skip to main content

Attention:

The NSF Public Access Repository (NSF-PAR) system and access will be unavailable from 5:00 PM ET until 11:00 PM ET on Friday, June 21 due to maintenance. We apologize for the inconvenience.


Title: Dual-Attention Recurrent Networks for Affine Registration of Neuroimaging Data
Neuroimaging data typically undergoes several preprocessing steps before further analysis and mining can be done. Affine image registration is one of the important tasks during preprocessing. Recently, several image registration methods which are based on Convolutional Neural Networks have been proposed. However, due to the high computational and memory requirements of CNNs, these methods cannot be used in real-time for large neuroimaging data like fMRI. In this paper, we propose a Dual-Attention Recurrent Network (DRN) which uses a hard attention mechanism to allow the model to focus on small, but task-relevant, parts of the input image – thus reducing computational and memory costs. Furthermore, DRN naturally supports inhomogeneity between the raw input image (e.g., functional MRI) and the image we want to align it to (e.g., anatomical MRI) so it can be applied to harder registration tasks such as fMRI coregistration and normalization. Extensive experiments on two different datasets demonstrate that DRN significantly reduces the computational and memory costs compared with other neural network-based methods without sacrificing the quality of image registration  more » « less
Award ID(s):
1718310
NSF-PAR ID:
10215776
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
Proceedings of the 2020 SIAM International Conference on Data Mining, SDM 2020
Page Range / eLocation ID:
379-387
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Cardiac Cine Magnetic Resonance (CMR) Imaging has made a significant paradigm shift in medical imaging technology, thanks to its capability of acquiring high spatial and temporal resolution images of different structures within the heart that can be used for reconstructing patient-specific ventricular computational models. In this work, we describe the development of dynamic patient-specific right ventricle (RV) models associated with normal subjects and abnormal RV patients to be subsequently used to assess RV function based on motion and kinematic analysis. We first constructed static RV models using segmentation masks of cardiac chambers generated from our accurate, memory-efficient deep neural architecture - CondenseUNet - featuring both a learned group structure and a regularized weight-pruner to estimate the motion of the right ventricle. In our study, we use a deep learning-based deformable network that takes 3D input volumes and outputs a motion field which is then used to generate isosurface meshes of the cardiac geometry at all cardiac frames by propagating the end-diastole (ED) isosurface mesh using the reconstructed motion field. The proposed model was trained and tested on the Automated Cardiac Diagnosis Challenge (ACDC) dataset featuring 150 cine cardiac MRI patient datasets. The isosurface meshes generated using the proposed pipeline were compared to those obtained using motion propagation via traditional non-rigid registration based on several performance metrics, including Dice score and mean absolute distance (MAD). 
    more » « less
  2. It is becoming increasingly common to collect multiple related neuroimaging datasets either from different modalities or from different tasks and conditions. In addition, we have non-imaging data such as cognitive or behavioral variables, and it is through the association of these two sets of data—neuroimaging and non-neuroimaging—that we can understand and explain the evolution of neural and cognitive processes, and predict outcomes for intervention and treatment. Multiple methods for the joint analysis or fusion of multiple neuroimaging datasets or modalities exist; however, methods for the joint analysis of imaging and non-imaging data are still in their infancy. Current approaches for identifying brain networks related to cognitive assessments are still largely based on simple one-to-one correlation analyses and do not use the cross information available across multiple datasets. This work proposes two approaches based on independent vector analysis (IVA) to jointly analyze the imaging datasets and behavioral variables such that multivariate relationships across imaging data and behavioral features can be identified. The simulation results show that our proposed methods provide better accuracy in identifying associations across imaging and behavioral components than current approaches. With functional magnetic resonance imaging (fMRI) task data collected from 138 healthy controls and 109 patients with schizophrenia, results reveal that the central executive network (CEN) estimated in multiple datasets shows a strong correlation with the behavioral variable that measures working memory, a result that is not identified by traditional approaches. Most of the identified fMRI maps also show significant differences in activations across healthy controls and patients potentially providing a useful signature of mental disorders. 
    more » « less
  3. Everyday experience requires processing external signals from the world around us and internal information retrieved from memory. To do both, the brain must fluctuate between states that are optimized for external versus internal attention. Here, we focus on the hippocampus as a region that may serve at the interface between these forms of attention and ask how it switches between prioritizing sensory signals from the external world versus internal signals related to memories and thoughts. Pharmacological, computational, and animal studies have identified input from the cholinergic basal forebrain as important for biasing the hippocampus toward processing external information, whereas complementary research suggests the dorsal attention network (DAN) may aid in allocating attentional resources toward accessing internal information. We therefore tested the hypothesis that the basal forebrain and DAN drive the hippocampus toward external and internal attention, respectively. We used data from 29 human participants (17 female) who completed two attention tasks during fMRI. One task (memory-guided) required proportionally more internal attention, and proportionally less external attention, than the other (explicitly instructed). We discovered that background functional connectivity between the basal forebrain and hippocampus was stronger during the explicitly instructed versus memory-guided task. In contrast, DAN–hippocampus background connectivity was stronger during the memory-guided versus explicitly instructed task. Finally, the strength of DAN–hippocampus background connectivity was correlated with performance on the memory-guided but not explicitly instructed task. Together, these results provide evidence that the basal forebrain and DAN may modulate the hippocampus to switch between external and internal attention.

    SIGNIFICANCE STATEMENTHow does the brain balance the need to pay attention to internal thoughts and external sensations? We focused on the human hippocampus, a region that may serve at the interface between internal and external attention, and asked how its functional connectivity varies based on attentional states. The hippocampus was more strongly coupled with the cholinergic basal forebrain when attentional states were guided by the external world rather than retrieved memories. This pattern flipped for functional connectivity between the hippocampus and dorsal attention network, which was higher for attention tasks that were guided by memory rather than external cues. Together, these findings show that distinct networks in the brain may modulate the hippocampus to switch between external and internal attention.

     
    more » « less
  4. Brain large-scale dynamics is constrained by the heterogeneity of intrinsic anatomical substrate. Little is known how the spatiotemporal dynamics adapt for the heterogeneous structural connectivity (SC). Modern neuroimaging modalities make it possible to study the intrinsic brain activity at the scale of seconds to minutes. Diffusion magnetic resonance imaging (dMRI) and functional MRI reveals the large-scale SC across different brain regions. Electrophysiological methods (i.e. MEG/EEG) provide direct measures of neural activity and exhibits complex neurobiological temporal dynamics which could not be solved by fMRI. However, most of existing multimodal analytical methods collapse the brain measurements either in space or time domain and fail to capture the spatio-temporal circuit dynamics. In this paper, we propose a novel spatio-temporal graph Transformer model to integrate the structural and functional connectivity in both spatial and temporal domain. The proposed method learns the heterogeneous node and graph representation via contrastive learning and multi-head attention based graph Transformer using multimodal brain data (i.e. fMRI, MRI, MEG and behavior performance). The proposed contrastive graph Transformer representation model incorporates the heterogeneity map constrained by T1-to-T2-weighted (T1w/T2w) to improve the model fit to structurefunction interactions. The experimental results with multimodal resting state brain measurements demonstrate the proposed method could highlight the local properties of large-scale brain spatio-temporal dynamics and capture the dependence strength between functional connectivity and behaviors. In summary, the proposed method enables the complex brain dynamics explanation for different modal variants. 
    more » « less
  5. Purpose

    The ability to register image data to a common coordinate system is a critical feature of virtually all imaging studies. However, in spite of the abundance of literature on the subject and the existence of several variants of registration algorithms, their practical utility remains problematic, as commonly acknowledged even by developers of these methods.

    Methods

    A new registration method is presented that utilizes a Hamiltonian formalism and constructs registration as a sequence of symplectomorphic maps in conjunction with a novel phase space regularization. For validation of the framework a panel of deformations expressed in analytical form is developed that includes deformations based on known physical processes in MRI and reproduces various distortions and artifacts typically present in images collected using these different MRI modalities.

    Results

    The method is demonstrated on the three different magnetic resonance imaging (MRI) modalities by mapping between high resolution anatomical (HRA) volumes, medium resolution diffusion weighted MRI (DW‐MRI) and HRA volumes, and low resolution functional MRI (fMRI) and HRA volumes.

    Conclusions

    The method has shown an excellent performance and the panel of deformations was instrumental to quantify its repeatability and reproducibility in comparison to several available alternative approaches.

     
    more » « less