Modeling Task fMRI Data via Deep Convolutional Autoencoder

Huang, Heng; Hu, Xintao; Makkie, Milad; Dong, Qinglin; Zhao, Yu; Han, Junwei; Liu, Tianming

doi:10.1007/978-3-319-59050-9_33

Cited by 32 publications

(37 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, the sparsity constraint has derived sparse representation and dictionary learning based methods [18], [27], [28], and the independence constraint has derived ICA based methods [12], [13]. Deep learning based methods [6], [29] have also been proposed for fMRI analysis. Power et al [6] proposed a restricted Boltzmann machine framework to investigate the complex structures in fMRI data.…”

Section: Introductionmentioning

confidence: 99%

“…Power et al [6] proposed a restricted Boltzmann machine framework to investigate the complex structures in fMRI data. In [29], a deep convolutional autoencoder (DCAE) model was proposed to identify the accurate brain activities in fMRI data. Although these methods greatly helped us improved the understanding of brain networks, possible limitations still existed.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Supervised Brain Network Learning Based on Deep Recurrent Neural Networks

et al. 2020

Self Cite

View full text Add to dashboard Cite

Task-based functional magnetic resonance imaging (tfMRI) is a widely used neuroimaging technique in exploring brain networks and functions associated with cognitive behaviors. Traditionally, the general linear model (GLM) is the most popular method in tfMRI data analysis due to its simpleness and robustness. This model-driven method adopts a canonical hemodynamic response function (HRF) and its various derivatives to construct regressors in the design matrix and estimate changes in the tfMRI data. However, a possible limitation of current model-driven methods is that the HRF is fixed and non-adaptive which may overlook other diverse and concurrent brain networks. In order to overcome these limitations, we proposed a novel hybrid framework, supervised brain network learning based on deep recurrent neural networks (SUDRNN), to reconstruct the diverse and concurrent functional brain networks. Specifically, this hybrid framework first takes advantage of the great capacity of deep recurrent neural networks (DRNN) in modeling sequential data to learn the diverse regressors from real tfMRI data. After that, it utilizes the effective supervised dictionary learning (SDL) method to reconstruct both the task-related functional brain networks and other latent brain networks simultaneously. Extensive experiment results on different tfMRI datasets from Human connectome project (HCP) demonstrated the superiority of the proposed framework. INDEX TERMS Task fMRI, brain networks, recurrent neural network, supervised dictionary learning.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Supervised Brain Network Learning Based on Deep Recurrent Neural Networks

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…By solving these sub-problems separately, the original reinforcement learning problems are finally solved [18,19]. It has also been proposed to use neural network structures to automate the hierarchical abstraction and learning of problems [20]. The literature [21] has proposed a reinforcement learning method that is based on path knowledge enlightenment, which uses the knowledge gained by online learning to reduce the blindness of the search, but it needs to set the robot action selection probability fictitiously.…”

Section: Introductionmentioning

confidence: 99%

Indoor robot path planning assisted by wireless network

Wang

Nie

Zhu

2019

J Wireless Com Network

View full text Add to dashboard Cite

Indoor robot global path planning needs to complete the motion between the starting point and the target point according to robot position command transmitted by the wireless network. Behavior dynamics and rolling windows in global path planning methods have limitations in their applications because the path may not be optimal, there could be a pseudo attractor or blind search in an environment with a large state space, there could be an environment where offline learning is not applicable to real-time changes, or there could be a need to set the probability of selecting the robot action. To solve these problems, we propose a behavior dynamics and rolling windows approach to a path planning which is based on online reinforcement learning. It applies Q learning to optimize the behavior dynamics model parameters to improve the performance, behavior dynamics guides the learning process of Q learning and improves learning efficiency, and each round of intensive learning action selection knowledge is gradually corrected as the Q table is updated. The learning process is optimized. The simulation results show that this method has achieved remarkable improvement in path planning. And, in the actual experiment, the robot obtains the target location information by wireless network, and plans an optimized and smooth global path online.

show abstract

“…The ground-truth annotation by experts is scarce for neuroimaging, especially functional MRI dataset. Towards producing `off-the-shelf' automatic disease diagnosis system, many publications, such as Huang et al (2017) and Makkie et al (2017) had addressed the need of ground-truth data.…”

Section: Introductionmentioning

confidence: 99%

Feature map size selection for fMRI classification on end-to-end deep convolutional neural networks

Suhaimi¹,

Htike²

2018

Int. j. adv. appl. sci.

View full text Add to dashboard Cite

The emergence of convolutional neural networks (CNN) in various fields has also paved numerous ways for advancement in the field of medical imaging. This paper focuses on functional magnetic resonance imaging (fMRI) in the field of neuroimaging. It has high temporal resolution and robust to control or non-control subjects. CNN analysis on structural magnetic resonance imaging (MRI) and fMRI datasets is compared to rule out one of the grey areas in building CNNs for medical imaging analysis. This study focuses on the feature map size selection on fMRI datasets with CNNs where the selected sizes are evaluated for their performances. Although few outstanding studies on fMRI have been published, the availability of diverse previous studies on MRI previous works impulses us to study to learn the pattern of feature map sizes for CNN configuration. Six configurations are analyzed with prominent public fMRI dataset, names as Human Connectome Project (HCP). This dataset is widely used for any type of fMRI classification. With three set of data divisions, the accuracy values for validation set of fMRI classification are assessed and discussed. Despite the fact that only one slice of every 118 subjects' temporal brain images is used in the study, the validation of classification for three training-excluded subjects known as validation set, has proven the need for feature map size selection. This paper emphasizes the indispensable step of selecting the feature map sizes when designing CNN for fMRI classification. In addition, we provide proofs that validation set should consist of distinct subjects for definite evaluation of any model performance.

show abstract

Modeling Task fMRI Data via Deep Convolutional Autoencoder

Cited by 32 publications

References 12 publications

Supervised Brain Network Learning Based on Deep Recurrent Neural Networks

Supervised Brain Network Learning Based on Deep Recurrent Neural Networks

Indoor robot path planning assisted by wireless network

Feature map size selection for fMRI classification on end-to-end deep convolutional neural networks

Contact Info

Product

Resources

About