Automated segmentation of cerebellar morphology in autism spectrum disorder and multi-center validation based on the ABIDE database

doi:10.3877/cma.j.issn.2095-655X.2026.01.003

Abstract

Abstract:

Objective

To examine cerebellar morphological differences between individuals with autism spectrum disorder (ASD) and typically developing (TD) controls using multicenter structural MRI data and validate the cross-site robustness of the results.

Methods

This study included 2 147 participants (aged 6-35 years; 1 030 cases of ASD, 1 117 cases of TD) from 28 sites. Intracranial volume effects were regressed out to obtain residual cerebellar volumes. A linear mixed-effects (LME) model was applied with group, sex, and age as fixed effects and scanning sites as random effects. Estimated marginal means (EMMeans) were derived from the LME model for group comparisons. Cross-site consistency was further evaluated using random-effects meta-analysis of effect sizes (Hedges′g).

Results

After controlling for age, sex, and site, the LME model revealed that the residual volume of the left cerebellum in the TD group was significantly larger than that in the ASD group (β=0.515, P=0.0345), while no significant intergroup difference was observed in the right cerebellum (β=0.368, P=0.1280). Age was negatively associated with bilateral cerebellar volumes (β=-0.094, -0.091, P<0.01), and females exhibited smaller volumes than males (β=-0.959, -0.925, P<0.01). The estimated marginal mean comparison further confirmed that the volume of the left cerebellum was significantly reduced in the ASD group compared to the TD group (EMMeans=-0.515, 95%CI: -0.992 to -0.037, P=0.0346), whereas no significant difference in the right cerebellum (EMMeans=-0.368, 95%CI: -0.842 to 0.106, P=0.1277). Meta-analysis demonstrated minimal cross-site heterogeneity (I²=0%) for both hemispheres, with pooled effect sizes of -0.06 (95%CI: -0.14 to 0.03) for the left cerebellum and -0.03 (95%CI: -0.12 to 0.05) for the right cerebellum, indicating the left cerebellar differences have good consistency.

Conclusions

Individuals with ASD exhibit specific reductions in the structure of the left cerebellum, a finding that is robust across multi-center data. These morphdogical changes in the left cerebellum may provide imaging evidence contributing to the understanding of neuropathological mechanisms in ASD.

Key words: Autism spectrum disorder, Cerebellum, Brain volume, Structural magnetic resonance imaging, Multicenter study

Jingjing Lin, Jiayu Han, Guanxun Cheng, Shifeng Li. Automated segmentation of cerebellar morphology in autism spectrum disorder and multi-center validation based on the ABIDE database[J]. Chinese Journal of Diagnostics(Electronic Edition), 2026, 14(01): 24-30.

/ / Recommend

Add to citation manager EndNote|Ris|BibTeX

URL: https://zhzdxdzzz.cma-cmc.com.cn/EN/10.3877/cma.j.issn.2095-655X.2026.01.003

https://zhzdxdzzz.cma-cmc.com.cn/EN/Y2026/V14/I01/24

Figures/Tables 6

References 24

[1]	American Psychiatric Association.Diagnostic and statistical manual of mental disorders[M].5th ed.Arlington，VA：American Psychiatric Publishing，2013.
[2]	Abrahams BS，Geschwind DH.Advances in autism genetics：on the threshold of a new neurobiology[J].Nat Rev Genet，2008，9(5)：341-355.DOI：10.1038/nrg2346.
[3]	Global Burden of Disease Study 2021 Autism Spectrum Collaborators.The global epidemiology and health burden of the autism spectrum：findings from the Global Burden of Disease Study 2021[J].Lancet Psychiatry，2025，12(2)：111-121.DOI：10.1016/S2215-0366(24)00363-8.
[4]	Liloia D，Zamfira DA，Tanaka M，et al.Disentangling the role of gray matter volume and concentration in autism spectrum disorder：a meta-analytic investigation of 25 years of voxel-based morphometry research[J].Neurosci Biobehav Rev，2024(164)：105791.DOI：10.1016/j.neubiorev.2024.105791.
[5]	Duan K，Eyler L，Pierce K，et al.Differences in regional brain structure in toddlers with autism are related to future language outcomes[J].Nat Commun，2024，15(1)：5075.DOI：10.1038/s41467-024-48952-4.
[6]	Ecker C，Rocha-Rego V，Johnston P，et al.Investigating the predictive value of whole-brain structural MR scans in autism：a pattern classification approach[J].Neuroimage，2010，49(1)：44-56.DOI：10.1016/j.neuroimage.2009.08.024.
[7]	Stoodley CJ.Distinct regions of the cerebellum show gray matter decreases in autism，ADHD，and developmental dyslexia[J].Front Syst Neurosci，2014(8)：92.DOI：10.3389/fnsys.2014.00092.
[8]	d′Oleire UquillasF，Sefik E，Li B，et al.Multimodal evidence for cerebellar influence on cortical development in autism：structural growth amidst functional disruption[J].Mol Psychiatry，2025，30(4)：1558-1572.DOI：10.1038/s41380-024-02769-1.
[9]	Romero JE，Coupé P，Giraud R，et al.CERES：a new cerebellum lobule segmentation method[J].Neuroimage，2017(147)：916-924.DOI：10.1016/j.neuroimage.2016.11.003.
[10]	Button KS，Ioannidis JP，Mokrysz C，et al.Power failure：why small sample size undermines the reliability of neuroscience[J].Nat Rev Neurosci，2013，14(5)：365-376.DOI：10.1038/nrn3475.
[11]	Varoquaux G．Cross-validation failure：small sample sizes lead to large error bars[J].Neuroimage，2018，180(Pt A)：68-77.DOI：10.1016/j.neuroimage.2017.06.061.
[12]	Takao H，Hayashi N，Ohtomo K．Effect of scanner in longitudinal studies of brain volume changes[J].J Magn Reson Imaging，2011，34(2)：438-444.DOI：10.1002/jmri.22636.
[13]	Glover GH，Mueller BA，Turner JA, et al. Function biomedical informatics research network recommendations for prospective multicenter functional MRI studies[J].J Magn Reson Imaging，2012，36(1)：39-54.DOI：10.1002/jmri.23572.
[14]	Hu M，Lou Y，Zhu C，et al.Evaluating the impact of intracranial volume correction approaches on the quantification of intracranial structures in MRI：a systematic analysis[J].J Magn Reson Imaging，2024，59(6)：2164-2177..DOI：10.1002/jmri.28974.
[15]	Marek S，Tervo-Clemmens B，Calabro FJ，et al.Reproducible brain-wide association studies require thousands of individuals[J].Nature，2022，603(7902)：654-660.DOI：10.1038/s41586-022-04492-9.
[16]	Courchesne E，Yeung-Courchesne R，Press GA，et al.Hypoplasia of cerebellar vermal lobules Ⅵ and Ⅶ in autism[J].N Engl J Med，1988，318(21)：1349-1354.DOI：10.1056/NEJM198805263182102
[17]	Kaufmann WE, Cooper KL, Mostofsky SH, et al. Specificity of cerebellar vermian abnormalities in autism：a quantitative magnetic resonance imaging study[J].J Child Neurol，2003，18(7)：463-470.DOI：10.1177/08830738030180070501.
[18]	Heemskerk R．Managing aggression in autism spectrum disorder：a cerebellar neuroscience approach[D].Utrecht：Utrecht University，2020.
[19]	Yoo S，Jang Y，Hong SJ，et al.Whole-brain structural connectome asymmetry in autism[J].Neuroimage，2024(288)：120534.DOI：10.1016/j.neuroimage.2024.120534.
[20]	Allen G，McColl R，Barnard H，et al.Magnetic resonance imaging of cerebellar-prefrontal and cerebellar-parietal functional connectivity[J].Neuroimage，2005，28(1)：39-48.DOI：10.1016/j.neuroimage.2005.06.013.
[21]	Courchesne E，Campbell K，Solso S．Brain growth across the life span in autism：age-specific changes in anatomical pathology[J].Brain Res，2010(1380)：138-145.DOI：10.1016/j.brainres.2010.09.101.
[22]	Işıklar S，Demir i，Özdemir ST，et al.Examination of the development and asymmetry of the cerebellum and its lobules in individuals aged 1-18 years：a retrospective MRI study[J].Brain Topogr，2023，36(6)：901-925.DOI：10.1007/s10548-023-00997-2.
[23]	Goldstein JM，Seidman LJ，Horton NJ，et al.Normal sexual dimorphism of the adult human brain assessed by in vivo magnetic resonance imaging[J].Cereb Cortex，2001，11(6)：490-497.DOI：10.1093/cercor/11.6.490.
[24]	Fortin JP，Cullen N，Sheline YI，et al.Harmonization of cortical thickness measurements across scanners and sites[J].Neuroimage，2018(167)：104-120.DOI：10.1016/j.neuroimage.2017.11.024.

组别	例数	年龄(岁，±s)	FIQ(±s)	男性[例(%)]
ASD组	1 030	15.77±8.71	106.04±16.83	892(86.60)
TD组	1 117	15.94±8.55	113.05±12.70	852(76.28)
统计量		t＝－0.44	t＝－10.43	χ²＝36.80
P值		0.662	<0.01	<0.01

组别	例数	年龄(岁，±s)	FIQ(±s)	男性[例(%)]
ASD组	1 030	15.77±8.71	106.04±16.83	892(86.60)
TD组	1 117	15.94±8.55	113.05±12.70	852(76.28)
统计量		t＝－0.44	t＝－10.43	χ²＝36.80
P值		0.662	<0.01	<0.01

脑区	固定效应	估计值	标准误	自由度	t值	P值
左侧	截距	2.078	0.540	86.40	3.849	0.0002
	诊断(TD vs ASD)	0.515	0.243	2125.41	2.115	0.0345
	性别(女vs男)	－0.959	0.318	2138.02	－3.014	0.0026
	年龄	－0.094	0.020	1078.44	－4.673	<0.0001
右侧	截距	2.111	0.500	95.35	4.226	<0.0001
	诊断(TD vs ASD)	0.368	0.242	2128.73	1.524	0.1280
	性别(女vs男)	－0.925	0.316	2141.46	－2.932	0.0034
	年龄	－0.091	0.020	834.83	－4.641	<0.0001

脑区	固定效应	估计值	标准误	自由度	t值	P值
左侧	截距	2.078	0.540	86.40	3.849	0.0002
	诊断(TD vs ASD)	0.515	0.243	2125.41	2.115	0.0345
	性别(女vs男)	－0.959	0.318	2138.02	－3.014	0.0026
	年龄	－0.094	0.020	1078.44	－4.673	<0.0001
右侧	截距	2.111	0.500	95.35	4.226	<0.0001
	诊断(TD vs ASD)	0.368	0.242	2128.73	1.524	0.1280
	性别(女vs男)	－0.925	0.316	2141.46	－2.932	0.0034
	年龄	－0.091	0.020	834.83	－4.641	<0.0001

区域	组别	EMMeans	标准误	自由度	t值	P值	EMMeans(95%CI)
左侧	ASD	0.116	0.432	45.7			－0.755~0.987
	TD	0.630	0.427	43.2			－0.231~1.491
	ASD-TD	－0.515	0.243	2126.3	－2.115	0.0346	－0.992~－0.037
右侧	ASD	0.212	0.388	49.3			－0.568~0.993
	TD	0.580	0.382	45.8			－0.189~1.350
	ASD-TD	－0.368	0.242	2129.5	－1.524	0.1277	－0.842~0.106

Please choose a citation manager

Content to export