Towards the interpretability of deep learning models for multi-modal neuroimaging: Finding structural changes of the ageing brain
Year: 2022
Type: article
Abstract: Brain-age (BA) estimates based on deep learning are increasingly used as neuroimaging biomarker for brain health; however, the underlying neural features have remained unclear. We combined ensembles of convolutional neural networks with Layer-wise Relevance Propagation (LRP) to detect which brain features contribute to BA. Trained on magnetic resonance imaging (MRI) data of a population-based study (n = 2637, 18-82 years), our models estimated age accurately based on single and multiple modalities, regionally restricted and whole-brain images (mean absolute errors 3.37-3.86 years). We find that BA estimates capture ageing at both small and large-scale changes, revealing gross enlargements of ventricles and subarachnoid spaces, as well as white matter lesions, and atrophies that appear throughout the brain. Divergence from expected ageing reflected cardiovascular risk factors and accelerated ageing was more pronounced in the frontal lobe. Applying LRP, our study demonstrates how superior deep learning models detect brain-ageing in healthy and at-risk individuals throughout adulthood.
Source: NeuroImage
Institutions: Fraunhofer Institute for Telecommunications, Heinrich Hertz Institute, University Hospital Leipzig, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig University, University Medical Center Hamburg-Eppendorf, Google (United States), Korea University, Max Planck Institute for Informatics, Berlin Institute for the Foundations of Learning and Data, Technische Universität Berlin, Humboldt-Universität zu Berlin, Center for Stroke Research Berlin, Charité - Universitätsmedizin Berlin
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $3,450
Awards: G1067038318, G1372115314, G1386765427, SFB 1052: Mechanismen der Adipositas, G1583199633, G1808290430, G1999169356, G2033548659, G2186864416, G2232035061, MALT3 - MAschinelles Lernen-The Tricks of the Trade, G2804209865, G2953128317, CompLS - Runde 2 - Verbundprojekt: Patho234 - Multidimensionale Bildanalyse von reaktiven und neoplastischen Lymphknoten durch maschinelles Lernen - Teilprojekt D, G3340689995, G3763983592, G4165376421, G4253563803, G430232657, Verbundprojekt: Bernstein Fokus Neurotechnologie - Nichtinvasive Neurotechnologie für Mensch-Maschine Interaktion - Teilprojekte A1, A3, A4, B4, W3, Zentrum, G434337789, G4839762492, G5007941840, G5113337935, Verbundprojekt BIFOLD-BZML: Berlin Institute for the Foundations of Learning and Data, D-JPN Verbund: Adaptive Gehirn-Computer-Schnittstellen (BCI) in nichtstationären Umgebungen, G6185874248, G6548623498, Intelligent Total Body Scanner for Early Detection of Melanoma, G743358325, Verbundprojekt BIFOLD-BBDC: Berlin Institute for the Foundations of Learning and Data, EXC 2046: MATH+: Berlin Mathematics Research Center, G7802408855, G7923577054, Verbundprojekt: TraMeExCo - Transparenter Begleiter für medizinische Anwendung, G8162341385, G8721601906, G8986525620
OpenAlex ID: W4286687906