Graph Neural Networks in Network Neuroscience

Bessadok, Alaa; Mahjoub, Mohamed Ali; Rekik, Islem

doi:10.48550/arxiv.2106.03535

Cited by 20 publications

(28 citation statements)

References 81 publications

(142 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In that way, we unprecedentedly enforce our teacher network to superresolve brain graphs while jointly solving two critical problems [2]: the inter-domain alignment and the topological property preservation of brain graphs (Fig. 1-B).…”

Section: Methodsmentioning

confidence: 99%

“…In addition, [18] proposed an autoencoder-based architecture for predicting 7T T1weighted MRI from its 3T counterparts by leveraging both spatial and wavelet domains. While several multi-resolution image synthesis works have been proposed, the road to superresolving brain graphs (i.e., connectomes) is still less traveled [2]. In a brain graph, nodes denote the region of interest (i.e., ROI) and edges denote the connectivity between pairs of ROIs.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Inter-Domain Alignment for Predicting High-Resolution Brain Networks Using Teacher-Student Learning

Demir¹,

Bessadok²,

Rekik³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Accurate and automated super-resolution image synthesis is highly desired since it has the great potential to circumvent the need for acquiring high-cost medical scans and a time-consuming preprocessing pipeline of neuroimaging data. However, existing deep learning frameworks are solely designed to predict high-resolution (HR) image from a low-resolution (LR) one, which limits their generalization ability to brain graphs (i.e., connectomes). A small body of works has focused on superresolving brain graphs where the goal is to predict a HR graph from a single LR graph. Although promising, existing works mainly focus on superresolving graphs belonging to the same domain (e.g., functional), overlooking the domain fracture existing between multimodal brain data distributions (e.g., morphological and structural). To this aim, we propose a novel inter-domain adaptation framework namely, Learn to SuperResolve Brain Graphs with Knowledge Distillation Network (L2S-KDnet), which adopts a teacher-student paradigm to superresolve brain graphs. Our teacher network is a graph encoder-decoder that firstly learns the LR brain graph embeddings, and secondly learns how to align the resulting latent representations to the HR ground truth data distribution using an adversarial regularization. Ultimately, it decodes the HR graphs from the aligned embeddings. Next, our student network learns the knowledge of the aligned brain graphs as well as the topological structure of the predicted HR graphs transferred from the teacher. We further leverage the decoder of the teacher to optimize the student network. In such a way, we are not only bringing the learned embeddings from both networks closer to each other but also their predicted HR graphs. L2S-KDnet presents the first TS architecture tailored for brain graph super-resolution synthesis that is based on inter-domain alignment. Our experimental results demonstrate substantial performance gains over benchmark methods.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Inter-Domain Alignment for Predicting High-Resolution Brain Networks Using Teacher-Student Learning

Demir¹,

Bessadok²,

Rekik³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, this problem is challenging due to the scarcity of longitudinal datasets. Thus, there is an increasing need to foresee longitudinal missing baby connectomic data given what already exists [4]. Network literature [5] shows that a typical brain connectome is defined as a graph of brain connectivities where nodes denote regions of interest (ROIs) and edges encode the pairwise connectivity strengths between ROI pairs.…”

Section: Introductionmentioning

confidence: 99%

A Few-shot Learning Graph Multi-Trajectory Evolution Network for Forecasting Multimodal Baby Connectivity Development from a Baseline Timepoint

Bessadok¹,

Nebli²,

Mahjoub³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Charting the baby connectome evolution trajectory during the first year after birth plays a vital role in understanding dynamic connectivity development of baby brains. Such analysis requires acquisition of longitudinal connectomic datasets. However, both neonatal and postnatal scans are rarely acquired due to various difficulties. A small body of works has focused on predicting baby brain evolution trajectory from a neonatal brain connectome derived from a single modality. Although promising, large training datasets are essential to boost model learning and to generalize to a multi-trajectory prediction from different modalities (i.e., functional and morphological connectomes). Here, we unprecedentedly explore the question: "Can we design a few-shot learningbased framework for predicting brain graph trajectories across different modalities?" To this aim, we propose a Graph Multi-Trajectory Evolution Network (GmTE-Net), which adopts a teacher-student paradigm where the teacher network learns on pure neonatal brain graphs and the student network learns on simulated brain graphs given a set of different timepoints. To the best of our knowledge, this is the first teacherstudent architecture tailored for brain graph multi-trajectory growth prediction that is based on few-shot learning and generalized to graph neural networks (GNNs). To boost the performance of the student network, we introduce a local topology-aware distillation loss that forces

show abstract

“…Nonetheless, SM-Net Fusion can only handle a population of graphs derived from a single view or neuroimaging modality, failing to generalize to multigraphs. To better model the complex interactions at the individual brain multigraph level (i.e., between ROIs) and at the population level (i.e., between multigraphs), one can leverage the power of the emerging graph neural networks (GNNs) [12,13,14] in learning end-to-end mapping for our CBT integration and evolution forecasting tasks. Very recently, [9] proposed deep graph normalizer (DGN), the state-of-the-art method to integrate a population of brain multigraphs into a representative CBT using GNNs.…”

Section: Introductionmentioning

confidence: 99%

Recurrent Multigraph Integrator Network for Predicting the Evolution of Population-Driven Brain Connectivity Templates

Demirbilek¹,

Rekik²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Learning how to estimate a connectional brain template (CBT) from a population of brain multigraphs, where each graph (e.g., functional) quantifies a particular relationship between pairs of brain regions of interest (ROIs), allows to pin down the unique connectivity patterns shared across individuals. Specifically, a CBT is viewed as an integral representation of a set of highly heterogeneous graphs and ideally meeting the centeredness (i.e., minimum distance to all graphs in the population) and discriminativeness (i.e., distinguishes the healthy from the disordered population) criteria. So far, existing works have been limited to only integrating and fusing a population of brain multigraphs acquired at a single timepoint. In this paper, we unprecedentedly tackle the question: "Given a baseline multigraph population, can we learn how to integrate and forecast its CBT representations at followup timepoints?" Addressing such question is of paramount in predicting common alternations across healthy and disordered populations. To fill this gap, we propose Recurrent Multigraph Integrator Network (ReMI-Net), the first graph recurrent neural network which infers the baseline CBT of an input population t1 and predicts its longitudinal evolution over time (ti > t1). Our ReMI-Net is composed of recurrent neural blocks with graph convolutional layers using a cross-node message passing to first learn hidden-states embeddings of each CBT node (i.e., brain region of interest) and then predict its evolution at the consecutive timepoint. Moreover, we design a novel time-dependent loss to regularize the CBT evolution trajectory over time and further introduce a cyclic recursion and learnable normalization layer to generate well-centered CBTs from time-dependent hidden-state embeddings. Finally, we derive the CBT adjacency matrix from the learned hidden state graph representation. ReMI-Net significantly outperformed benchmark methods in both centeredness and discriminative connectional biomarker discovery criteria in demented patients. Our ReMI-Net GitHub code is available at https://github.com/basiralab/ReMI-Net.

show abstract

Graph Neural Networks in Network Neuroscience

Cited by 20 publications

References 81 publications

Inter-Domain Alignment for Predicting High-Resolution Brain Networks Using Teacher-Student Learning

Inter-Domain Alignment for Predicting High-Resolution Brain Networks Using Teacher-Student Learning

A Few-shot Learning Graph Multi-Trajectory Evolution Network for Forecasting Multimodal Baby Connectivity Development from a Baseline Timepoint

Recurrent Multigraph Integrator Network for Predicting the Evolution of Population-Driven Brain Connectivity Templates

Contact Info

Product

Resources

About