TRAMFIX: TRavelling Across Melbourne for FIXel-based analysis (a reproducibility study)

TRAMFIX: TRavelling Across Melbourne for FIXel-based analysis (a reproducibility study) | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results TRAMFIX: TRavelling Across Melbourne for FIXel-based analysis (a reproducibility study) View ORCID Profile Remika Mito , View ORCID Profile Sila Genc , Jocelyn Halim , View ORCID Profile Joseph Yuan-Mou Yang , View ORCID Profile Jacques-Donald Tournier , Michael Kean , Chris Kokkinos , Richard McIntyre , View ORCID Profile Maria A. Di Biase , View ORCID Profile Robert Smith , View ORCID Profile Andrew Zalesky doi: https://doi.org/10.1101/2025.07.21.666058 Remika Mito 1 Department of Psychiatry, The University of Melbourne , Parkville, VIC, Australia 2 Florey Department of Neuroscience and Mental Health, The University of Melbourne , Parkville, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Remika Mito For correspondence: remika.mito{at}unimelb.edu.au Sila Genc 3 Developmental Imaging, Clinical Sciences, Murdoch Children’s Research Institute , Parkville, VIC, Australia 4 Neuroscience Advanced Clinical Imaging Service (NACIS), Department of Neurosurgery, The Royal Children’s Hospital , Parkville, VIC, Australia 5 Neuroscience Research, Murdoch Children’s Research Institute , Parkville, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Sila Genc Jocelyn Halim 1 Department of Psychiatry, The University of Melbourne , Parkville, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Joseph Yuan-Mou Yang 4 Neuroscience Advanced Clinical Imaging Service (NACIS), Department of Neurosurgery, The Royal Children’s Hospital , Parkville, VIC, Australia 5 Neuroscience Research, Murdoch Children’s Research Institute , Parkville, VIC, Australia 6 Department of Paediatrics, The University of Melbourne , Parkville, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Joseph Yuan-Mou Yang Jacques-Donald Tournier 7 Department of Early Life Imaging, School of Biomedical Engineering & Imaging Sciences, King’s College London , London, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Jacques-Donald Tournier Michael Kean 3 Developmental Imaging, Clinical Sciences, Murdoch Children’s Research Institute , Parkville, VIC, Australia 6 Department of Paediatrics, The University of Melbourne , Parkville, VIC, Australia 8 Department of Medical Imaging, The Royal Children’s Hospital , Parkville, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Chris Kokkinos 9 Florey Institute of Neuroscience and Mental Health , Heidelberg, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Richard McIntyre 10 Monash Biomedical Imaging, Monash University , Clayton, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site Maria A. Di Biase 1 Department of Psychiatry, The University of Melbourne , Parkville, VIC, Australia 11 Department of Anatomy and Physiology, The University of Melbourne , VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Maria A. Di Biase Robert Smith 2 Florey Department of Neuroscience and Mental Health, The University of Melbourne , Parkville, VIC, Australia 9 Florey Institute of Neuroscience and Mental Health , Heidelberg, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Robert Smith Andrew Zalesky 1 Department of Psychiatry, The University of Melbourne , Parkville, VIC, Australia 12 Department of Biomedical Engineering, The University of Melbourne , Parkville, VIC, Australia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Andrew Zalesky Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract Fixel-based analysis (FBA) has gained substantial interest for its ability to probe fibre-specific changes in the brain’s white matter from diffusion-weighted imaging data. However, the reproducibility and reliability of fixel-based measures across different scanners remains largely unknown. In this work, we present TRAMFIX: a multisite dataset of travelling participants (n=10 healthy adults) scanned across four 3T MRI scanners using a harmonised multi-shell DWI protocol. DWI data were processed using two pipelines that can adopted when performing multi-site FBA studies (site-specific versus pooled processing). We extracted fixel-based measures of fibre density (FD), fibre cross-section (FC), and fibre density and cross-section (FDC), as well as diffusion tensor imaging (DTI)-based fractional anisotropy (FA) and mean diffusivity (MD) from the harmonised protocol. Within-subject coefficients of variation (CV ws ) and intraclass correlation coefficients (ICC) of FBA and DTI measures were computed at multiple resolutions of computation and analysis: (i) the whole-brain averaged level, (ii) tract-level, and (iii) fixel- or voxel-level. Fixel-based metrics demonstrated high reproducibility and reliability at the whole-brain level (CV ws ranging between 0.51% to 1.57% and ICC between 0.93 and 0.996). While reproducibility and reliability remained high for tract-averaged FBA measures (particularly the FC and FDC metrics), some tracts exhibited lower ICC values < 0.8 for the FD measure. When examining fixel-level reliability and reproducibility clear spatial patterns emerged, with lower ICC across subcortical and cerebellar regions, and higher CV ws at the cortical boundaries. Across all levels, tensor-based metrics demonstrated slightly lower reproducibility and reliability metrics than FBA measures, suggesting that FBA metrics are just as, if not more, reproducible than DTI metrics across scanners when using identical protocols. Our findings provide support for the reproducibility and reliability of fixel-based measures, highlighting their potential for use in multi-site FBA studies. Future work examining protocol-related differences, as well as appropriate harmonisation strategies when pooling data across sites and scanners, will be valuable. 1. Introduction Diffusion-weighted imaging (DWI; or diffusion MRI) is currently the only tool available to non-invasively map the human brain’s complex white matter architecture. While the development of diffusion tensor imaging (DTI) was a revolutionary step in our ability to quantify changes to the brain’s white matter connections ( Basser et al., 1994 ), known limitations to the technique’s modelling assumptions and biological interpretations ( Jones et al., 2013 ) have led to major improvements in DWI acquisition and modelling approaches. The past few decades have seen a rapid rise in the development and application of these advanced DWI techniques, enabling improved characterisation of the brain’s complex white matter microstructure. Fixel-Based Analysis (FBA; ( Raffelt et al., 2017 )) is one such framework that can provide more specific insight into the brain’s white matter fibre pathways. By modelling high-angular resolution diffusion imaging (HARDI) data with constrained spherical deconvolution (CSD; ( Dell’Acqua & Tournier, 2019 )), FBA can derive quantitative measures even in the presence of multiple fibre directions within voxels (known as ‘fixels’), enabling fibre-specific analysis of white matter pathways. The FBA framework has now been widely adopted across a range of clinical applications (see Dhollander et al., (2021) for a review), including in neurodegeneration (Ahmadi et al., 2024; Mito et al., 2018 ), psychopathology ( Grazioplene et al., 2022 ; Kristensen et al., 2023 , 2024 ; Lyon et al., 2019 ) and neurological disorders ( Mito et al., 2022 , 2024 ; Raffelt et al., 2017 ; Vaughan et al., 2017 ), as well as in healthy development ( Genc et al., 2018 ; Genc, Malpas, et al., 2020) and aging ( Kelley et al., 2021 ; Tinney et al., 2024 ). A key advantage of the technique is in its ability to identify fibre tract-specific changes, overcoming some of the key limitations to DTI and voxel-aggregate scalars ( Mito et al., 2018 ; Raffelt et al., 2017 ). Despite the technique’s growing popularity, FBA studies have thus far been largely limited to single-site designs. A key barrier for multi-site diffusion MRI studies more broadly is that quantitative measures are highly susceptible to differences between scanners and protocols, which can pose a major issue when pooling data across sites ( Pinto et al., 2020 ). FBA poses additional challenges for data harmonisation, as recommended pipelines necessitate the construction of study-specific templates, and CSD modelling approaches may differ based on acquisition (e.g., multi-shell vs single-shell) ( Dhollander & Connelly, 2016 ; Jeurissen et al., 2014 ). Although the extent of scanner or protocol related variability has been well documented for other DWI-based measures ( Besseling et al., 2012 ; Cai et al., 2021 ; Fan et al., 2021 ; Grech-Sollars et al., 2015 ; Koller et al., 2021 ; Luque Laguna et al., 2020 ; Schilling et al., 2021 ; Vollmar et al., 2010 ), the extent to which site or scanner differences impact fixel-based measures is unclear. The goal of this study was to assess the reproducibility and reliability of fixel-based measures using participants scanned across multiple sites (a ‘travelling heads’ study). Here, we present the TRAMFIX (TRavelling Across Melbourne for FIXel-based analysis) cohort, in which ten healthy adults were scanned across four 3 T MRI scanners located at three different sites with an identical DWI protocol. We computed fixel-based metrics using two pipelines that could be used when combining data across sites (site-specific versus pooled processing). Reproducibility and reliability of these measures across sites were assessed at three different levels of analysis: (i) averaged across the whole-brain; (ii) averaged across white matter tracts; and (iii) at the individual fixel-level. While the terms reproducibility and reliability are often used interchangeably, they represent different properties ( Shoukri et al., 2008 ), and were assessed using within-subject coefficient of variation (CV ws ) and intraclass correlation coefficient (ICC), respectively. For comparison, we also examined the reproducibility and reliability of tensor-based metrics at these equivalent three levels (whole-brain, tract-averaged, and voxel-level). 2. Methods 2.1 Participants Ten neurologically healthy adult volunteers (7 female, age range: 19-30 years) were recruited into the study via internal email advertisement and notice boards at participating institutions (University of Melbourne, the Florey Institute of Neuroscience and Mental Health, and Monash University). All volunteers were screened for MRI safety and enrolled into the study if they had no contraindications to MRI, and were provided monetary compensation for travel and participation. Participants were included in the analysis if all MRI scans were acquired within a 6-month period. All scans were completed between September to December 2024, with an average interval between first and last scan of 17.8 days (range: 2 to 64 days). Informed written consent was obtained from all participants for being included in the study, and ethical approval was granted by the University of Melbourne’s Human Research Ethics Committee (HREC #28396). 2.2 MRI scanner characteristics Brain images were acquired on four 3T MRI scanners at three different sites across Melbourne, Australia (see Table 1 ). These included a 3T Magnetom Skyra, Vida, and two Prisma Fits (Siemens Healthcare, Erlangen, Germany). All systems were equipped with a 32-channel head coil. View this table: View inline View popup Download powerpoint Table 1: Site and scanner information 2.3 MRI data acquisition Each MRI session lasted 1 hour, with approximately 40-50 mins of scan time at each site. Diffusion-weighted images and structural images were acquired at each site, with all sites acquiring one consistent harmonisation protocol for DWI, along with additional site-specific protocols not used in the present study. The main harmonisation protocol was a multi-shell DWI acquired with a spin-echo EPI sequence (TE/TR = 113/3345 ms, voxel size = 2 mm 3 , matrix size:116 x 116 x 76, multi-band factor 4, no in-plane parallel acceleration). The complete diffusion gradient table consisted of: b =0 s/mm 2 (10 volumes), b =300 s/mm 2 , (13 directions), b =1000 s/mm 2 (24 directions), b =2000 s/mm 2 (42 directions), and b =3000 s/mm 2 (60 directions). The diffusion encoding scheme was optimally split, with half the volumes acquired in the anterior-posterior (AP) phase-encoding direction, and half the volumes acquired in the posterior-anterior (PA) direction (designed to achieve optimal sampling across the AP and PA directions using a similar implementation originally developed for the developing Human Connectome Project (dHCP) ( Tournier et al., 2020 )). Acquisition time for this protocol was ∼9 minutes. 2.4 DWI data processing Here, we discuss the processing steps for the common DWI harmonisation protocol. All DWI data were preprocessed according to the recommended MRtrix3 pipeline for multi-tissue fixel-based analysis. This included MP-PCA denoising data ( Veraart et al., 2016 ), Gibbs ringing removal ( Kellner et al., 2016 ), EPI susceptibility and motion correction using FSL topup and eddy ( Andersson et al., 2003 ; Andersson & Sotiropoulos, 2016 ) with eddy QC ( Bastiani et al., 2019 ), bias field correction using ANTs N4ITK ( Tustison et al., 2010 ), and then upsampling all data to a voxel resolution of 1.25 mm 3 . For susceptibility field estimation, only leading b =0 volumes were extracted and included, as interleaved b =0 volumes were observed to contain residual eddy current distortions from prior volumes. To mimic two potential strategies for data analysis when combining DWI data across sites for fixel-based analysis, we then processed the cohort using two distinct pipelines (see Figure 1 ). The first pipeline (Pipeline 1: “site-specific processing”) processed DWI data for each site independently, only pooling data across sites for final analysis. The second pipeline (Pipeline 2: “pooled processing”) instead pooled DWI data across all sites for processing and analysis. Download figure Open in new tab Figure 1: Processing pipelines for extracting fixel-based metrics. Here, we used two alternative approaches that might be taken when combining multi-site data. All data preprocessing, tensor fitting, brain mask estimation, and subject-specific response function estimation were performed in common (once per scanning session). Pipeline 1 describes the scenario where FOD modelling might be performed at a site-specific level (i.e., using site-averaged response functions (RFs) and site-specific templates). Pipeline 2 describes the scenario where data are pooled from different sites prior to FOD modelling (i.e., using globally averaged RFs and a global population template). In both cases, after DWI data preprocessing, response functions were estimated for white matter, grey matter and CSF ( Dhollander et al., 2019 ) for each participant. Fibre orientation distribution (FOD) functions were then estimated using multi-shell multi-tissue CSD (MSMT-CSD) ( Jeurissen et al., 2014 ). For the first pipeline, site-specific response functions (RFs) were computed (average response functions across all participants at a given site) and used for FOD estimation, while in the second pipeline, response functions averaged across all participants and scanning sessions (i.e., incorporating data from all individuals across all sites) were used for FOD modelling. For each pipeline, a study-specific FOD population template was built using an iterative registration and averaging approach ( Raffelt et al., 2011 ). For Pipeline 1, this was done by first building site-specific templates, then a common population template space using site-specific templates as input; for Pipeline 2, FOD images from all scanning sessions were used in a single template construction step. Each individual FOD image was then registered to the template using FOD-guided non-linear registration ( Raffelt et al., 2012 ). Measures of fibre density (FD), fibre cross-section (FC), and fibre-density and cross-section (FDC) were then computed according to the fixel-based analysis pipeline ( Raffelt et al., 2017 ). Whole-brain tractography was performed on the population template image, by first generating 20 million streamlines, which were subsequently filtered to 2 million streamlines by performing SIFT ( Smith et al., 2013 ). Smoothing was applied to fixel-based measures using the fixel-fixel connectivity matrix computed from this tractogram ( Raffelt et al., 2015 ). To compare the reproducibility of fixel-based measures with more conventional tensor-based metrics, we additionally computed measures of fractional anisotropy (FA) and mean diffusivity (MD) from the diffusion tensor. Here, the b =1000 s/mm 2 shell was extracted from the multi-shell harmonisation dataset, and the tensor fit performed using an iterated weighted least-squares approach ( Veraart et al., 2013 ). FA and MD were estimated for each image, and then transformed to population template space (using non-linear transformations computed through registration of FOD images). In order to compare reproducibility of FA and MD to the fixel-based measures, voxel masks were computed as the set of all voxels containing at least one fixel, and tensor measures averaged within those voxel masks. Given that these masks are more extensive than what is typically included within DTI-based analyses, we additionally computed the mean values of these tensor measures within a DTI-specific white matter mask, including only template voxels with a mean FA > 0.2. We additionally investigated the reproducibility and reliability of fixel-based measures when computed from single-shell data. We extracted b = 1000, 2000, and 3000 s/mm 2 shells from the full multi-shell dataset after preprocessing. Response functions were estimated for each b -shell, followed by FOD estimation using single-shell 3-tissue CSD (SS3T-CSD; using MRtrix3Tissue) ( Dhollander & Connelly, 2016 ). As we used the same population template image for these analyses and the fibre cross-section (FC) metric is a measure of deformation to template space, we examined only the reproducibility of the fibre density (FD) measure. All processing steps (unless otherwise specified) were performed using MRtrix3 (version 3.0.4) ( Tournier et al., 2019 ). 2.5 Tract segmentation Tracts of interest were delineated by performing TractSeg ( Wasserthal et al., 2018 ) on the population template FOD peaks image. Mean FD, FC, and FDC were computed for each tract, including only fixels with correspondence to the local tract direction. Mean voxel-based FA and MD were also computed for each tract, averaging across those voxels containing a fixel ascribed to the tract. 2.6 Statistical analysis To assess the reproducibility and reliability of FBA and tensor-based metrics across the different sites, we computed the coefficients of variation (CV) and intraclass correlation coefficient (ICC). CV is commonly used to assess the reproducibility of measures, and is broadly defined as the standard deviation of the group (σ) divided by the mean of the group (μ) (often expressed as a percentage). When assessing CV across sites with the same subjects, the within-subject variability (CV ws ) is considered a more appropriate measure for reproducibility than pooled CV ( Shoukri et al., 2008 ). CV ws was calculated following recommendations from the quantitative imaging biomarkers alliance (QIBA) for DWI biomarkers (Shukla-Dave et al., 2018). Here, squared CV ws was computed for each subject (dividing variance by squared mean), and average squared CV ws across subjects computed. CV ws was therefore expressed as: While we focus here on within-subject variability to assess reproducibility, we additionally computed between-subject CV (CV bs ), which captures biological variability between subjects. CV bs was computed by first taking the within-subject means ( μ i ), then computing Intraclass correlation coefficient was computed using ICC(3,1), representing two-way mixed effects, consistency, single rater/measurement ( Koo & Li,.2016 ), calculated as: Intraclass correlation coefficient was computed using ICC(3,1), representing two-way mixed MS R is the mean square for rows (i.e., subjects), while MS E is the mean square for error. While CV ws captures the reproducibility of metrics, ICC was used to assess reliability of these measures (the overall consistency of measures across individuals) (Luque Laguna et al., 2020 ; Shoukri et al., 2008 ). Lower CV ws values reflect better reproducibility, while higher ICC values reflect higher reliability. Generally, ICC values less than 0.5 indicate poor reliability, values between 0.5 and 0.75 reflect moderate reliability, values between 0.75 and 0.9 indicate good reliability, and values greater than 0.9 reflect excellent reliability ( Koo & Li, 2016 ). CV ws and ICC were assessed for each of the fixel-based measures (FD, FC, and FDC) at: (i) the whole brain level, where measures were averaged across all white matter fixels; (ii) at the tract level, where measures were averaged across tracts of interest; and (iii) at the fixel level, where reproducibility and reliability were assessed at each fixel in template space. These metrics were also computed for tensor-based scalars (FA and MD) at the same three levels. To compute estimates of sample size requirements for tract-level analyses, we performed power calculations (at α = 0.05, power = 0.8) at different expected observable effect sizes (Cohen’s d ), using a correction formula implemented in adjusting for group-level correlations ( Bliese, 1998 ), where: Here, ICC was used as a proxy for reliability, and the computed ICC values from the tract-level results were used. Three specific tracts (corticospinal tract (CST), arcuate fasciculus (AF), and fornix (FX)) were selected a priori for sample size calculation to capture different types of tracts. This was based on inclusion in previous repeatability work ( Koller et al., 2021 ). This selection includes commonly studied tracts due to their clinical relevance (CST and AF), along with a tract that is more technically challenging to reconstruct consistently (FX). 3. Results 3.1 Reproducibility and reliability of fixel-based measures 3.1.1 Comparing whole-brain averaged measures Table 2 shows the reproducibility (indexed by CV ws ) and reliability (indexed by ICC) of fixel-based measures (as well as DTI-based measures; see Section 3.2 ) when averaged across the whole-brain white matter. Across both pipelines, these whole-brain averaged fixel-based measures demonstrated significant evidence of reliability (ICC p-value 0.9 across both pipelines). Variability within subject (CV ws ) was lower than between subject (CV bs ) for all measures, ranging between 0.51% to 1.57% for within-subject, and 2.26% to 8.27% for between-subject variability. Reproducibility and reliability metrics were highly comparable between the two pipelines ( Table 2 ). View this table: View inline View popup Table 2: Reproducibility and reliability of fixel-based and DTI-based measures across the whole brain white matter Figure 2 shows boxplots of the mean fixel-based measures across the four sites. Download figure Open in new tab Figure 2: Reproducibility of mean fixel-based measures across the whole-brain. Boxplots show whole-brain fixel-wise measures: (A and D) fibre density (FD); (B and E) fibre cross-section (FC); (C and F) fibre density and cross-section (FDC). The top row shows results for Pipeline 1 (site-specific processing), and bottom row shows results for Pipeline 2 (pooled processing). Points (in grey) represent whole-brain fixel-wise averages for each participant, with grey lines connecting points from the same individual across the four sites. Site-specific mean values and standard deviations (std) are shown for each measure. 3.1.2 Comparing tract-level measures When comparing tract-averaged fixel-based measures, reproducibility and reliability remained high across tracts. Reproducibility was high for the FD measure for both Pipeline 1 (CV ws ranged between 0.74 – 3.63%) and Pipeline 2 (CV ws ranged between 0.78 – 4.01%). Similarly, reproducibility was high for the FC metric (Pipeline 1 CV ws : 0.54 – 1.87%; Pipeline 2 CV ws : 0.60 – 1.76%) and for the FDC metric (Pipeline 1 CV ws : 1.23 – 5.03%; Pipeline 2 CV ws : 1.22 – 5.66%). Of the fixel-based measures, reliability was most variable for the FD measure (Pipeline 1: ICC ranged between 0.652 – 0.980; Pipeline 2: ICC ranged between 0.633 – 0.980). Reliability was high across both pipelines for the FC measure (Pipeline 1 ICC: 0.966-0.988; Pipeline 2 ICC: 0.967-0.998), and the FDC measure (Pipeline 1 ICC: 0.916-0.995; Pipeline 2 ICC: 0.911-0.994). Figure 3 shows the ICC values for the FD measure for each white matter tract with Pipeline 1. We focus here on the FD metric, as it was the most affected by site of the fixel-based metrics. Despite the FD measure having the largest variability in reliability, all but two tracts had an ICC > 0.8 (with 65 of 72 tracts showing excellent reliability with ICC > 0.9). Supplementary Figure S1 shows the ICC per tract for the FD measure with Pipeline 2 (same as Figure 3 , but with Pipeline 2). Download figure Open in new tab Figure 3: Intraclass correlation coefficient (ICC(3,1)) for the FD measure per tract using Pipeline 1. (A): Bar plot showing ICC for each tract (ordered from highest to lowest). Most tracts showed a very high ICC (> 0.9), with the bilateral fornices (FX) showing the highest of all tracts, and bilateral striato-fronto-orbital tracts (STFO) showing the lowest ICC values. (B): Tract trajectories coloured by their ICC. The top image shows a number of key tracts, showing high (FX), moderate (corticospinal tracts; CST), and lower ICC (striato-fronto-orbital; ST-FO and rostrum of corpus callosum; CC 1). The bottom images show different ICC values for different tract classifications. See Wasserthal et al. (2018) for full tract definitions. Supplementary Tables S1-S3 report the reproducibility and reliability of fixel-based measures for each white matter tract with Pipeline 1 (site-specific processing). 3.1.3 Comparing fixel-level measures Figure 4 shows the ICC for each individual fixel in template space (Pipeline 1), for each of the fixel-based measures. In general, reliability was high, with mean ICC across all fixels being 0.83 for FD, 0.95 for FC, and 0.86 for FDC. Figure 5 shows a thresholded image to highlight regions with poor reliability for both FBA and tensor-based metrics (for which results are reported in 3.2). Fixels with lower reliability (ICC < 0.5) tended to be in cerebellar and subcortical grey matter regions, rather than in core white matter tracts. ICC was similarly high at the fixel-level when computing FODs using globally averaged response functions (Pipeline 2), with very comparable results to Pipeline 1 (see Table 3 for summary statistics on ICC across different pipelines and metrics). Download figure Open in new tab Figure 4: Reliability of FBA metrics at the fixel-level (Pipeline 1). Images show axial slices at 15 mm increments, with white matter fixels overlaid on the FOD template. Fixels are coloured by their ICC value. Top row shows ICC for the FD measure; middle row shows ICC for the FC measure; bottom row shows ICC for the FDC measures. Download figure Open in new tab Figure 5: Thresholded ICC image showing regions of lower reliability. ICC is shown for fixel-based analysis (FBA) metrics (top) and tensor-based metrics (bottom). In both cases, the ICC images are thresholded to only show voxels with poor reliability (ICC < 0.5), with bright areas corresponding to lower ICC values and displayed on coronal slices at 18mm increments. For FBA metrics, each voxel is coloured by the minimum value of ICC across the fixels within that voxel (to enable easier visual comparison with tensor-based metrics). View this table: View inline View popup Download powerpoint Table 3: Summary statistics for ICC at the fixel- or voxel-level Table 4 shows the reproducibility (within-subject variability; CV ws ) of fixel-based measures, summarised across all fixels within the white matter mask (density distribution plots shown in Supplementary Figure S3). Overall, mean CV ws was lowest for the FC measure (mean CV ws < 3%), while substantially higher for the FD and FDC measures. Higher average CV ws across all fixels for these two measures were likely driven by very large variability in a small number of fixels, with the mean CV ws around 15%, and median CV ws around 10% across all fixels. Figure 6 shows CV ws for each of the metrics on a single axial slice. Core white matter pathways tended to have high reproducibility, while lower reproducibility (higher CV ws ) was observed in the FD and FDC metrics in fixels closer to the cortical surface, and fixels crossing major white matter regions into subcortical regions (e.g., crossing fibre structures in the anterior limb of the internal capsule). Download figure Open in new tab Figure 6: Fixel-level reproducibility. CV ws is shown for each of the fixel-based measures: fibre density (FD), fibre cross-section (FC), and fibre density and cross-section (FDC) on a single axial slice. View this table: View inline View popup Table 4: Summary statistics for CV ws at the fixel- or voxel-level 3.2 Reproducibility and reliability of tensor-based measures In addition to comparing fixel-based measures, we also examined the reliability reproducibility of tensor-based fractional anisotropy (FA) and mean diffusivity (MD). When using a white matter voxel mask equivalent to the fixel mask used for FBA analyses, mean FA and MD showed high reliability and reproducibility, with high ICC (FA ICC = 0.923; MD ICC = 0.859) and low CV ws (FA CV ws = 1.24%, MD CV ws = 1.01%). Reliability and reproducibility were slightly lower when FA and MD were averaged across a constrained WM mask (see Table 2 ). Boxplots in Supplementary Figure S4 show mean whole-brain FA and MD for participants across sites. Taking tract-averaged FA and MD using the same TractSeg tract definitions used for FBA metrics, we examined the reproducibility of measures averaged at each white matter tract. Within-subject variability (CV ws ) for FA ranged from 0.65% to 4.58%, and intraclass correlation coefficient (ICC) from 0.486 to 0.963, with the lowest ICC reported in cerebellar tracts (middle cerebellar peduncle (MCP) and bilateral inferior cerebellar peduncles (ICP)). The CV ws for MD ranged from 0.90% to 1.98%, and ICC between 0.465 and 0.969. Figure 7 shows ICC for each of the TractSeg tracts for the FA measure. In general, reproducibility was lower for the FA measure than for the FD measure, and tracts showing lower/higher reproducibility did not necessarily overlap with the FD measure (see Figure 3 for comparison). Supplementary Tables S4 and S5 show results for each tract for FA and MD, respectively. Download figure Open in new tab Figure 7: Intraclass correlation coefficient (ICC(3,1)) for the FA measure per tract. (A): Bar plot showing ICC for each tract (ordered from highest to lowest), using same colour scaling as Figure 3. (B): Glass brain images show tracts coloured by their ICC value, for a selection of tracts with the highest and lowest ICC (top), and for different tract classifications (bottom images). Here, reproducibility was lowest for the cerebellar white matter tracts (middle cerebellar peduncle (MCP) and bilateral cerebellar peduncles (ICP)), and highest in the motor system tracts and thalamic radiations. When examining ICC across white matter voxels independently, a similar pattern was observed to fixel-based measures, with ICC lowest in cerebellar and subcortical white matter regions. However, mean ICC was substantially lower for tensor-based metrics than for fixel-based measures, with mean ICC across all WM voxels at 0.60 for FA and 0.55 for MD ( Table 3 ). Given that the fixel-based white matter mask may be more extensive than white matter masks used for DTI-based analyses, we also computed a DTI-specific white matter mask (constraining to voxels where the FA > 0.2). However, results were similar when using the constrained WM mask, with mean ICC across all WM voxels at 0.59 for FA, and 0.48 for MD. Figure 5 shows voxels with poor reliability for both the fixel-based and tensor-based measures, demonstrating the greater extent of lower reliability voxels in tensor-based metrics, which included some key white matter regions (e.g., corpus callosum, internal capsule). The reproducibility of tensor-based metrics at the voxel level was generally high within core white matter tracts, with CV ws similar to that of FBA metrics. Supplementary Figure S3 shows distribution of voxel-wise ICC and CV ws for all metrics, while Supplementary Figure S4 shows ICC and CV ws for the FA and MD measures on a single axial slice, alongside whole-brain boxplots. 3.3 Reproducibility of fibre density measure at different b-values The fibre density (FD) measure is known to be dependent on b -value, and as such, we also examined whether reproducibility of this measure would be different at different b -values (see Supplementary Figure S5). Here, we extracted each b -shell from the full multi-shell DWI, and modelled FODs using the single-shell data only. The FD measure was indeed dependent on b -value, with higher mean FD across the whole brain at b =1000 s/mm 2 and b =2000 s/mm 2 than at b =3000 s/mm 2 . Reproducibility of the FD measure was high across all b -shells, though was highest when using the b =1000 s/mm 2 data (ICC of 0.921), followed by the b =3000 s/mm 2 data (ICC of 0.914), and lowest for the b =2000 s/mm 2 shell (ICC of 0.879) (see Supplementary Table S6). Reproducibility of FD was high across all b -shells, with all demonstrating within-subject variability (CV ws ) around 1%. 3.4 Sample size estimates for tract-level analyses Sample size estimates were computed for tract-level analyses, by using the ICC values reported in sections 3.1.2 and 3.2 to estimate observable effect sizes at different expected true effects. Figure 8 shows sample size requirements per group for an independent sample t-test, with a = 0.05 and power = 0.8, for three select tracts (fornix, arcuate fasciculus, and corticospinal tract), for both the fibre density (FD) and fractional anisotropy (FA) metrics across a range of expected true effect sizes (0.5 < d < 1.0). Sample size requirements were similar for fixel-based FD to tensor-based FA, but smaller tracts like the fornix had lower sample size requirements for FD, due to the greater impact of site variability on the FA measure than for the FD measure. Download figure Open in new tab Figure 8: Sample size estimates for tract-level analyses for fibre density (FD) and fractional anisotropy (FA). Glass brain in top left shows the three tracts selected for sample size estimates (fornix (FX); arcuate fasciculus (AF); corticospinal tract (CST)). Bar plots show sample size requirements for each of these tracts, using the ICC values that were computed from the tract-level reproducibility analyses. Given the lower reproducibility of the FA measure for the fornix, sample size requirements to detect significant effects in this tract were higher than for the FD measure. For other tracts, sample size estimates were comparable. Of note, these sample size estimates are based on independent tract analyses. More stringent alpha thresholds that might be used when conducting multiple tract analyses would result in higher sample size requirements. 4. Discussion In this work, we present the TRAMFIX dataset, a travelling heads dataset designed to probe the reproducibility and reliability of fixel-based analysis (FBA) metrics derived from DWI data. We demonstrate that whole-brain and tract-averaged FBA measures are highly reproducible and reliable across scanners when using consistent protocols. However, when examining reproducibility and reliability at individual fixels, it was evident that certain white matter regions are more greatly impacted by scanner variability than others, which may be underappreciated at the whole-brain and tract-averaged level. This highlights the importance of careful interpretation in certain brain regions when conducting multi-site studies, and the potential need for appropriate harmonisation approaches. Of note, both reproducibility and reliability of FBA metrics was comparable, if not higher, than DTI-based measures, particularly for small fibre structures like the fornix, and in subcortical and cerebellar white matter. However, we also highlight that the b -value dependence of fixel-based measures should be taken into account when pooling data across different protocols. We additionally provide recommendations for sample size estimates when performing tract-level analyses with multi-site datasets, in light of the observed variability. Together, our findings provide novel insight into the reproducibility and reliability of fixel-based measures, which will potentially guide future studies that expand FBA beyond the single-site, and into multi-site designs. 4.1 Reproducibility and reliability of FBA measures There has been growing interest in the repeatability, reproducibility, and reliability of DWI-derived measures, as quantitative MRI measures are known to demonstrate some level of variability even under identical conditions ( Jansen et al., 2007 ; Koller et al., 2021 ; Vollmar et al., 2010 ), and to be highly susceptible to differences in scanner hardware and data acquisition ( Mirzaalian et al., 2016 ; Tax et al., 2019 ). There have now been a number of studies examining the reproducibility of tensor-based metrics ( Besseling et al., 2012 ; Koller et al., 2021 ; Luque Laguna et al., 2020 ; Magnotta et al., 2012 ; Palacios et al., 2017 ; Vollmar et al., 2010 ), measures from multicompartment models ( Andica et al., 2020 ; Fan et al., 2021 ; Veraart et al., 2021 ), and fibre tractography ( Schilling et al., 2021 ). This work extends our understanding of DWI reproducibility to fixel-based metrics, which are being increasingly adopted in research studies due to their ability to identify fibre tract-specific differences in clinical cohorts. Our findings demonstrate that FBA measures are highly reproducible and reliable across sites when using identical imaging protocols. FBA metrics demonstrated lower within-subject variation (CV ws ) than between-subject variation (CV bs ), and high intraclass correlation coefficient (ICC) across the different scanners, with comparable results to previous literature in other DWI metrics. Within our study, we similarly observed comparable reproducibility in FBA metrics as with DTI-based FA and MD, with slightly higher reliability evident for FBA measures particularly at the tract- and voxel-level. Our findings may be taken to support the expansion of fixel-based analysis studies beyond predominantly single-site into multi-site study designs. Indeed, reliability and reproducibility of fixel-based measures was very high when looking at whole-brain averaged or tract-averaged metrics, with high ICC values and low within-subject variability across most white matter bundles. This was in keeping with other DWI metrics, where inter-scanner variabilities have generally been reported to be less than 10% within select regions or tracts of interest ( Cai et al., 2021 ; Grech-Sollars et al., 2015 ; Palacios et al., 2017 ; Vollmar et al., 2010 ). DTI-based studies have increasingly been expanded to multi-site designs, and the comparable reproducibility of FBA metrics reported here demonstrates promise for studies that may wish to explore fibre-specific white matter differences in cohorts with multi-site data collection. To this end, we provide sample size estimates when studies may wish to probe tract-specific differences in multi-site cohorts. However, we do recommend caution when interpreting findings in brain regions that may be more susceptible to scanner differences. Importantly, in this study, we additionally examined reliability and reproducibility of fixel-based metrics at the individual fixel-level (as well as DTI-metrics at the voxel-level). Of note, previous studies examining variability of DWI measures across sessions or sites have predominantly focused on quantifying reliability (ICC) or reproducibility (CV ws ) in whole-brain, tract-, or ROI-averaged metrics. While this is a valuable way to assess and summarise DWI metrics, many clinical research studies aim to conduct group analyses at the more granular voxel-level (or in the case of FBA, at the fixel-level). Studies that have examined variability in DWI metrics at the voxel-level have generally assessed correlation (using Pearson’s R) or absolute deviation between paired tests (e.g., test-retest pairs) ( Fan et al., 2021 ; Koller et al., 2021 ). Here, we provide ICC and CV ws estimates at each individual fixel or voxel, enabling easy visualisation of brain regions that are affected by scanner differences. Indeed, our fixel- and voxel-level results indicated clear spatial patterns to reliability, with much lower ICC in some brain regions than was observable at the tract or whole-brain averaged level. The spatial pattern of low ICC within subcortical structures was similar to previous reports that have shown lower ICC in voxel-based measures (Luque Laguna et al., 2020 ). Areas of low reliability ( Figure 5 ) were predominantly in the subcortex and cerebellum, likely reflecting scanner-related differences in achievable SNR in these regions. However, we also note that for DTI-based measures, lower ICC was observed across many core white matter voxels. Reproducibility tended to be high across most white matter fixels for FBA metrics, and voxels for tensor-based metrics, with greater within-subject variability at cortical boundaries or in minor crossing fixels. Providing estimates of reliability and reproducibility at the individual voxel- or fixel-level may be informative for studies conducting multi-site DWI analyses, particularly when using scanners with different gradient systems. Of note, we investigated two potential pipelines that might be used when pooling data for fixel-based analyses across sites. Modelling data using constrained spherical deconvolution (CSD) relies on first estimating a response function from the DWI data. As fibre orientation distribution functions are estimated from the derived response functions (by deconvolution of the response function from the measured DWI signal ( Tournier et al., 2007 ), the use of standardised (for example, group-averaged) response functions is recommended when performing fixel-based analysis (Raffelt, Tournier, Rose, et al., 2012; Raffelt et al., 2017 ). We tested reproducibility of fixel-based measures using both site-specific response functions and FOD estimation, and pooled average response function estimation and FOD estimation. Although we expected combined processing (using pooled average RFs and FOD estimation) would reduce variability and improve reproducibility and reliability across sites, the results were in fact highly similar across the two pipelines. This may suggest that when combining data for FBA across multiple sites, data processing (that is, DWI modelling steps prior to template registration) could reasonably be performed at a site-specific level prior to pooling data across sites for further analysis (coregistration to a common template, and statistical analysis), at least where DWI data have been collected using matched protocols. However, we note that for this particular study, we made significant effort to utilise identical protocols with matched parameters across all sites, while including scanners from the same vendor, resulting in low variability across sites (CV ws < 5% across all white matter tracts). Where there is greater variability across sites, group-based processing may still help to improve reproducibility and reliability of fixel-based measures. Finally, we note that when assessing reproducibility of diffusion-based measures, coefficient of variation is often used ( Cai et al., 2021 ; Grech-Sollars et al., 2015 ; Koller et al., 2021 ; Luque Laguna et al., 2020 ; Magnotta et al., 2012 ; Pfefferbaum et al., 2003 ; Vollmar et al., 2010 ). Although CV is quite simply defined as the standard deviation of measurements divided by the mean, when used to assess reproducibility across individuals and sites, separating the variability into within-subject and between-subject variability is valuable for differentiating site-related variability (i.e., reproducibility) from between-subject differences ( Shoukri et al., 2008 ). However, even when computing CV ws , studies will often used a pooled (grand) mean ( Kurokawa et al., 2021 ; Luque Laguna et al., 2020 ), which still captures between subject differences. Here, we implement a robust measure of CV ws , as recommended by the quantitative imaging biomarkers alliance (QIBA) ( Shukla-Dave et al., 2019 ), and advocate for future studies to similarly quantify reproducibility in this way. We also echo previous recommendations to assess both reproducibility (generally indexed with within-subject variability) and reliability (indexed with ICC) in such studies, as they reflect separate properties ( Shoukri et al., 2008 ; Luque Laguna et al., 2020 ). 4.2 Limitations and future directions To our knowledge, this is the first study to probe the reproducibility and reliability of fixel-based analysis (FBA) measures, making use of a travelling heads cohort. Here, we were particularly interested in assessing scanner-related variability across local neuroimaging research centres, and as such, focused on these inter-scanner differences while making use of an identical harmonised DWI protocol across sites (that is, we focused on reproducibility rather than repeatability ( Plesser, 2018 ; Raunig et al., 2015 )). We were unable to assess inter-session (test-retest) repeatability, as well as potential protocol-related variability in fixel-based measures. As has been previously demonstrated (Genc, Tax, et al., 2020), we note the b -value dependence of the apparent fibre density (FD) measure, highlighting that protocol-related variability may be much more pronounced than scanner-related differences. We also note that the use of multi-shell versus single-shell data and subsequent FOD modelling could result in further variability in derived metrics. To this end, we hope that the release of the TRAMFIX dataset will enable researchers to additionally characterise protocol-related variability for fixel-based measures. Another limitation to our work was the inclusion of scanners from the same vendor. Previous work has highlighted that vendor-related variability in diffusion-derived features may be much greater than that from different scanner models of the same vendor ( Andica et al., 2020 ; Schilling et al., 2021 ). Of note, even with identical protocols and scanner model (Scanner 1 and 4), we observed some variability in derived measures, as has been previously reported for DTI-based metrics. However, we note that future work probing the impact of different vendors on fixel-based measures may also be valuable. Finally, while we focused in this study on assessing the reproducibility and reliability of fixel-based analysis metrics, to bring FBA studies into multi-site designs, future work evaluating the performance of different harmonisation approaches may be valuable. There are now a range of potential diffusion harmonisation approaches available ( Pinto et al., 2020 ; Tax et al., 2019 ), including those that harmonise on the diffusion signal ( Mirzaalian et al., 2016 ), and batch harmonisation approaches for derived metrics ( Fortin et al., 2017 ). These may be particularly pertinent when performing whole-brain analyses rather than examining tract-averaged metrics, given the variable reliability across fixels. The extent to which different harmonisation approaches perform on fixel-based measures remains only briefly characterised ( Mito et al., 2023 ). Travelling subject datasets can be valuable in this way, as they enable biological variability to be kept at a minimum, while assessing the performance of different harmonisation approaches. We hope that the TRAMFIX dataset will be a valuable resource to future studies that may wish to assess the performance of harmonisation techniques, and provide this as a resource alongside other travelling or test-retest datasets available in the community ( Cai et al., 2021 ; Koller et al., 2021 ; Kurokawa et al., 2021 ; Tax et al., 2019 ). 4.3 Conclusion Here, we present a travelling heads data for fixel-based metrics – the TRAMFIX dataset (TRavelling Across Melbourne for FIXel-based analysis). We demonstrate high reliability and reproducibility of fixel-based measures across scanners from the same vendor and sites, when using identical protocols, with similar results to that reported from other DWI measures. Future studies that examine the impact of different protocols as well as performance of data harmonisation approaches are likely to be valuable. To this end, we provide the TRAMFIX dataset as a resource for future work. These findings may help to guide large multi-site study designs, and improve the reach of more advanced diffusion approaches like fixel-based analysis. Data and Code Availability The code used to process DWI data, and to perform statistical analyses are available on GitHub ( https://github.com/remikamito/TRAMFIX ). Author Contributions Conceptualisation: RM, SG Methodology: RM, RS Data curation: RM, JH, MK, CK, RM Investigation: RM, SG, JYMY, RS Formal analysis: RM Software: RM, RS Writing – Original Draft: RM Writing – Review & Editing: RM, SG, JH, JYMY, JDT, MAD, RS, AZ Funding acquisition: RM Funding - Funder: Australian Research Council (ARC) - Award ID: DE240101035 - Principal Award Recipient: Remika Mito Declaration of Competing Interests The authors report no competing interests. Acknowledgments We are very grateful to the participants for volunteering their time for this study. The authors acknowledge the facilities and scientific and technical assistance of the National Imaging Facility (NIF), a National Collaborative Research Infrastructure Strategy (NCRIS) capability. The authors also thank the radiographers and facility staff at each of the research sites for their expertise and support: the Florey MRI facility, Monash Biomedical Imaging, and Royal Children’s Hospital. We are also very grateful to Dr Michelle Wu for her assistance in the study. R.M. is the recipient of an Australian Research Council Discovery Early Career Researcher Award (project number: DE240101035) and University of Melbourne Establishment Grant. S.G. and J.Y-M.Y. acknowledge funding from the Royal Children’s Hospital Foundation (RCHF 2022-1402). S.G. acknowledges support from the Melbourne Children’s Clinician-Scientist Fellowship. J.Y-M.Y. acknowledges support from The Kids’ Cancer Project (TKCP) Col Reynolds Fellowship. JDT is supported by core funding from the Wellcome/EPSRC Centre for Medical Engineering [WT203148/Z/16/Z] and by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Guy’s and St Thomas’ NHS Foundation Trust and King’s College London and/or the NIHR Clinical Research Facility. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care. A.Z. is supported by an Australian Research Council Future Fellowship and Rebecca L. Cooper Foundation Fellowship. Funder Information Declared Australian Research Council, https://ror.org/05mmh0f86 , DE240101035 References ↵ Andersson , J. L. R. , Skare , S. , & Ashburner , J . ( 2003 ). How to correct susceptibility distortions in spin-echo echo-planar images: Application to diffusion tensor imaging . NeuroImage , 20 ( 2 ), 870 – 888 . doi: 10.1016/S1053-8119(03)00336-7 OpenUrl CrossRef PubMed Web of Science ↵ Andersson , J. L. R. , & Sotiropoulos , S. N . ( 2016 ). An integrated approach to correction for off-resonance effects and subject movement in diffusion MR imaging . NeuroImage , 125 , 1063 – 1078 . doi: 10.1016/j.neuroimage.2015.10.019 OpenUrl CrossRef PubMed ↵ Andica , C. , Kamagata , K. , Hayashi , T. , Hagiwara , A. , Uchida , W. , Saito , Y. , Kamiya , K. , Fujita , S. , Akashi , T. , Wada , A. , Abe , M. , Kusahara , H. , Hori , M. , & Aoki , S . ( 2020 ). Scan–rescan and inter-vendor reproducibility of neurite orientation dispersion and density imaging metrics . Neuroradiology , 62 ( 4 ), 483 – 494 . doi: 10.1007/s00234-019-02350-6 OpenUrl CrossRef PubMed ↵ Basser , P. J. , Mattiello , J. , & LeBihan , D . ( 1994 ). MR diffusion tensor spectroscopy and imaging . Biophysical Journal , 66 ( 1 ), 259 – 267 . doi: 10.1016/S0006-3495(94)80775-1 OpenUrl CrossRef PubMed Web of Science ↵ Bastiani , M. , Cottaar , M. , Fitzgibbon , S. P. , Suri , S. , Alfaro-Almagro , F. , Sotiropoulos , S. N. , Jbabdi , S. , & Andersson , J. L. R . ( 2019 ). Automated quality control for within and between studies diffusion MRI data using a non-parametric framework for movement and distortion correction . NeuroImage , 184 , 801 – 812 . doi: 10.1016/j.neuroimage.2018.09.073 OpenUrl CrossRef PubMed ↵ Besseling , R. M. H. , Jansen , J. F. A. , Overvliet , G. M. , Vaessen , M. J. , Braakman , H. M. H. , Hofman , P. A. M. , Aldenkamp , A. P. , & Backes , W. H . ( 2012 ). Tract Specific Reproducibility of Tractography Based Morphology and Diffusion Metrics . PLOS ONE , 7 ( 4 ), e34125 . doi: 10.1371/journal.pone.0034125 OpenUrl CrossRef PubMed ↵ Bliese , P. D . ( 1998 ). Group Size, ICC Values, and Group-Level Correlations: A Simulation . Organizational Research Methods , 1 ( 4 ), 355 – 373 . doi: 10.1177/109442819814001 OpenUrl CrossRef Web of Science ↵ Cai , L. Y. , Yang , Q. , Kanakaraj , P. , Nath , V. , Newton , A. T. , Edmonson , H. A. , Luci , J. , Conrad , B. N. , Price , G. R. , Hansen , C. B. , Kerley , C. I. , Ramadass , K. , Yeh , F.-C. , Kang , H. , Garyfallidis , E. , Descoteaux , M. , Rheault , F. , Schilling , K. G. , & Landman , B. A . ( 2021 ). MASiVar: Multisite, multiscanner, and multisubject acquisitions for studying variability in diffusion weighted MRI . Magnetic Resonance in Medicine , 86 ( 6 ), 3304 – 3320 . doi: 10.1002/mrm.28926 OpenUrl CrossRef PubMed ↵ Dell’Acqua , F. , & Tournier , J.-D . ( 2019 ). Modelling white matter with spherical deconvolution: How and why? NMR in Biomedicine , 32 ( 4 ), e3945 . doi: 10.1002/nbm.3945 OpenUrl CrossRef PubMed ↵ Dhollander , T. , Clemente , A. , Singh , M. , Boonstra , F. , Civier , O. , Duque , J. D. , Egorova , N. , Enticott , P. , Fuelscher , I. , Gajamange , S. , Genc , S. , Gottlieb , E. , Hyde , C. , Imms , P. , Kelly , C. , Kirkovski , M. , Kolbe , S. , Liang , X. , Malhotra , A. , … Caeyenberghs , K . ( 2021 ). Fixel-based Analysis of Diffusion MRI: Methods, Applications, Challenges and Opportunities . NeuroImage , 241 , 118417 . doi: 10.1016/j.neuroimage.2021.118417 OpenUrl CrossRef PubMed ↵ Dhollander , T. , & Connelly , A . ( 2016 ). A novel iterative approach to reap the benefits of multi-tissue CSD from just single-shell (+ b= 0) diffusion MRI data . 24 , 3010 . OpenUrl ↵ Dhollander , T. , Mito , R. , Raffelt , D. , & Connelly , A. ( 2019 ). Improved white matter response function estimation for 3-tissue constrained spherical deconvolution . ↵ Fan , Q. , Polackal , M. N. , Tian , Q. , Ngamsombat , C. , Nummenmaa , A. , Witzel , T. , Klawiter , E. C. , & Huang , S. Y . ( 2021 ). Scan-rescan repeatability of axonal imaging metrics using high-gradient diffusion MRI and statistical implications for study design . NeuroImage , 240 , 118323 . doi: 10.1016/j.neuroimage.2021.118323 OpenUrl CrossRef ↵ Fortin , J.-P. , Parker , D. , Tunç , B. , Watanabe , T. , Elliott , M. A. , Ruparel , K. , Roalf , D. R. , Satterthwaite , T. D. , Gur , R. C. , Gur , R. E. , Schultz , R. T. , Verma , R. , & Shinohara , R. T . ( 2017 ). Harmonization of multi-site diffusion tensor imaging data . NeuroImage , 161 , 149 – 170 . doi: 10.1016/j.neuroimage.2017.08.047 OpenUrl CrossRef PubMed Genc , S. , Malpas , C. B. , Gulenc , A. , Sciberras , E. , Efron , D. , Silk , T. J. , & Seal , M. L . ( 2020 ). Longitudinal patterns of white matter fibre density and morphology in children are associated with age and pubertal stage . Developmental Cognitive Neuroscience , 45 , 100853 . doi: 10.1016/j.dcn.2020.100853 OpenUrl CrossRef PubMed ↵ Genc , S. , Smith , R. E. , Malpas , C. B. , Anderson , V. , Nicholson , J. M. , Efron , D. , Sciberras , E. , Seal , M. L. , & Silk , T. J . ( 2018 ). Development of white matter fibre density and morphology over childhood: A longitudinal fixel-based analysis . NeuroImage , 183 , 666 – 676 . doi: 10.1016/j.neuroimage.2018.08.043 OpenUrl CrossRef PubMed Genc , S. , Tax , C. M. W. , Raven , E. P. , Chamberland , M. , Parker , G. D. , & Jones , D. K . ( 2020 ). Impact of b-value on estimates of apparent fibre density . Human Brain Mapping , 41 ( 10 ), 2583 – 2595 . doi: 10.1002/hbm.24964 OpenUrl CrossRef PubMed ↵ Grazioplene , R. G. , DeYoung , C. G. , Hampson , M. , Anticevic , A. , & Pittenger , C . ( 2022 ). Obsessive compulsive symptom dimensions are linked to altered white-matter microstructure in a community sample of youth . Translational Psychiatry , 12 ( 1 ), 328 . doi: 10.1038/s41398-022-02013-w OpenUrl CrossRef PubMed ↵ Grech-Sollars , M. , Hales , P. W. , Miyazaki , K. , Raschke , F. , Rodriguez , D. , Wilson , M. , Gill , S. K. , Banks , T. , Saunders , D. E. , Clayden , J. D. , Gwilliam , M. N. , Barrick , T. R. , Morgan , P. S. , Davies , N. P. , Rossiter , J. , Auer , D. P. , Grundy , R. , Leach , M. O. , Howe , F. A. , … Clark , C. A . ( 2015 ). Multi-centre reproducibility of diffusion MRI parameters for clinical sequences in the brain . NMR in Biomedicine , 28 ( 4 ), 468 – 485 . doi: 10.1002/nbm.3269 OpenUrl CrossRef PubMed ↵ Jansen , J. F. A. , Kooi , M. E. , Kessels , A. G. H. , Nicolay , K. , & Backes , W. H . ( 2007 ). Reproducibility of Quantitative Cerebral T2 Relaxometry, Diffusion Tensor Imaging, and 1H Magnetic Resonance Spectroscopy at 3.0 Tesla . Investigative Radiology , 42 ( 6 ), 327 . doi: 10.1097/01.rli.0000262757.10271.e5 OpenUrl CrossRef PubMed ↵ Jeurissen , B. , Tournier , J.-D. , Dhollander , T. , Connelly , A. , & Sijbers , J . ( 2014 ). Multi-tissue constrained spherical deconvolution for improved analysis of multi-shell diffusion MRI data . NeuroImage , 103 , 411 – 426 . doi: 10.1016/j.neuroimage.2014.07.061 OpenUrl CrossRef PubMed ↵ Jones , D. K. , Knösche , T. R. , & Turner , R . ( 2013 ). White matter integrity, fiber count, and other fallacies: The do’s and don’ts of diffusion MRI . NeuroImage , 73 , 239 – 254 . doi: 10.1016/j.neuroimage.2012.06.081 OpenUrl CrossRef PubMed Web of Science ↵ Kelley , S. , Plass , J. , Bender , A. R. , & Polk , T. A . ( 2021 ). Age-Related Differences in White Matter: Understanding Tensor-Based Results Using Fixel-Based Analysis . Cerebral Cortex , 31 ( 8 ), 3881 – 3898 . doi: 10.1093/cercor/bhab056 OpenUrl CrossRef PubMed ↵ Kellner , E. , Dhital , B. , Kiselev , V. G. , & Reisert , M . ( 2016 ). Gibbs-ringing artifact removal based on local subvoxel-shifts . Magnetic Resonance in Medicine , 76 ( 5 ), 1574 – 1581 . doi: 10.1002/mrm.26054 OpenUrl CrossRef PubMed ↵ Koller , K. , Rudrapatna , U. , Chamberland , M. , Raven , E. P. , Parker , G. D. , Tax , C. M. W. , Drakesmith , M. , Fasano , F. , Owen , D. , Hughes , G. , Charron , C. , Evans , C. J. , & Jones , D. K . ( 2021 ). MICRA: Microstructural image compilation with repeated acquisitions . NeuroImage , 225 , 117406 . doi: 10.1016/j.neuroimage.2020.117406 OpenUrl CrossRef PubMed ↵ Koo , T. K. , & Li , M. Y . ( 2016 ). A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research . Journal of Chiropractic Medicine , 15 ( 2 ), 155 – 163 . doi: 10.1016/j.jcm.2016.02.012 OpenUrl CrossRef PubMed ↵ Kristensen , T. D. , Ambrosen , K. S. , Raghava , J. M. , Syeda , W. T. , Dhollander , T. , Lemvigh , C. K. , Bojesen , K. B. , Barber , A. D. , Nielsen , M. Ø. , Rostrup , E. , Pantelis , C. , Fagerlund , B. , Glenthøj , B. Y. , & Ebdrup , B. H . ( 2024 ). Structural and functional connectivity in relation to executive functions in antipsychotic-naïve patients with first episode schizophrenia . Schizophrenia , 10 ( 1 ), 72 . doi: 10.1038/s41537-024-00487-9 OpenUrl CrossRef ↵ Kristensen , T. D. , Raghava , J. M. , Skjerbæk , M. W. , Dhollander , T. , Syeda , W. , Ambrosen , K. S. , Bojesen , K. B. , Nielsen , M. Ø. , Pantelis , C. , Glenthøj , B. Y. , & Ebdrup , B. H . ( 2023 ). Fibre density and fibre-bundle cross-section of the corticospinal tract are distinctly linked to psychosis-specific symptoms in antipsychotic-naïve patients with first-episode schizophrenia . European Archives of Psychiatry and Clinical Neuroscience , 273 ( 8 ), 1797 – 1812 . doi: 10.1007/s00406-023-01598-7 OpenUrl CrossRef PubMed ↵ Kurokawa , R. , Kamiya , K. , Koike , S. , Nakaya , M. , Uematsu , A. , Tanaka , S. C. , Kamagata , K. , Okada , N. , Morita , K. , Kasai , K. , & Abe , O . ( 2021 ). Cross-scanner reproducibility and harmonization of a diffusion MRI structural brain network: A traveling subject study of multi-b acquisition . NeuroImage , 245 , 118675 . doi: 10.1016/j.neuroimage.2021.118675 OpenUrl CrossRef PubMed ↵ Luque Laguna , P. A. , Combes , A. J. E. , Streffer , J. , Einstein , S. , Timmers , M. , Williams , S. C. R. , & Dell’Acqua , F. ( 2020 ). Reproducibility, reliability and variability of FA and MD in the older healthy population: A test-retest multiparametric analysis . NeuroImage: Clinical , 26 , 102168 . doi: 10.1016/j.nicl.2020.102168 OpenUrl CrossRef PubMed ↵ Lyon , M. , Welton , T. , Varda , A. , Maller , J. J. , Broadhouse , K. , Korgaonkar , M. S. , Koslow , S. H. , Williams , L. M. , Gordon , E. , Rush , A. J. , & Grieve , S. M . ( 2019 ). Gender-specific structural abnormalities in major depressive disorder revealed by fixel-based analysis . NeuroImage: Clinical , 21 , 101668 . doi: 10.1016/j.nicl.2019.101668 OpenUrl CrossRef PubMed ↵ Magnotta , V. A. , Matsui , J. T. , Liu , D. , Johnson , H. J. , Long , J. D. , Bolster , B. D. , Mueller , B. A. , Lim , K. , Mori , S. , Helmer , K. G. , Turner , J. A. , Reading , S. , Lowe , M. J. , Aylward , E. , Flashman , L. A. , Bonett , G. , & Paulsen , J. S . ( 2012 ). MultiCenter Reliability of Diffusion Tensor Imaging . Brain Connectivity , 2 ( 6 ), 345 – 355 . doi: 10.1089/brain.2012.0112 OpenUrl CrossRef PubMed ↵ Mirzaalian , H. , Ning , L. , Savadjiev , P. , Pasternak , O. , Bouix , S. , Michailovich , O. , Grant , G. , Marx , C. E. , Morey , R. A. , Flashman , L. A. , George , M. S. , McAllister , T. W. , Andaluz , N. , Shutter , L. , Coimbra , R. , Zafonte , R. D. , Coleman , M. J. , Kubicki , M. , Westin , C. F. , … Rathi , Y . ( 2016 ). Inter-site and inter-scanner diffusion MRI data harmonization . NeuroImage , 135 , 311 – 323 . doi: 10.1016/j.neuroimage.2016.04.041 OpenUrl CrossRef PubMed ↵ Mito , R. , Pardoe , H. , Smith , R. , Tournier , J.-D. , Vaughan , D. , Pedersen , M. , & Jackson , G . ( 2023 ). ComBat scanner harmonization for fixel-based analysis . Proc Int Soc Mag Res Med , 3243 . ↵ Mito , R. , Pedersen , M. , Pardoe , H. , Parker , D. , Smith , R. E. , Cameron , J. , Scheffer , I. E. , Berkovic , S. F. , Vaughan , D. N. , & Jackson , G. D . ( 2024 ). Exploring individual fixel-based white matter abnormalities in epilepsy . Brain Communications , 6 ( 1 ), fcad352 . doi: 10.1093/braincomms/fcad352 OpenUrl CrossRef ↵ Mito , R. , Raffelt , D. , Dhollander , T. , Vaughan , D. N. , Tournier , J.-D. , Salvado , O. , Brodtmann , A. , Rowe , C. C. , Villemagne , V. L. , & Connelly , A . ( 2018 ). Fibre-specific white matter reductions in Alzheimer’s disease and mild cognitive impairment . Brain , 141 ( 3 ), 888 – 902 . doi: 10.1093/brain/awx355 OpenUrl CrossRef PubMed ↵ Mito , R. , Vaughan , D. N. , Semmelroch , M. , Connelly , A. , & Jackson , G. D . ( 2022 ). Bilateral Structural Network Abnormalities in Epilepsy Associated With Bottom-of-Sulcus Dysplasia . Neurology , 98 ( 2 ), e152 – e163 . doi: 10.1212/WNL.0000000000013006 OpenUrl CrossRef PubMed ↵ Palacios , E. M. , Martin , A. J. , Boss , M. A. , Ezekiel , F. , Chang , Y. S. , Yuh , E. L. , Vassar , M. J. , Schnyer , D. M. , MacDonald , C. L. , Crawford , K. L. , Irimia , A. , Toga , A. W. , & Mukherjee , P . ( 2017 ). Toward Precision and Reproducibility of Diffusion Tensor Imaging: A Multicenter Diffusion Phantom and Traveling Volunteer Study . American Journal of Neuroradiology , 38 ( 3 ), 537 – 545 . doi: 10.3174/ajnr.A5025 OpenUrl Abstract / FREE Full Text ↵ Pfefferbaum , A. , Adalsteinsson , E. , & Sullivan , E. V . ( 2003 ). Replicability of diffusion tensor imaging measurements of fractional anisotropy and trace in brain . Journal of Magnetic Resonance Imaging , 18 ( 4 ), 427 – 433 . doi: 10.1002/jmri.10377 OpenUrl CrossRef PubMed Web of Science ↵ Pinto , M. S. , Paolella , R. , Billiet , T. , Van Dyck , P. , Guns , P.-J. , Jeurissen , B. , Ribbens , A. , den Dekker , A. J. , & Sijbers , J. ( 2020 ). Harmonization of Brain Diffusion MRI: Concepts and Methods . Frontiers in Neuroscience , 14 . doi: 10.3389/fnins.2020.00396 OpenUrl CrossRef PubMed ↵ Plesser , H. E . ( 2018 ). Reproducibility vs. Replicability: A Brief History of a Confused Terminology . Frontiers in Neuroinformatics , 11 . doi: 10.3389/fninf.2017.00076 OpenUrl CrossRef ↵ Raffelt , D. A. , Smith , R. E. , Ridgway , G. R. , Tournier , J.-D. , Vaughan , D. N. , Rose , S. , Henderson , R. , & Connelly , A . ( 2015 ). Connectivity-based fixel enhancement: Whole-brain statistical analysis of diffusion MRI measures in the presence of crossing fibres . Neuroimage , 117 , 40 – 55 . OpenUrl CrossRef PubMed ↵ Raffelt , D. A. , Tournier , J.-D. , Smith , R. E. , Vaughan , D. N. , Jackson , G. , Ridgway , G. R. , & Connelly , A . ( 2017 ). Investigating white matter fibre density and morphology using fixel-based analysis . NeuroImage , 144 , 58 – 73 . doi: 10.1016/j.neuroimage.2016.09.029 OpenUrl CrossRef PubMed ↵ Raffelt , D. , Tournier , J.-D. , Crozier , S. , Connelly , A. , & Salvado , O . ( 2012 ). Reorientation of fiber orientation distributions using apodized point spread functions . Magnetic Resonance in Medicine , 67 ( 3 ), 844 – 855 . doi: 10.1002/mrm.23058 OpenUrl CrossRef PubMed ↵ Raffelt , D. , Tournier , J.-D. , Fripp , J. , Crozier , S. , Connelly , A. , & Salvado , O . ( 2011 ). Symmetric diffeomorphic registration of fibre orientation distributions . Neuroimage , 56 ( 3 ), 1171 – 1180 . OpenUrl CrossRef PubMed Raffelt , D. , Tournier , J.-D. , Rose , S. , Ridgway , G. R. , Henderson , R. , Crozier , S. , Salvado , O. , & Connelly , A . ( 2012 ). Apparent Fibre Density: A novel measure for the analysis of diffusion-weighted magnetic resonance images . NeuroImage , 59 ( 4 ), 3976 – 3994 . doi: 10.1016/j.neuroimage.2011.10.045 OpenUrl CrossRef PubMed ↵ Raunig , D. L. , McShane , L. M. , Pennello , G. , Gatsonis , C. , Carson , P. L. , Voyvodic , J. T. , Wahl , R. L. , Kurland , B. F. , Schwarz , A. J. , Gönen , M. , Zahlmann , G. , Kondratovich , M. V. , O’Donnell , K. , Petrick , N. , Cole , P. E. , Garra , B. , & Sullivan , D. C . ( 2015 ). Quantitative imaging biomarkers: A review of statistical methods for technical performance assessment . Statistical Methods in Medical Research , 24 ( 1 ), 27 – 67 . doi: 10.1177/0962280214537344 OpenUrl CrossRef PubMed ↵ Schilling , K. G. , Tax , C. M. W. , Rheault , F. , Hansen , C. , Yang , Q. , Yeh , F.-C. , Cai , L. , Anderson , A. W. , & Landman , B. A . ( 2021 ). Fiber tractography bundle segmentation depends on scanner effects, vendor effects, acquisition resolution, diffusion sampling scheme, diffusion sensitization, and bundle segmentation workflow . NeuroImage , 242 , 118451 . doi: 10.1016/j.neuroimage.2021.118451 OpenUrl CrossRef PubMed ↵ Shoukri , M. M. , Colak , D. , Kaya , N. , & Donner , A . ( 2008 ). Comparison of two dependent within subject coefficients of variation to evaluate the reproducibility of measurement devices . BMC Medical Research Methodology , 8 ( 1 ), 24 . doi: 10.1186/1471-2288-8-24 OpenUrl CrossRef PubMed ↵ Shukla-Dave , A. , Obuchowski , N. A. , Chenevert , T. L. , Jambawalikar , S. , Schwartz , L. H. , Malyarenko , D. , Huang , W. , Noworolski , S. M. , Young , R. J. , Shiroishi , M. S. , Kim , H. , Coolens , C. , Laue , H. , Chung , C. , Rosen , M. , Boss , M. , & Jackson , E. F . ( 2019 ). Quantitative imaging biomarkers alliance (QIBA) recommendations for improved precision of DWI and DCE-MRI derived biomarkers in multicenter oncology trials . Journal of Magnetic Resonance Imaging , 49 ( 7 ), e101 – e121 . doi: 10.1002/jmri.26518 OpenUrl CrossRef PubMed ↵ Smith , R. E. , Tournier , J.-D. , Calamante , F. , & Connelly , A . ( 2013 ). SIFT: Spherical-deconvolution informed filtering of tractograms . Neuroimage , 67 , 298 – 312 . OpenUrl CrossRef PubMed ↵ Tax , C. MW. , Grussu , F. , Kaden , E. , Ning , L. , Rudrapatna , U. , John Evans , C. , St-Jean , S. , Leemans , A. , Koppers , S. , Merhof , D. , Ghosh , A. , Tanno , R. , Alexander , D. C. , Zappalà , S. , Charron , C. , Kusmia , S. , Linden , D. EJ. , Jones , D. K. , & Veraart , J . ( 2019 ). Cross-scanner and cross-protocol diffusion MRI data harmonisation: A benchmark database and evaluation of algorithms . NeuroImage , 195 , 285 – 299 . doi: 10.1016/j.neuroimage.2019.01.077 OpenUrl CrossRef PubMed ↵ Tinney , E. M. , Warren , A. E. L. , Ai , M. , Morris , T. P. , O’Brien , A. , Odom , H. , Sutton , B. P. , Jain , S. , Kang , C. , Huang , H. , Wan , L. , Oberlin , L. , Burns , J. M. , Vidoni , E. D. , McAuley , E. , Kramer , A. F. , Erickson , K. I. , & Hillman , C. H . ( 2024 ). Understanding Cognitive Aging Through White Matter: A Fixel-Based Analysis . Human Brain Mapping , 45 ( 18 ), e70121 . doi: 10.1002/hbm.70121 OpenUrl CrossRef PubMed ↵ Tournier , J.-D. , Calamante , F. , & Connelly , A . ( 2007 ). Robust determination of the fibre orientation distribution in diffusion MRI: Non-negativity constrained super-resolved spherical deconvolution . NeuroImage , 35 ( 4 ), 1459 – 1472 . doi: 10.1016/j.neuroimage.2007.02.016 OpenUrl CrossRef PubMed Web of Science ↵ Tournier , J.-D. , Christiaens , D. , Hutter , J. , Price , A. N. , Cordero-Grande , L. , Hughes , E. , Bastiani , M. , Sotiropoulos , S. N. , Smith , S. M. , Rueckert , D. , Counsell , S. J. , Edwards , A. D. , & Hajnal , J. V . ( 2020 ). A data-driven approach to optimising the encoding for multi-shell diffusion MRI with application to neonatal imaging . NMR in Biomedicine , 33 ( 9 ), e4348 . doi: 10.1002/nbm.4348 OpenUrl CrossRef PubMed ↵ Tournier , J.-D. , Smith , R. , Raffelt , D. , Tabbara , R. , Dhollander , T. , Pietsch , M. , Christiaens , D. , Jeurissen , B. , Yeh , C.-H. , & Connelly , A . ( 2019 ). MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation . Neuroimage , 202 , 116137 . OpenUrl CrossRef PubMed ↵ Tustison , N. J. , Avants , B. B. , Cook , P. A. , Zheng , Y. , Egan , A. , Yushkevich , P. A. , & Gee , J. C . ( 2010 ). N4ITK: Improved N3 Bias Correction . IEEE Transactions on Medical Imaging , 29 ( 6 ), 1310 – 1320 . doi: 10.1109/TMI.2010.2046908 OpenUrl CrossRef PubMed Web of Science ↵ Vaughan , D. N. , Raffelt , D. , Curwood , E. , Tsai , M.-H. , Tournier , J.-D. , Connelly , A. , & Jackson , G. D . ( 2017 ). Tract-specific atrophy in focal epilepsy: Disease, genetics, or seizures? Annals of Neurology , 81 ( 2 ), 240 – 250 . doi: 10.1002/ana.24848 OpenUrl CrossRef PubMed ↵ Veraart , J. , Novikov , D. S. , Christiaens , D. , Ades-aron , B. , Sijbers , J. , & Fieremans , E . ( 2016 ). Denoising of diffusion MRI using random matrix theory . NeuroImage , 142 , 394 – 406 . doi: 10.1016/j.neuroimage.2016.08.016 OpenUrl CrossRef PubMed ↵ Veraart , J. , Raven , E. P. , Edwards , L. J. , Weiskopf , N. , & Jones , D. K . ( 2021 ). The variability of MR axon radii estimates in the human white matter . Human Brain Mapping , 42 ( 7 ), 2201 – 2213 . doi: 10.1002/hbm.25359 OpenUrl CrossRef PubMed ↵ Veraart , J. , Sijbers , J. , Sunaert , S. , Leemans , A. , & Jeurissen , B . ( 2013 ). Weighted linear least squares estimation of diffusion MRI parameters: Strengths, limitations, and pitfalls . NeuroImage , 81 , 335 – 346 . doi: 10.1016/j.neuroimage.2013.05.028 OpenUrl CrossRef PubMed ↵ Vollmar , C. , O’Muircheartaigh , J. , Barker , G. J. , Symms , M. R. , Thompson , P. , Kumari , V. , Duncan , J. S. , Richardson , M. P. , & Koepp , M. J . ( 2010 ). Identical, but not the same: Intra-site and inter-site reproducibility of fractional anisotropy measures on two 3.0 T scanners . NeuroImage , 51 ( 4 ), 1384 – 1394 . doi: 10.1016/j.neuroimage.2010.03.046 OpenUrl CrossRef PubMed Web of Science ↵ Wasserthal , J. , Neher , P. , & Maier-Hein , K. H . ( 2018 ). TractSeg-Fast and accurate white matter tract segmentation . NeuroImage , 183 , 239 – 253 . OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted July 25, 2025. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following TRAMFIX: TRavelling Across Melbourne for FIXel-based analysis (a reproducibility study) Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share TRAMFIX: TRavelling Across Melbourne for FIXel-based analysis (a reproducibility study) Remika Mito , Sila Genc , Jocelyn Halim , Joseph Yuan-Mou Yang , Jacques-Donald Tournier , Michael Kean , Chris Kokkinos , Richard McIntyre , Maria A. Di Biase , Robert Smith , Andrew Zalesky bioRxiv 2025.07.21.666058; doi: https://doi.org/10.1101/2025.07.21.666058 Share This Article: Copy Citation Tools TRAMFIX: TRavelling Across Melbourne for FIXel-based analysis (a reproducibility study) Remika Mito , Sila Genc , Jocelyn Halim , Joseph Yuan-Mou Yang , Jacques-Donald Tournier , Michael Kean , Chris Kokkinos , Richard McIntyre , Maria A. Di Biase , Robert Smith , Andrew Zalesky bioRxiv 2025.07.21.666058; doi: https://doi.org/10.1101/2025.07.21.666058 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Neuroscience Subject Areas All Articles Animal Behavior and Cognition (7629) Biochemistry (17660) Bioengineering (13881) Bioinformatics (41912) Biophysics (21436) Cancer Biology (18578) Cell Biology (25482) Clinical Trials (138) Developmental Biology (13372) Ecology (19889) Epidemiology (2067) Evolutionary Biology (24302) Genetics (15599) Genomics (22483) Immunology (17728) Microbiology (40365) Molecular Biology (17163) Neuroscience (88540) Paleontology (666) Pathology (2830) Pharmacology and Toxicology (4821) Physiology (7637) Plant Biology (15130) Scientific Communication and Education (2045) Synthetic Biology (4290) Systems Biology (9818) Zoology (2269)