Format

Send to

Choose Destination
Front Neurosci. 2014 Aug 20;8:229. doi: 10.3389/fnins.2014.00229. eCollection 2014.

Deep learning for neuroimaging: a validation study.

Author information

1
The Mind Research Network Albuquerque, NM, USA.
2
Department of Computer Science, University of New Mexico Albuquerque, NM, USA.
3
Department of Computer Science, University of Toronto Toronto, ON, Canada.
4
The Mind Research Network Albuquerque, NM, USA ; Department of Biological and Medical Psychology, University of Bergen Bergen, Norway.
5
Advanced Biomedical Informatics Group, LLC, University of Iowa Iowa City, IA, USA.
6
Department of Psychiatry, Carver College of Medicine, University of Iowa Iowa City, IA, USA ; Department of Biostatistics, College of Public Health, University of Iowa Iowa City, IA, USA.
7
Department of Psychiatry, Carver College of Medicine, University of Iowa Iowa City, IA, USA ; Department of Biomedical Engineering, College of Engineering, University of Iowa Iowa City, IA, USA.
8
Department of Psychiatry, Carver College of Medicine, University of Iowa Iowa City, IA, USA ; Department of Psychology, Neuroscience Institute, University of Iowa Iowa City, IA, USA ; Department of Neurology, Carver College of Medicine, University of Iowa Iowa City, IA, USA.
9
Department of Psychology, Neuroscience Institute, Georgia State University Atlanta, GA, USA.
10
The Mind Research Network Albuquerque, NM, USA ; Department of Computer Science, University of New Mexico Albuquerque, NM, USA ; Department of Electrical and Computer Engineering, University of New Mexico Albuquerque, NM, USA.

Abstract

Deep learning methods have recently made notable advances in the tasks of classification and representation learning. These tasks are important for brain imaging and neuroscience discovery, making the methods attractive for porting to a neuroimager's toolbox. Success of these methods is, in part, explained by the flexibility of deep learning models. However, this flexibility makes the process of porting to new areas a difficult parameter optimization problem. In this work we demonstrate our results (and feasible parameter ranges) in application of deep learning methods to structural and functional brain imaging data. These methods include deep belief networks and their building block the restricted Boltzmann machine. We also describe a novel constraint-based approach to visualizing high dimensional data. We use it to analyze the effect of parameter choices on data transformations. Our results show that deep learning methods are able to learn physiologically important representations and detect latent relations in neuroimaging data.

KEYWORDS:

MRI; classification; fMRI; intrinsic networks; unsupervised learning

Supplemental Content

Full text links

Icon for Frontiers Media SA Icon for PubMed Central
Loading ...
Support Center