Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Efficient 4D fMRI ASD Classification using Spatial-Temporal-Omics-based Learning Framework

Ziqiao Weng, Weidong Cai, Bo Zhou

TL;DR

This paper tackles ASD classification from resting-state fMRI by addressing the trade-off between preserving spatial detail and maintaining computational efficiency. It introduces Spatial-Temporal-Omics (STO), a dual-branch framework that combines STVOmics (voxel-level 3D time-domain derivatives via a lightweight 3D CNN) and STROmics (upper-triangle functional connectivity via a DiagNet-inspired encoder–decoder) to form rich spatio-temporal representations. On the ABIDE benchmark, STO achieves state-of-the-art AUC across data proportions while reducing model size and inference time, with ablations showing that the two omics are complementary and that four 3D derivatives provide the strongest voxel-level features. The approach offers a scalable, integrable path for ASD studies and potentially broader neuroimaging tasks requiring efficient spatio-temporal feature learning, validated on multiple atlases and with strong cross-data-proportion performance.

Abstract

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder impacting social and behavioral development. Resting-state fMRI, a non-invasive tool for capturing brain connectivity patterns, aids in early ASD diagnosis and differentiation from typical controls (TC). However, previous methods, which rely on either mean time series or full 4D data, are limited by a lack of spatial information or by high computational costs. This underscores the need for an efficient solution that preserves both spatial and temporal information. In this paper, we propose a novel, simple, and efficient spatial-temporal-omics learning framework designed to efficiently extract spatio-temporal features from fMRI for ASD classification. Our approach addresses these limitations by utilizing 3D time-domain derivatives as the spatial-temporal inter-voxel omics, which preserve full spatial resolution while capturing diverse statistical characteristics of the time series at each voxel. Meanwhile, functional connectivity features serve as the spatial-temporal inter-regional omics, capturing correlations across brain regions. Extensive experiments and ablation studies on the ABIDE dataset demonstrate that our framework significantly outperforms previous methods while maintaining computational efficiency. We believe our research offers valuable insights that will inform and advance future ASD studies, particularly in the realm of spatial-temporal-omics-based learning.

Efficient 4D fMRI ASD Classification using Spatial-Temporal-Omics-based Learning Framework

TL;DR

This paper tackles ASD classification from resting-state fMRI by addressing the trade-off between preserving spatial detail and maintaining computational efficiency. It introduces Spatial-Temporal-Omics (STO), a dual-branch framework that combines STVOmics (voxel-level 3D time-domain derivatives via a lightweight 3D CNN) and STROmics (upper-triangle functional connectivity via a DiagNet-inspired encoder–decoder) to form rich spatio-temporal representations. On the ABIDE benchmark, STO achieves state-of-the-art AUC across data proportions while reducing model size and inference time, with ablations showing that the two omics are complementary and that four 3D derivatives provide the strongest voxel-level features. The approach offers a scalable, integrable path for ASD studies and potentially broader neuroimaging tasks requiring efficient spatio-temporal feature learning, validated on multiple atlases and with strong cross-data-proportion performance.

Abstract

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder impacting social and behavioral development. Resting-state fMRI, a non-invasive tool for capturing brain connectivity patterns, aids in early ASD diagnosis and differentiation from typical controls (TC). However, previous methods, which rely on either mean time series or full 4D data, are limited by a lack of spatial information or by high computational costs. This underscores the need for an efficient solution that preserves both spatial and temporal information. In this paper, we propose a novel, simple, and efficient spatial-temporal-omics learning framework designed to efficiently extract spatio-temporal features from fMRI for ASD classification. Our approach addresses these limitations by utilizing 3D time-domain derivatives as the spatial-temporal inter-voxel omics, which preserve full spatial resolution while capturing diverse statistical characteristics of the time series at each voxel. Meanwhile, functional connectivity features serve as the spatial-temporal inter-regional omics, capturing correlations across brain regions. Extensive experiments and ablation studies on the ABIDE dataset demonstrate that our framework significantly outperforms previous methods while maintaining computational efficiency. We believe our research offers valuable insights that will inform and advance future ASD studies, particularly in the realm of spatial-temporal-omics-based learning.

Paper Structure

This paper contains 10 sections, 1 figure, 2 tables.

Table of Contents

  1. Introduction
  2. Method
  3. Spatial-Temporal-Omics-based Framework
  4. Dataset and Preprocessing
  5. Evaluation Metrics and Baseline
  6. Implementation
  7. Experiments and Results
  8. Discussion and Conclusion
  9. Acknowledgments
  10. Compliance with Ethical Standards

Figures (1)

  • Figure 1: Overview of our STO framework, consisting of (a) spatial-temporal inter-regional omics (STROmics) and (b) spatial-temporal inter-voxel omics (STVOmics).