Deep Learning for neuroimaging data is a promising but challenging direction. The high dimensionality of 3D MRI scans makes this endeavor compute and data-intensive. Most conventional 3D neuroimaging methods use 3D-CNN-based architectures with a large number of parameters and require more time and data to train. Recently, 2D-slice-based models have received increasing attention as they have fewer parameters and may require fewer samples to achieve comparable performance. In this paper, we propose a new architecture for BrainAGE prediction. The proposed architecture works by encoding each 2D slice in an MRI with a deep 2D-CNN model. Next, it combines the information from these 2D-slice encodings using set networks or permutation invariant layers. Experiments on the BrainAGE prediction problem, using the UK Biobank dataset, showed that the model with the permutation invariant layers trains faster and provides better predictions compared to other state-of-the-art approaches.
BrainAge (a subject’s apparent age predicted from neuroimaging data) is an important biomarker of brain aging. The deviation of BrainAge from true age has been associated with psychiatric and neurological disease, and has proven effective in predicting conversion from mild cognitive impairment (MCI) to dementia. Conventionally, 3D convolutional neural networks and their variants are used for brain age prediction. However, these networks have a larger number of parameters and take longer to train than their 2D counterparts. Here we propose a 2D slice-based recurrent neural network model, which takes in an ordered sequence of sagittal slices as input to predict the brain age. The model consists of two components: a 2D convolutional neural network (CNN), which encodes the relevant features from the slices, and a recurrent neural network (RNN) that learns the relationship between slices. We compare our method to other recently proposed methods, including 3D deep convolutional regression networks, information theoretic models, and bag-of-features (BoF) models (such as BagNet) - where the classification is based on the occurrences of local features, without taking into consideration their global spatial ordering. In our experiments, our proposed model performs comparably to, or better than, the current state of the art models, with nearly half the number of parameters and a lower convergence time.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.