Brain structural and genetic correlates of motor coordination and learning behaviours: modelling developmental coordination disorder
Jason P. LerchiD, David George AshbrookiD, Kamaldeep K. GilliD, Jeffy Rajan Soundara RajaniD, Jacob EllegoodiD, John G. SlediD, Brian J. NiemaniD, Jill Glennis ZwickeriD, Dan GoldowitziD
Abstract
Developmental coordination disorder (DCD) is a common neurodevelopmental condition characterized by impaired motor coordination and learning, yet its neurobiological and genetic bases remain poorly understood. Here, we leverage the BXD recombinant inbred mouse panel to model the polygenic architecture of DCD and link behaviour, brain structure, and genotype. High-resolution ex vivo MRI across 14 strains revealed that DCD-like mice have modestly reduced total brain volume, with a distinct neuroanatomical profile characterized by enlarged cortical regions alongside reduced cerebellar, thalamic, and other subcortical volumes. These structural differences closely mirror findings reported in human DCD. Across strains, variation in brain structure strongly correlated with motor behaviours, with coordinated patterns linking increased cortical and decreased subcortical volumes to poorer motor coordination, while more focal associations were observed for motor learning. Multivariate analysis identified a dominant brain-behaviour axis capturing this cortical-subcortical trade-off. Quantitative trait locus (QTL) mapping revealed multiple loci influencing regional brain volumes, including a prominent locus on chromosome 12 regulating cerebellar structures, but did not identify single loci driving the main multivariate brain-behaviour relationships, consistent with a distributed genetic architecture. Together, these findings demonstrate that DCD-like motor impairments arise from coordinated alterations across distributed brain systems under polygenic control. This work establishes a translational framework linking genetic variation to brain organization and motor function, and suggests that DCD reflects the extreme of a continuous spectrum of neurobiological variation rather than a discrete condition.
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 1
- Organism
- Assay
- Research Focus
- Approach
In-depth summary
Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (4 Aug 2026) · CC BY 4.0
In 14 BXD mouse strains, those modelling developmental coordination disorder have relatively larger cortex and smaller cerebellar/subcortical volumes; this axis tracks motor coordination but maps to no single QTL.
Developmental coordination disorder (DCD) is heritable but genetically complex, its brain and genetic bases remain unclear, and mouse models are less often used for such complex, polygenic conditions. Building on earlier work that used behaviour to identify five DCD-like strains in the BXD recombinant inbred panel, this study adds whole-brain MRI and quantitative trait locus (QTL) mapping to connect genotype, neuroanatomy and motor performance.
Key results
SuggestedStrains classified as DCD-like have slightly smaller brains. Once total volume is accounted for, they have relatively larger cortical areas and cortical white-matter bundles but a smaller thalamus, smaller cerebellar nuclei and several smaller subcortical tracts.
5 DCD-like vs 9 non-DCD-like strains: total brain volume 2.6% smaller (t=-2.6, p=0.014). Relative to brain size: larger cingulate areas 24b/29c, orbital area, parietal association cortex, primary somatosensory cortex, primary/secondary motor cortex, corpus callosum and cingulum. Smaller thalamus, cerebellar nuclei, stria terminalis, stria medullaris and mammillothalamic tract. · RESULTS; Figure legends
SuggestedAcross strain means, regional volumes correlate strongly with motor scores. Gait measures rise with cortical motor and corpus callosum volume and fall with cerebellar, pallidal and thalamic volume. Open-field speed shows the opposite pattern, and learning measures link to fewer, more focal regions.
Pearson correlations of 590 regions (adjusted for total brain volume) against 25 scores from 7 assays over 14 strain means; several |r| > 0.9. The authors note that correlations across individual mice top out near |r| 0.5. Examples: skilled reaching with a smaller corpus callosum; complex wheel and accelerated rotarod gains with a smaller thalamus. · RESULTS; Discussion; Figure legends
SuggestedA sparse canonical correlation analysis (CCA) finds one dominant brain-behaviour axis: relatively larger cortex with smaller cerebellar and subcortical regions. Of the first four components, it is the only one that significantly separates DCD-like from non-DCD-like strains.
8 canonical components kept r ≥ 0.5. Component 1 loaded positively on step cycle, stance and swing duration, posterior extreme position and duty factor, and negatively on body speed, velocity and skilled reaching learning rate. Among the first four components, only component 1 separated the groups (p=0.017); otherwise the strains were interspersed. · RESULTS; Figure legends
SuggestedQTL mapping of regional volumes finds many loci, including a chromosome 12 cluster that influences several cerebellar hemisphere lobules. Several of these regions reach significance in both the trained and the untrained subsets of mice.
Strain means for all, untrained and trained mice gave 201 suggestive (LOD > 3.5) and 92 nominally significant (LOD > 4) QTL. The chr12 110.7-120 Mb locus affects ansiform lobule crus 1 and 2 and their white matter, hemispheric regions, the paramedian lobule and the simple lobule. · QTL mapping; QTL analyses; Tables
SuggestedThe dominant brain-behaviour components have no detectable QTL. The authors read this as distributed, polygenic control rather than a single locus driving the DCD-like network.
QTL scans on CCA components: none for components 1-3; component 4 LOD = 4 at chr7:59.99-74.52 Mb; component 8 LOD = 6.1 at chr8:121.59-125.16 Mb. · QTL analyses; Discussion
Methods, data and reproducibility
Mice from 12 BXD strains and the C57BL/6J and DBA/2J parents were grouped as DCD-like or not from a motor test battery. All were imaged by ex vivo 7 T MRI and segmented into 590 regions. Strain-mean volumes were related to 25 motor scores by Pearson correlation and sparse CCA, and mapped to loci with R/qtl2 using a kinship matrix.
- BXD recombinant inbred strains (12) plus parental C57BL/6J and DBA/2J, ex vivo T2-weighted MRI at 0.040 mm isotropic resolution · mouse · n = 193 mice, 14 strains (101 behaviour-tested, 92 MRI only)
- Motor battery: gait, open field, rotarod, accelerated rotarod, complex wheel, horizontal rung, skilled reaching (25 scores from 7 assays) · mouse · n = 101 mice
- BXD genotypes derived from whole-genome sequencing, downloaded from GeneNetwork.org · mouse · n = 21,056 markers
- Methods & tools
- MAGeT, mixOmics (R), R/qtl2, GeneNetwork.org, Agilent multi-channel 7.0 T MRI scanner, ProHance gadolinium contrast
- Data & accessions
- http://genenetwork.org/
- Reproducibility gaps
- The text does not say whether the MRI images, regional volumes or behavioural data are deposited anywhere. The QTL results are cited to a supplementary table that is not part of the text.
Limitations (5)
- Only 14 strains were used, which limits the power to detect QTL and raises the risk of false positives. The detected QTL intervals are too wide to pinpoint individual genes or variants. · stated by the authors
- MRI-derived volumes give unbiased whole-brain coverage comparable to human studies, but they do not readily give cellular or molecular explanations. · stated by the authors
- Imaging and behavioural testing were done in young adult mice, so the study cannot address developmental explanations for DCD. · stated by the authors
- QTL thresholds are reported inconsistently (suggestive is -logP > 3 in Methods but LOD > 3.5 in Results). Hundreds of regional traits were scanned in three groupings, yet the loci are only 'nominally' significant and no correction across traits is stated. · noted in this summary
- The very high |r| values come from only 14 strain means screened across 590 regions and 25 scores. The text does not say how many of these correlations survive FDR correction, so individual large correlations may be inflated. · noted in this summary
Who should read:Read this if you study the brain or genetic basis of motor disorders, or want an example of combining BXD strain-mean MRI volumes with behaviour and QTL mapping.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Developmental Coordination Disorder: What Can We Learn From Recombinant Inbred Mice Using Motor Learning Tasks and Quantitative Trait Locus Analysis
Kamaldeep K. Gill, Jeffy Rajan Soundara Rajan, … Dan Goldowitz·bioRxiv·May 18, 2022
- Investigating mouse motor coordination using quantitative trait locus analysis to model the genetic underpinnings of developmental coordination disorder
Kamaldeep K. Gill, Jeffy Rajan Soundara Rajan, … Dan Goldowitz·bioRxiv·Jun 7, 2022
- Body schema plasticity is altered in Developmental Coordination Disorder
Marie Martel, Véronique Boulenger A, … Alice Catherine Roy·NeuropsychologiaPeer-reviewed·Dec 23, 2021
- Developmental coordination disorder affects the pre-ordering of sequential movements
Helena Wright-Wieckowski, Kate Wilmut, Katja Kornysheva·bioRxiv·Jun 12, 2026
- Motor and non-motor sequence prediction is equally affected in children with Developmental Coordination Disorder
Bertram Opitz, Daniel Brady, Hayley C. Leonard·PLoS ONEPeer-reviewed·Nov 9, 2020
- Distinct impact modes of polygenic disposition to dyslexia in the adult brain
Sourena Soheili‐Nezhad, Dick Schijven, … Clyde Francks·Science AdvancesPeer-reviewed·Dec 18, 2024