Unraveling the Autism spectrum heterogeneity: Insights from ABIDE I Database using data/model-driven permutation testing approaches

TL;DR

This study interrogates whether structural MRI reveals consistent brain-region differences between individuals with Autism Spectrum Condition and neurotypical controls within the multicenter ABIDE I database. Using permutation testing with two mapping frameworks—Statistical Agnostic Mapping (SAM) and Statistical Parametric Mapping (SPM)—the authors assess differences by condition and by acquisition site, while examining the impact of outliers. Across analyses, they observe no statistically robust region-level differences, largely attributing variance to center effects, small sample sizes, and data heterogeneity; SAM results align more closely with prior literature than SPM, which is notably conservative. The findings suggest that ABIDE I’s structural MRI data, given its heterogeneity, are unlikely to yield reliable structural biomarkers for ASC without harmonized, larger-scale data collection. The work highlights the need for standardized protocols and multi-center coordination to improve biomarker discovery in heterogeneous disorders such as autism.

Abstract

Autism Spectrum Condition (ASC) is a neurodevelopmental condition characterized by impairments in communication, social interaction and restricted or repetitive behaviors. Extensive research has been conducted to identify distinctions between individuals with ASC and neurotypical individuals. However, limited attention has been given to comprehensively evaluating how variations in image acquisition protocols across different centers influence these observed differences. This analysis focuses on structural magnetic resonance imaging (sMRI) data from the Autism Brain Imaging Data Exchange I (ABIDE I) database, evaluating subjects' condition and individual centers to identify disparities between ASC and control groups. Statistical analysis, employing permutation tests, utilizes two distinct statistical mapping methods: Statistical Agnostic Mapping (SAM) and Statistical Parametric Mapping (SPM). Results reveal the absence of statistically significant differences in any brain region, attributed to factors such as limited sample sizes within certain centers, noise effects and the problem of multicentrism in a heterogeneous condition such as autism. This study indicates limitations in using the ABIDE I database to detect structural differences in the brain between neurotypical individuals and those diagnosed with ASC. Furthermore, results from the SAM mapping method show greater consistency with existing literature.

Figures (16)

• Figure 1: The block diagram illustrates the MRI analysis process using SPM and its three main steps. The diagram specifically focuses on the comparison between HC and individuals with ASC. For the comparison between healthy controls themselves (HC vs HC), the group division involves creating two separate HC groups while maintaining the remaining stages of the analysis process.
• Figure 2: The block diagram illustrates the MRI analysis process incorporating SAM and the feature extraction and selection stage. Specifically, the diagram represents the comparison between HC and individuals with ASC. As for the comparison between HC themselves (HC vs HC), the group division will involve creating two separate HC groups, while retaining the remaining stages of the analysis process.
• Figure 3: Block diagram about structure of the Autism Brain Imaging Data Exchange I (ABIDE I) dataset analysis.
• Figure 4: Estimated probability of detection through permutation testing using the SAM mapping method. Each colormap rectangle represents one of the 116 brain regions defined in Table 2 of sun_mining_2009. The color map depicts the probability of detection of significant differences observed in the regions during the permutation test.
• Figure 5: Brain mosaics from different centers to assess the presence of defective images. Specifically, the mosaics of three centers with distinct issues in their images are shown:(a) CALTECH's cerebral mosaic, (b) YALE's cerebral mosaic and (c) UM_1's cerebral mosaic.