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

TL;DR

This paper introduces GmTE-Net, a novel few-shot learning framework using a teacher-student graph neural network architecture to predict multimodal brain connectivity development trajectories in infants, addressing data scarcity issues.

Contribution

It presents the first teacher-student GNN architecture for multi-trajectory brain graph prediction using few-shot learning, with a topology-aware distillation loss.

Findings

GmTE-Net outperforms benchmark methods in predicting brain graph trajectories.

The model effectively generalizes to multimodal and atypical brain development data.

Experimental results demonstrate significant performance improvements.

Abstract

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 learning-based 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 teacher-student 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 the predicted graph topology of the student network to be consistent with the teacher network. Experimental results demonstrate substantial performance gains over benchmark methods. Hence, our GmTE-Net can be leveraged to predict atypical brain connectivity trajectory evolution across various modalities. Our code is available at https: //github.com/basiralab/GmTE-Net.