 33%
• Multi-feature 33%
• Neuroimaging 100%
• Neuroimaging Data 100%
• Occlusion Sensitivity 33%
• Origin Identification 33%
• Out-of-sample Performance 33%
• Robust Classification 33%
• Sample Heterogeneity 33%

INIS

• classification 50%
• computers 25%
• control 25%
• data 100%
• decision making 25%
• diseases 50%
• females 25%
• genetics 25%
• hippocampus 25%
• images 25%
• latin america 50%
• learning 100%
• males 25%
• nervous system diseases 100%
• neural networks 25%
• nmr imaging 50%
• origin 25%
• pipelines 25%
• psychological behavior 25%
• sensitivity 25%
• signals 25%
• specificity 25%
• validation 25%
• vision 25%

Neuroscience

• Alzheimer's Disease 33%
• Computer Vision 33%
• Decision-Making 33%
• Frontal Variant Frontotemporal Dementia 100%
• Genetics 33%
• Hippocampus 33%
• Magnetic Resonance Imaging 66%
• Neural Network 33%