Dataset Information

Modeling task-based fMRI data via deep belief network with neural architecture search.

ABSTRACT: It has been shown that deep neural networks are powerful and flexible models that can be applied on fMRI data with superb representation ability over traditional methods. However, a challenge of neural network architecture design has also attracted attention: due to the high dimension of fMRI volume images, the manual process of network model design is very time-consuming and not optimal. To tackle this problem, we proposed an unsupervised neural architecture search (NAS) framework on a deep belief network (DBN) that models volumetric fMRI data, named NAS-DBN. The NAS-DBN framework is based on Particle Swarm Optimization (PSO) where the swarms of neural architectures can evolve and converge to a feasible optimal solution. The experiments showed that the proposed NAS-DBN framework can quickly find a robust architecture of DBN, yielding a hierarchy organization of functional brain networks (FBNs) and temporal responses. Compared with 3 manually designed DBNs, the proposed NAS-DBN has the lowest testing loss of 0.0197, suggesting an overall performance improvement of up to 47.9 %. For each task, the NAS-DBN identified 260 FBNs, including task-specific FBNs and resting state networks (RSN), which have high overlap rates to general linear model (GLM) derived templates and independent component analysis (ICA) derived RSN templates. The average overlap rate of NAS-DBN to GLM on 20 task-specific FBNs is as high as 0.536, indicating a performance improvement of up to 63.9 % in respect of network modeling. Besides, we showed that the NAS-DBN can simultaneously generate temporal responses that resemble the task designs very well, and it was observed that widespread overlaps between FBNs from different layers of NAS-DBN model form a hierarchical organization of FBNs. Our NAS-DBN framework contributes an effective, unsupervised NAS method for modeling volumetric task fMRI data.

SUBMITTER: Qiang N

PROVIDER: S-EPMC7412935 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Feedforward deep neural networks (DNNs), artificial neural networks with multiple hidden layers, have recently demonstrated a record-breaking performance in multiple areas of applications in computer vision and speech processing. Following the success, DNNs have been applied to neuroimaging modalities including functional/structural magnetic resonance imaging (MRI) and positron-emission tomography data. However, no study has explicitly applied DNNs to 3D whole-brain fMRI volumes and thereby extracted hidden volumetric representations of fMRI that are discriminative for a task performed as the fMRI volume was acquired. Our study applied fully connected feedforward DNN to fMRI volumes collected in four sensorimotor tasks (i.e., left-hand clenching, right-hand clenching, auditory attention, and visual stimulus) undertaken by 12 healthy participants. Using a leave-one-subject-out cross-validation scheme, a restricted Boltzmann machine-based deep belief network was pretrained and used to initialize weights of the DNN. The pretrained DNN was fine-tuned while systematically controlling weight-sparsity levels across hidden layers. Optimal weight-sparsity levels were determined from a minimum validation error rate of fMRI volume classification. Minimum error rates (mean±standard deviation; %) of 6.9 (±3.8) were obtained from the three-layer DNN with the sparsest condition of weights across the three hidden layers. These error rates were even lower than the error rates from the single-layer network (9.4±4.6) and the two-layer network (7.4±4.1). The estimated DNN weights showed spatial patterns that are remarkably task-specific, particularly in the higher layers. The output values of the third hidden layer represented distinct patterns/codes of the 3D whole-brain fMRI volume and encoded the information of the tasks as evaluated from representational similarity analysis. Our reported findings show the ability of the DNN to classify a single fMRI volume based on the extraction of hidden representations of fMRI volumes associated with tasks across multiple hidden layers. Our study may be beneficial to the automatic classification/diagnosis of neuropsychiatric and neurological diseases and prediction of disease severity and recovery in (pre-) clinical settings using fMRI volumes without requiring an estimation of activation patterns or ad hoc statistical evaluation.

Project description:Perturbing a signaling system with a serial of single gene deletions and then observing corresponding expression changes in model organisms, such as yeast, is an important and widely used experimental technique for studying signaling pathways. People have developed different computational methods to analyze the perturbation data from gene deletion experiments for exploring the signaling pathways. The most popular methods/techniques include K-means clustering and hierarchical clustering techniques, or combining the expression data with knowledge, such as protein-protein interactions (PPIs) or gene ontology (GO), to search for new pathways. However, these methods neither consider nor fully utilize the intrinsic relation between the perturbation of a pathway and expression changes of genes regulated by the pathway, which served as the main motivation for developing a new computational method in this study. In our new model, we first find gene transcriptomic modules such that genes in each module are highly likely to be regulated by a common signal. We then use the expression status of those modules as readouts of pathway perturbations to search for up-stream pathways. Systematic evaluation, such as through gene ontology enrichment analysis, has provided evidence that genes in each transcriptomic module are highly likely to be regulated by a common signal. The PPI density analysis and literature search revealed that our new perturbation modules are functionally coherent. For example, the literature search revealed that 9 genes in one of our perturbation module are related to cell cycle and all 10 genes in another perturbation module are related by DNA damage, with much evidence from the literature coming from in vitro or/and in vivo verifications. Hence, utilizing the intrinsic relation between the perturbation of a pathway and the expression changes of genes regulated by the pathway is a useful method of searching for signaling pathways using genetic perturbation data. This model would also be suitable for analyzing drug experiment data, such as the CMap data, for finding drugs that perturb the same pathways.

Dataset Information

Modeling task-based fMRI data via deep belief network with neural architecture search.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets