Tractography-Based Score for Learning Effective Connectivity From Multimodal Imaging Data Using Dynamic Bayesian Networks
TL;DR: A novel anatomically informed (AI) score that evaluates fitness of a given connectivity structure to both DTI and fMRI data simultaneously is proposed that is employed in structure learning of DBN given the data.
Abstract: Objective: Effective connectivity (EC) is the methodology for determining functional-integration among the functionally active segregated regions of the brain. By definition [1] EC is “the causal influence exerted by one neuronal group on another” which is constrained by anatomical connectivity (AC) (axonal connections). AC is necessary for EC but does not fully determine it, because synaptic communication occurs dynamically in a context-dependent fashion. Although there is a vast emerging evidence of structure–function relationship using multimodal imaging studies, till date only a few studies have done joint modeling of the two modalities: functional MRI (fMRI) and diffusion tensor imaging (DTI). We aim to propose a unified probabilistic framework that combines information from both sources to learn EC using dynamic Bayesian networks (DBNs). Method: DBNs are probabilistic graphical temporal models that learn EC in an exploratory fashion. Specifically, we propose a novel anatomically informed (AI) score that evaluates fitness of a given connectivity structure to both DTI and fMRI data simultaneously. The AI score is employed in structure learning of DBN given the data. Results: Experiments with synthetic-data demonstrate the face validity of structure learning with our AI score over anatomically uninformed counterpart. Moreover, real-data results are cross-validated by performing classification-experiments. Conclusion: EC inferred on real fMRI-DTI datasets is found to be consistent with previous literature and show promising results in light of the AC present as compared to other classically used techniques such as Granger-causality. Significance: Multimodal analyses provide a more reliable basis for differentiating brain under abnormal/diseased conditions than the single modality analysis.
Citations
More filters
TL;DR: The experimental results indicate that the proposed variational inference method for the KDBN can provide reliable PIs for the generation flow of BFG and it exhibits shorter computing time than the stochastic based one.
Abstract: Prediction intervals ( PIs ) for industrial time series can provide useful guidance for workers. Given that the failure of industrial sensors may cause the missing point in inputs, the existing kernel dynamic Bayesian networks ( KDBN ) , serving as an effective method for PIs construction, suffer from high computational load using the stochastic algorithm for inference. This study proposes a variational inference method for the KDBN for the purpose of fast inference, which avoids the time-consuming stochastic sampling. The proposed algorithm contains two stages. The first stage involves the inference of the missing inputs by using a local linearization based variational inference, and based on the computed posterior distributions over the missing inputs the second stage sees a Gaussian approximation for probability over the nodes in future time slices. To verify the effectiveness of the proposed method, a synthetic dataset and a practical dataset of generation flow of blast furnace gas ( BFG ) are employed with different ratios of missing inputs. The experimental results indicate that the proposed method can provide reliable PIs for the generation flow of BFG and it exhibits shorter computing time than the stochastic based one.
9 citations
Cites methods from "Tractography-Based Score for Learni..."
...To verify the performance of the proposed method, this study compares the experimental results of several other methods of PIs construction, including the KDBN with weighted likelihood inference (KDBN-WL) [25], the Bayesian multiple layer perceptron (Bayesian MLP) [13], and the bootstrap-based echo state networks (Bootstrap ESN) [20]....
[...]
...Besides, in [20], an ensemble model containing a number of reservoir computing networks was employed by using the bootstrap techniques, which was applied to the prediction of practical industrial data....
[...]
...Similarly, to further verify the performance of the proposed method for the BFG data, this study compares the experimental results of several other methods of PIs construction, including the KDBN-WL [25], the Bayesian MLP [13], and the bootstrap ESN [20]....
[...]
TL;DR: A new method, called ACOEC-FD, to learn EC networks from fMRI and diffusion tensor imaging (DTI) using ant colony optimization (ACO), which achieves multi-modal imaging data integration by incorporating anatomical constraint information into the heuristic function of probabilistic transition rules.
Abstract: Identifying brain effective connectivity (EC) networks from neuroimaging data has become an effective tool that can evaluate normal brain functions and the injuries associated with neurodegenerative diseases. So far, there are many methods used to identify EC networks. However, most of the research currently focus on learning EC networks from single modal imaging data such as functional magnetic resonance imaging (fMRI) data. This paper proposes a new method, called ACOEC-FD, to learn EC networks from fMRI and diffusion tensor imaging (DTI) using ant colony optimization (ACO). First, ACOEC-FD uses DTI data to acquire some positively correlated relations among regions of interest (ROI), and takes them as anatomical constraint information to effectively restrict the search space of candidate arcs in an EC network. ACOEC-FD then achieves multi-modal imaging data integration by incorporating anatomical constraint information into the heuristic function of probabilistic transition rules to effectively encourage ants more likely to search for connections between structurally connected regions. Through simulation studies on generated datasets and real fMRI-DTI datasets, we demonstrate that the proposed approach results in improved inference results on EC compared to some methods that only used fMRI data.
7 citations
TL;DR: The presented methodology appears to be well-suited for fMRI, particularly given its lack of explicit dependence on temporal lag structure, and is readily applicable to whole-brain effective connectivity estimation.
Abstract: In functional magnetic resonance imaging (fMRI), functional connectivity is conventionally characterized by correlations between fMRI time series, which are intrinsically undirected measures of connectivity. Yet, some information about the directionality of network connections can nevertheless be extracted from the matrix of pairwise temporal correlations between all considered time series, when expressed in the frequency-domain as a cross-spectral density matrix. Using a sparsity prior, it then becomes possible to determine a unique directed network topology that best explains the observed undirected correlations, without having to rely on temporal precedence relationships that may not be valid in fMRI. Applying this method on simulated data with 100 nodes yielded excellent retrieval of the underlying directed networks under a wide variety of conditions. Importantly, the method did not depend on temporal precedence to establish directionality, thus reducing susceptibility to hemodynamic variability. The computational efficiency of the algorithm was sufficient to enable whole-brain estimations, thus circumventing the problem of missing nodes that otherwise occurs in partial-brain analyses. Applying the method to real resting-state fMRI data acquired with a high temporal resolution, the inferred networks showed good consistency with structural connectivity obtained from diffusion tractography in the same subjects. Interestingly, this agreement could also be seen when considering high-frequency rather than low-frequency connectivity (average correlation: r = 0.26 for f < 0.3 Hz, r = 0.43 for 0.3 < f < 5 Hz). Moreover, this concordance was significantly better (p < 0.05) than for networks obtained with conventional functional connectivity based on correlations (average correlation r = 0.18). The presented methodology thus appears to be well-suited for fMRI, particularly given its lack of explicit dependence on temporal lag structure, and is readily applicable to whole-brain effective connectivity estimation.
7 citations
Cites methods from "Tractography-Based Score for Learni..."
...SC is commonly used to constrain the estimation (Gilson et al., 2016; Crimi et al., 2017; Dang et al., 2018), or may be used independently to validate the estimated EC (Uddin et al., 2011; Bringmann et al., 2013)....
[...]
...SC is commonly used to constrain the estimation (Gilson et al., 2016; Crimi et al., 2017; Dang et al., 2018), or may be used independently to validate the estimated EC (Uddin et al....
[...]
TL;DR: Discrete Dynamic Bayesian Network (dDBN) is used in many challenging causal modelling applications, such as human brain connectivity, due to its multivariate, non-deterministic, and nonlinear capab...
Abstract: Discrete Dynamic Bayesian Network (dDBN) is used in many challenging causal modelling applications, such as human brain connectivity, due to its multivariate, non-deterministic, and nonlinear capab...
5 citations
TL;DR: Four machine learning tools are applied to address the challenge in the IFDI of cutting arms and the experimental results show that the support vector machines based on dynamic cuckoo outperform the other methods.
Abstract: Incipient fault detection and identification (IFDI) of cutting arms is a crucial guarantee for the smooth operation of a roadheader. However, the shortage of fault samples restricts the application of the fault diagnosis technique, and the data analysis tools should be optimized efficiently. In this study, four machine learning tools (the back-propagation neural network based on genetic algorithm optimization, the naive Bayes based on genetic algorithm optimization, the support vector machines based on particle swarm optimization, and the support vector machines based on dynamic cuckoo) are applied to address the challenge in the IFDI of cutting arms. The commonly measured current and vibration data cutting arms are used in the IFDI. The experimental results show that the support vector machines based on dynamic cuckoo outperform the other methods. Besides, the performance of the four methods under different operating conditions is compared. The fault cause of cutting arms of the roadheader is analyzed and the design improvement scheme for cutting arms is provided. This study provides a reference for improving the fault diagnosis of the roadheader.
3 citations
References
More filters
TL;DR: The resulting probabilistic atlases represent a comprehensive cohort of functionally-defined white matter regions that can be used in future brain imaging studies to ascribe DTI or other white matter changes to particular functional brain networks, and compliment resting state fMRI or other functional connectivity analyses.
Abstract: Diffusion tensor imaging (DTI) is a powerful MRI technique that can be used to estimate both the microstructural integrity and the trajectories of white matter pathways throughout the central nervous system. This fiber tracking (aka, “tractography”) approach is often carried out using anatomically-defined seed points to identify white matter tracts that pass through one or more structures, but can also be performed using functionally-defined regions of interest (ROIs) that have been determined using functional MRI (fMRI) or other methods. In this study, we performed fMRI-guided DTI tractography between all of the previously defined nodes within each of six common resting-state brain networks, including the: dorsal Default Mode Network (dDMN), ventral Default Mode Network (vDMN), left Executive Control Network (lECN), right Executive Control Network (rECN), anterior Salience Network (aSN), and posterior Salience Network (pSN). By normalizing the data from 32 healthy control subjects to a standard template – using high-dimensional, non-linear warping methods – we were able to create probabilistic white matter atlases for each tract in stereotaxic coordinates. By investigating all 198 ROI-to-ROI combinations within the aforementioned resting-state networks (for a total of 6336 independent DTI tractography analyses), the resulting probabilistic atlases represent a comprehensive cohort of functionally-defined white matter regions that can be used in future brain imaging studies to: 1) ascribe DTI or other white matter changes to particular functional brain networks, and 2) compliment resting state fMRI or other functional connectivity analyses.
40 citations
"Tractography-Based Score for Learni..." refers background or result in this paper
...We found several anatomical connections consistent with previous studies, such as strong bi-directional connections were found between: (1) mPFC (both anterior and ventral regions) and pCC/RSC, (2) left and right superior frontal cortex (SFC), (3) retrosplenial cortex (RSC) and parahippocampal (pHIP) gyrus (both left and right) [43]–[47]....
[...]
...Most of the relatively strong AC was related to the regions pCC/RSC, mPFC, and SFC, and some of which have also been reported or examined by other studies, such as left SFC- right SFC [43], mPFC-PCC/RSC [45]–[47], [58], and pHIP-RSC [59]....
[...]
...Both the left and right SFC received uni-directional influence from anterior mPFC; and the right SFC was bi-directionally connected with pCC and right LPC, although with low probability of connection....
[...]
...pCC/RSC, (2) left and right superior frontal cortex (SFC), (3) retrosplenial cortex (RSC) and parahippocampal (pHIP) gyrus (both left and right) [43]–[47]....
[...]
...of the relatively strong AC was related to the regions pCC/RSC, mPFC, and SFC, and some of which have also been reported or examined by other studies, such as left SFC- right SFC [43],...
[...]
TL;DR: An FC measure is introduced that rests upon assessments of functional coherence between regional brain activity identified from functional magnetic resonance imaging (fMRI) data and a unified Bayesian framework for analyzing FC utilizing the knowledge of associated structural connections is presented.
Abstract: Recent innovations in neuroimaging technology have provided opportunities for researchers to investigate connectivity in the human brain by examining the anatomical circuitry as well as functional relationships between brain regions. Existing statistical approaches for connectivity generally examine resting-state or task-related functional connectivity (FC) between brain regions or separately examine structural linkages. As a means to determine brain networks, we present a unified Bayesian framework for analyzing FC utilizing the knowledge of associated structural connections, which extends an approach by Patel et al.(2006a) that considers only functional data. We introduce an FC measure that rests upon assessments of functional coherence between regional brain activity identified from functional magnetic resonance imaging (fMRI) data. Our structural connectivity (SC) information is drawn from diffusion tensor imaging (DTI) data, which is used to quantify probabilities of SC between brain regions. We formulate a prior distribution for FC that depends upon the probability of SC between brain regions, with this dependence adhering to structural-functional links revealed by our fMRI and DTI data. We further characterize the functional hierarchy of functionally connected brain regions by defining an ascendancy measure that compares the marginal probabilities of elevated activity between regions. In addition, we describe topological properties of the network, which is composed of connected region pairs, by performing graph theoretic analyses. We demonstrate the use of our Bayesian model using fMRI and DTI data from a study of auditory processing. We further illustrate the advantages of our method by comparisons to methods that only incorporate functional information.
38 citations
"Tractography-Based Score for Learni..." refers background in this paper
...[30] have examined distribution of functional coherence at lower and higher levels of voxel-wise AC across 20910 region pairs based on 205 brain regions for a group-study....
[...]
..., overfitting) leading to poor generalizability of the model [30]....
[...]
TL;DR: The results indicated that the network efficiency of the DMN may play a role in the memory and anxiety, and the strong structural connectivity between multiple brain regions withinDMN may reflect that theDMN has certain structural basis.
Abstract: The default mode network (DMN) is one of the most widely studied resting state functional networks. The structural basis for the DMN is of particular interest and has been studied by several researchers using diffusion tensor imaging (DTI). Most of these previous studies focused on a few regions or white matter tracts of the DMN so that the global structural connectivity pattern and network properties of the DMN remain unclear. Moreover, evidences indicate that the DMN is involved in both memory and emotion, but how the DMN regulates memory and anxiety from the perspective of the whole DMN structural network remains unknown. We used multimodal neuroimaging methods to investigate the structural connectivity pattern of the DMN and the association of its network properties with memory and anxiety in 205 young healthy subjects with age ranging from 18 to 29 years old. The Group ICA method was used to extract the DMN component from functional magnetic resonance imaging (fMRI) data and a probabilistic fiber tractography technique based on DTI data was applied to construct the global structural connectivity pattern of the DMN. Then we used the graph theory method to analyze the DMN structural network and found that memory quotient (MQ) score was significantly positively correlated with the global and local efficiency of the DMN whereas anxiety was found to be negatively correlated with the efficiency. The strong structural connectivity between multiple brain regions within DMN may reflect that the DMN has certain structural basis. Meanwhile, the results we found that the network efficiency of the DMN were related to memory and anxiety measures, indicated that the DMN may play a role in the memory and anxiety.
33 citations
TL;DR: A rotation and translation invariant model is proposed that represents the spatial relationship between fiber tracts and anatomic and functional landmarks and can be used for detection and prediction of fiber tracts based on landmarks.
22 citations
"Tractography-Based Score for Learni..." refers background in this paper
...tions), namely: anatomically-weighted FC [7], track-weighted FC [8], landmark distance model [9], and probabilistic fusion...
[...]
TL;DR: A new BN method based on Gaussian assumptions, termed Gaussian DBN, is proposed to capture the temporal characteristics of connectivity with less associated loss of information.
Abstract: Two techniques based on the Bayesian network (BN), Gaussian Bayesian network and discrete dynamic Bayesian network (DBN), have recently been used to determine the effective connectivity from functional magnetic resonance imaging (fMRI) data in an exploratory manner and to provide a new method for exploring the interactions among brain regions. However, Gaussian BN ignores the temporal relationships of interactions among brain regions, while discrete DBN loses a great deal of information by discretizing the data. To overcome these limitations, the current study proposes a new BN method based on Gaussian assumptions, termed Gaussian DBN, to capture the temporal characteristics of connectivity with less associated loss of information. A set of synthetic data were generated to measure the robustness of this method to noise, and the results were compared with discrete DBN. In addition, real fMRI data obtained from twelve normal subjects in the resting state was used to further demonstrate and validate the effectiveness of the Gaussian DBN method. The results demonstrated that the Gaussian DBN was more robust than discrete DBN and an improvement over BN.
18 citations
"Tractography-Based Score for Learni..." refers methods in this paper
...used for fMRI data [18]–[20] and also for other domains [21], [22]....
[...]