Dataset Information

Tractography dissection variability: What happens when 42 groups dissect 14 white matter bundles on the same dataset?

ABSTRACT: White matter bundle segmentation using diffusion MRI fiber tractography has become the method of choice to identify white matter fiber pathways in vivo in human brains. However, like other analyses of complex data, there is considerable variability in segmentation protocols and techniques. This can result in different reconstructions of the same intended white matter pathways, which directly affects tractography results, quantification, and interpretation. In this study, we aim to evaluate and quantify the variability that arises from different protocols for bundle segmentation. Through an open call to users of fiber tractography, including anatomists, clinicians, and algorithm developers, 42 independent teams were given processed sets of human whole-brain streamlines and asked to segment 14 white matter fascicles on six subjects. In total, we received 57 different bundle segmentation protocols, which enabled detailed volume-based and streamline-based analyses of agreement and disagreement among protocols for each fiber pathway. Results show that even when given the exact same sets of underlying streamlines, the variability across protocols for bundle segmentation is greater than all other sources of variability in the virtual dissection process, including variability within protocols and variability across subjects. In order to foster the use of tractography bundle dissection in routine clinical settings, and as a fundamental analytical tool, future endeavors must aim to resolve and reduce this heterogeneity. Although external validation is needed to verify the anatomical accuracy of bundle dissections, reducing heterogeneity is a step towards reproducible research and may be achieved through the use of standard nomenclature and definitions of white matter bundles and well-chosen constraints and decisions in the dissection process.

SUBMITTER: Schilling KG

PROVIDER: S-EPMC8855321 | biostudies-literature | 2021 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Tractography dissection variability: What happens when 42 groups dissect 14 white matter bundles on the same dataset?

Schilling Kurt G KG Rheault François F Petit Laurent L Hansen Colin B CB Nath Vishwesh V Yeh Fang-Cheng FC Girard Gabriel G Barakovic Muhamed M Rafael-Patino Jonathan J Yu Thomas T Fischi-Gomez Elda E Pizzolato Marco M Ocampo-Pineda Mario M Schiavi Simona S Canales-Rodríguez Erick J EJ Daducci Alessandro A Granziera Cristina C Innocenti Giorgio G Thiran Jean-Philippe JP Mancini Laura L Wastling Stephen S Cocozza Sirio S Petracca Maria M Pontillo Giuseppe G Mancini Matteo M Vos Sjoerd B SB Vakharia Vejay N VN Duncan John S JS Melero Helena H Manzanedo Lidia L Sanz-Morales Emilio E Peña-Melián Ángel Á Calamante Fernando F Attyé Arnaud A Cabeen Ryan P RP Korobova Laura L Toga Arthur W AW Vijayakumari Anupa Ambili AA Parker Drew D Verma Ragini R Radwan Ahmed A Sunaert Stefan S Emsell Louise L De Luca Alberto A Leemans Alexander A Bajada Claude J CJ Haroon Hamied H Azadbakht Hojjatollah H Chamberland Maxime M Genc Sila S Tax Chantal M W CMW Yeh Ping-Hong PH Srikanchana Rujirutana R Mcknight Colin D CD Yang Joseph Yuan-Mou JY Chen Jian J Kelly Claire E CE Yeh Chun-Hung CH Cochereau Jerome J Maller Jerome J JJ Welton Thomas T Almairac Fabien F Seunarine Kiran K KK Clark Chris A CA Zhang Fan F Makris Nikos N Golby Alexandra A Rathi Yogesh Y O'Donnell Lauren J LJ Xia Yihao Y Aydogan Dogu Baran DB Shi Yonggang Y Fernandes Francisco Guerreiro FG Raemaekers Mathijs M Warrington Shaun S Michielse Stijn S Ramírez-Manzanares Alonso A Concha Luis L Aranda Ramón R Meraz Mariano Rivera MR Lerma-Usabiaga Garikoitz G Roitman Lucas L Fekonja Lucius S LS Calarco Navona N Joseph Michael M Nakua Hajer H Voineskos Aristotle N AN Karan Philippe P Grenier Gabrielle G Legarreta Jon Haitz JH Adluru Nagesh N Nair Veena A VA Prabhakaran Vivek V Alexander Andrew L AL Kamagata Koji K Saito Yuya Y Uchida Wataru W Andica Christina C Abe Masahiro M Bayrak Roza G RG Wheeler-Kingshott Claudia A M Gandini CAMG D'Angelo Egidio E Palesi Fulvia F Savini Giovanni G Rolandi Nicolò N Guevara Pamela P Houenou Josselin J López-López Narciso N Mangin Jean-François JF Poupon Cyril C Román Claudio C Vázquez Andrea A Maffei Chiara C Arantes Mavilde M Andrade José Paulo JP Silva Susana Maria SM Calhoun Vince D VD Caverzasi Eduardo E Sacco Simone S Lauricella Michael M Pestilli Franco F Bullock Daniel D Zhan Yang Y Brignoni-Perez Edith E Lebel Catherine C Reynolds Jess E JE Nestrasil Igor I Labounek René R Lenglet Christophe C Paulson Amy A Aulicka Stefania S Heilbronner Sarah R SR Heuer Katja K Chandio Bramsh Qamar BQ Guaje Javier J Tang Wei W Garyfallidis Eleftherios E Raja Rajikha R Anderson Adam W AW Landman Bennett A BA Descoteaux Maxime M

NeuroImage 20210822

White matter bundle segmentation using diffusion MRI fiber tractography has become the method of choice to identify white matter fiber pathways in vivo in human brains. However, like other analyses of complex data, there is considerable variability in segmentation protocols and techniques. This can result in different reconstructions of the same intended white matter pathways, which directly affects tractography results, quantification, and interpretation. In this study, we aim to evaluate and q ...[more]

PMID: 34433094

Similar Datasets

Project description:The segmentation of brain structures is a key component of many neuroimaging studies. Consistent anatomical definitions are crucial to ensure consensus on the position and shape of brain structures, but segmentations are prone to variation in their interpretation and execution. White-matter (WM) pathways are global structures of the brain defined by local landmarks, which leads to anatomical definitions being difficult to convey, learn, or teach. Moreover, the complex shape of WM pathways and their representation using tractography (streamlines) make the design and evaluation of dissection protocols difficult and time-consuming. The first iteration of Tractostorm quantified the variability of a pyramidal tract dissection protocol and compared results between experts in neuroanatomy and nonexperts. Despite virtual dissection being used for decades, in-depth investigations of how learning or practicing such protocols impact dissection results are nonexistent. To begin to fill the gap, we evaluate an online educational tractography course and investigate the impact learning and practicing a dissection protocol has on interrater (groupwise) reproducibility. To generate the required data to quantify reproducibility across raters and time, 20 independent raters performed dissections of three bundles of interest on five Human Connectome Project subjects, each with four timepoints. Our investigation shows that the dissection protocol in conjunction with an online course achieves a high level of reproducibility (between 0.85 and 0.90 for the voxel-based Dice score) for the three bundles of interest and remains stable over time (repetition of the protocol). Suggesting that once raters are familiar with the software and tasks at hand, their interpretation and execution at the group level do not drastically vary. When compared to previous work that used a different method of communication for the protocol, our results show that incorporating a virtual educational session increased reproducibility. Insights from this work may be used to improve the future design of WM pathway dissection protocols and to further inform neuroanatomical definitions.

Project description:Anatomic tracing is recognized as a critical source of knowledge on brain circuitry that can be used to assess the accuracy of diffusion MRI (dMRI) tractography. However, most prior studies that have performed such assessments have used dMRI and tracer data from different brains and/or have been limited in the scope of dMRI analysis methods allowed by the data. In this work, we perform a quantitative, voxel-wise comparison of dMRI tractography and anatomic tracing data in the same macaque brain. An ex vivo dMRI acquisition with high angular resolution and high maximum b-value allows us to compare a range of q-space sampling, orientation reconstruction, and tractography strategies. The availability of tracing in the same brain allows us to localize the sources of tractography errors and to identify axonal configurations that lead to such errors consistently, across dMRI acquisition and analysis strategies. We find that these common failure modes involve geometries such as branching or turning, which cannot be modeled well by crossing fibers. We also find that the default thresholds that are commonly used in tractography correspond to rather conservative, low-sensitivity operating points. While deterministic tractography tends to have higher sensitivity than probabilistic tractography in that very conservative threshold regime, the latter outperforms the former as the threshold is relaxed to avoid missing true anatomical connections. On the other hand, the q-space sampling scheme and maximum b-value have less of an impact on accuracy. Finally, using scans from a set of additional macaque brains, we show that there is enough inter-individual variability to warrant caution when dMRI and tracer data come from different animals, as is often the case in the tractography validation literature. Taken together, our results provide insights on the limitations of current tractography methods and on the critical role that anatomic tracing can play in identifying potential avenues for improvement.

Dataset Information

Tractography dissection variability: What happens when 42 groups dissect 14 white matter bundles on the same dataset?

Publications

Tractography dissection variability: What happens when 42 groups dissect 14 white matter bundles on the same dataset?

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets