<?xml version="1.0"?>
<response><xml version="1.0" encoding="UTF-8"><resource xmlns="http://datacite.org/schema/kernel-4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4/metadata.xsd"><identifier identifierType="DOI">10.60964/rnd-vnfn-1492</identifier><creators><creator><creatorName nameType="Personal">Yona G</creatorName><givenName>Guy</givenName><familyName>Yona</familyName><nameIdentifier nameIdentifierScheme="ORCID" schemeURI="https://orcid.org">https://orcid.org/0000-0003-0695-0827</nameIdentifier></creator></creators><titles><title xml:lang="en">ClustME: a MATLAB Toolbox for Fast Cluster-Based Permutation Testing with Linear Mixed-Effects Models</title></titles><resourceType resourceTypeGeneral="Dataset">ClustME: a MATLAB Toolbox for Fast Cluster-Based Permutation Testing with Linear Mixed-Effects Models</resourceType><publisher>University of Oxford</publisher><publicationYear>2026</publicationYear><dates><date dateType="Issued">2026</date></dates><language>en</language><descriptions><description xml:lang="en" descriptionType="TechnicalInfo"><![CDATA[ClustME is a MATLAB toolbox for fast, family-wise error rate (FWER)-controlled cluster inference in hierarchical time-series data.It combines linear mixed-effects (LME) modelling with cluster-based permutation testing, allowing users to analyse repeated observations, subject -level structure, and trial-level predictors without collapsing observations to subject-level averages.This toolbox is available on GitHub. This DOI relates to release v1.0.0.Directly combining LMEs with cluster-based permutation testing is usually computationally prohibitive, because a full mixed-effects model would need to be refitted at every time sample and for every randomisation. ClustME avoids this by estimating a pooled, static estimate of the marginal covariance matrix V&nbsp;and reusing it during null generation.At each time sample, the test statistic is a Generalized Least Squares (GLS) contrast derived from the specified LME model. GLS uses the estimated covariance structure to account for hierarchical dependence that ordinary least-squares approaches would otherwise ignore. This makes mixed-effects cluster inference feasible on standard hardware, provided that the selected randomisation method matches the design's exchangeability structure.Early release noteClustME is currently in early release. The core v1.0.0 functionality has been validated, but user feedback will help improve the toolbox across a wider range of datasets and designs. Please use the GitHub repository or email guy.yona@ndcn.ox.ac.uk to ask questions, report issues, or suggest improvements.]]></description></descriptions><fundingReferences><fundingReference><funderName>Medical Research Council, UKRI</funderName><awardNumber>MC_UU_00003/5</awardNumber></fundingReference></fundingReferences></resource></xml></response>
