CoChat
Sign Up
Cited by 70Advanced Neuroimaging Techniques and ApplicationsFunctional Brain Connectivity StudiesAdvanced MRI Techniques and Applications

Diffusion MRI-based cortical connectome reconstruction: dependency on tractography procedures and neuroanatomical characteristics

Open access
Sinke, Michel R. T., Otte, Willem M., Christiaens, Daan, Schmitt, Oliver, Leemans, Alexander, van der Toorn, Annette, Sarabdjitsingh, R. Angela, Joëls, Marian +1 more·2018·Brain Structure and Function
DOI 10.1007/s00429-018-1628-yPMC PMC5968063
Open this paper in CoChat — free accountOpen PDF

Abstract

Diffusion MRI (dMRI)-based tractography offers unique abilities to map whole-brain structural connections in human and animal brains. However, dMRI-based tractography indirectly measures white matter tracts, with suboptimal accuracy and reliability. Recently, sophisticated methods including constrained spherical deconvolution (CSD) and global tractography have been developed to improve tract reconstructions through modeling of more complex fiber orientations. Our study aimed to determine the accuracy of connectome reconstruction for three dMRI-based tractography approaches: diffusion tensor (DT)-based, CSD-based and global tractography. Therefore, we validated whole brain structural connectome reconstructions based on ten ultrahigh-resolution dMRI rat brain scans and 106 cortical regions, from which varying tractography parameters were compared against standardized neuronal tracer data. All tested tractography methods generated considerable numbers of false positive and false negative connections. There was a parameter range trade-off between sensitivity: 0.06-0.63 interhemispherically and 0.22-0.86 intrahemispherically; and specificity: 0.99-0.60 interhemispherically and 0.99-0.23 intrahemispherically. Furthermore, performance of all tractography methods decreased with increasing spatial distance between connected regions. Similar patterns and trade-offs were found, when we applied spherical deconvolution informed filtering of tractograms, streamline thresholding and group-based average network thresholding. Despite the potential of CSD-based and global tractography to handle complex fiber orientations at voxel level, reconstruction accuracy, especially for long-distance connections, remains a challenge. Hence, connectome reconstruction benefits from varying parameter settings and combination of tractography methods to account for anatomical variation of neuronal pathways.

AI snapshot

Written from the paper’s full text.

TL;DR

This study validated whole-brain diffusion MRI tractography connectomes in rat brain against neuronal tracer data and found that DT-based, CSD-based, and global tractography all produce substantial false positive and false negative connections. Reconstruction accuracy depends strongly on tractography method, parameter settings, hemisphere, and connection distance.

Key findings

  • Across all tested tractography methods and parameters, sensitivity ranged from 0.06–0.63 for interhemispheric and 0.22–0.86 for intrahemispheric connections, while specificity ranged from 0.99–0.60 interhemispherically and 0.99–0.23 intrahemispherically.
  • CSD-based tractography gave the highest sensitivity but lowest specificity; global tractography gave the lowest sensitivity but highest specificity; DT-based tractography generally had lower Jaccard index than CSD-based or global tractography.
  • Performance declined with increasing spatial distance between connected regions: sensitivity and Jaccard index decreased non-linearly, and specificity increased non-linearly, especially for long-distance connections; global tractography had particularly low sensitivity (<0.10) and Jaccard index (<0.05) for long-distance fibers.
  • SIFT, streamline thresholding, and group-based average network thresholding did not remove the sensitivity–specificity trade-off; SIFT decreased sensitivity, increased specificity, and decreased Jaccard index for both DT- and CSD-based tractography.

Methods

  • Ten healthy adult male Wistar rat brains were scanned postmortem with ultrahigh-resolution HARDI at 9.4 T (60 diffusion directions, b = 3842 s/mm², voxel resolution 150 × 148 × 150 µm³), and 106 bilateral cortical regions from the Paxinos and Watson atlas were used as connectome nodes.
  • DT-based, CSD-based, and global tractography were performed in MRtrix3 with varied parameters, including step size, FA/FOD threshold, angular threshold, streamline number, particle potential, and connection potential; SIFT, streamline thresholds, and group incidence thresholds were also tested.
  • Tractography-based connectivity was compared against a rat neuronal tracer database containing more than 450,237 ipsilateral and 175,654 contralateral connections from more than 6,183 publications, with tracer-based network density of 0.27; performance was measured by sensitivity, specificity, and Jaccard index, including by inter- vs. intrahemispheric connections and Euclidean distance categories.

Limitations

  • The paper states that mismatch with tracers may partly reflect incompleteness of neuroanatomical tracer data, that multi-synaptic connections not easily detected by tracing might be misclassified as false positives, and that this challenges neuronal tracing as a gold standard.
  • The paper states that subcortical regions were not assessed, biasing conclusions toward cortical connectomics; connectivity was defined only by streamline endpoints, ignoring three-dimensional pathway structure; and connection weights were not considered.
  • Evident from the text: validation was limited to postmortem rat brain and 106 cortical regions, so findings may not directly generalize to in vivo human connectomes or subcortical networks.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s00429-018-1628-y

Cites

75 of 75 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

22 of 41 cited sources checked · 11 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    At standard spatial resolutions around a few mm, multiple fiber orientations affect the diffusivity patterns in up to 90% of white matter voxels in the human brain.

    “At standard spatial resolutions around a few mm, multiple fiber orientations affect the diffusivity patterns in up to 90% of white matter voxels in the human brain (Jeurissen et al. 2013).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Jeurissen et al. 2013 source
  • Premisecitation

    DT-based deterministic tractography is particularly hampered by voxels containing multiple fiber orientations such as crossing, bending or fanning fibers.

    “This method is particularly hampered by voxels containing multiple fiber orientations such as crossing, bending or fanning fibers (Jbabdi and Johansen-Berg 2011).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Jbabdi and Johansen-Berg 2011 source
  • Premisecitation

    False positive tract reconstructions may lead to biased characterization of connectome topology.

    “This may lead to biased characterization of connectome topology, particularly in the presence of false positive tract reconstructions (Drakesmith et al. 2015; Zalesky et al. 2016).”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Drakesmith et al. 2015 source
    • cites Zalesky et al. 2016 source
  • Interpretationcitation

    The overall results confirm previously reported trade-offs between sensitivity and specificity of tract reconstructions across tractography methods.

    “As such, our overall results confirm previously reported trade-offs between sensitivity and specificity of tract reconstructions across tractography methods (Bastiani et al. 2012; Thomas et al. 2014; Chen et al. 2015; Donahue et al. 2016).”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites Bastiani et al. 2012 source
    • cites Thomas et al. 2014 source
    • cites Chen et al. 2015 source

      The Results (§3.1) explicitly report a sensitivity–specificity trade-off when tuning fiber-number thresholds and DTI tractography parameters, and say it agrees with previous findings. Chen et al. do not establish that trade-off across different tractography methods; their analysis concerns parameter choices within DTI tractography. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
  • Findingcitation

    The FA threshold had significantly more effect on connectome reconstruction sensitivity and specificity than the step size and angle threshold.

    “The FA threshold had significantly more effect on connectome reconstruction sensitivity and specificity than the step size and angle threshold, in line with previous research (Azadbakht et al. 2015; Chen et al. 2015).”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites Azadbakht et al. 2015 source

      The Results sections on FA termination and streamline curvature report that higher FA thresholds reduced connection accuracy, while curvature constraints had no clear effect. However, the paper does not report a statistical comparison showing a significantly greater effect on sensitivity and specificity, and the Discussion refers to a default step size rather than a tested step-size comparison. Page numbers are unavailable in our copy. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
    • cites Chen et al. 2015 source

      Results §3.1 says the FA threshold had more impact on the reconstructed connectome than the angular threshold, and describes a sensitivity–specificity tradeoff when parameters were varied. Methods §2.2 evaluates FA and angular thresholds, but not step size; the cited passages also do not establish that the difference was statistically significant. Page numbers are unavailable in our copy. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
  • May combine two claims. 47
  • Classified as probably not a substantial claim. 11
  • The quote only partly states this claim. 2

  • May combine two claims.

    Premisecitation

    Long-distance connections are more difficult to track, as reconstruction error accumulates with each tracking step.

    “Consequently, long-distance connections are more difficult to track, as the reconstruction error accumulates with each tracking step (Zalesky and Fornito 2009; Jbabdi et al. 2015; Reveley et al. 2015; Donahue et al. 2016).”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites Zalesky and Fornito 2009 source
    • cites Jbabdi et al. 2015 source

      Section 3.1 says tractography uses local orientation information to infer long-range connections, can take wrong turns, and is affected by the distance a pathway travels. It does not state that reconstruction error accumulates with each tracking step, so the claimed mechanism is not established by this source. Page numbers are unavailable in our copy. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
    • cites Reveley et al. 2015 source

      The Results report that track lines often become trapped in superficial white matter and fail to reach distant targets, supporting the difficulty of tracking long-range connections. The paper attributes that difficulty to local fiber structure; the passages do not establish that reconstruction error accumulates with each tracking step. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    The anatomical accuracy of diffusion-weighted tractography continues to be a matter of controversy, and reconstructing connectomes with both high sensitivity and high specificity remains challenging.

    “However, the anatomical accuracy of diffusion-weighted tractography, which infers neuronal pathways from tissue water diffusion direction, continues to be a matter of controversy, and reconstructing connectomes with both high sensitivity and high specificity remains challenging (Sporns et al. 2005; Bastiani et al. 2012; Jbabdi et al. 2013, 2015; Thomas et al. 2014; Azadbakht et al. 2015; Calabrese”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites Sporns et al. 2005 source

      In “Macroscale: Brain Regions and Pathways” (page unavailable), the paper says the relationship between diffusion-imaging results, fiber tracts, and neuronal connections remains controversial, and it describes limitations in acquisition and processing. It also says no single mapping method is likely to be sufficiently comprehensive. It does not specifically address the challenge of achieving both high sensitivity and high specificity in reconstructed connectomes. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites Bastiani et al. 2012 source
    • cites Jbabdi et al. 2013 source

      The Discussion (Multimodal comparative anatomy; page unavailable) identifies a tractography failure and explains that small fascicles embedded in grey matter are difficult to track. This supports concern about anatomical accuracy, but the paper does not establish the broader claim that reconstructing connectomes with both high sensitivity and high specificity remains challenging. (LLM, conf=0.84)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    CSD-based tractography solves some typical tractography errors, but false positive tract reconstructions may still arise from noisy peaks or from ambiguous fanning and bending fiber populations.

    “This solves some of the typical tractography errors, but false positive tract reconstructions may still arise from noisy (spurious) peaks or from ambiguous fanning and bending fiber populations (Jbabdi and Johansen-Berg 2011).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Jbabdi and Johansen-Berg 2011 source

      In “Low level modelling” (page unavailable), the paper says spherical-deconvolution fibre models have improved tractography, while unmodelled fanning can produce false positives. It also describes errors associated with sharp bends and noise-related uncertainty. It does not specifically identify noisy or spurious peaks as a cause of false-positive reconstructions, or establish the claimed improvement specifically for CSD-based tractography. (LLM, conf=0.83)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    The angle threshold might still have affected specific pathway reconstructions in the brain.

    “However, the angle threshold might still have affected specific pathway reconstructions in the brain (Thomas et al. 2014).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Thomas et al. 2014 source

  • May combine two claims.

    Interpretationown data

    The large number of invalid tract reconstructions observed stresses the relatively ill-posed nature of current tractography approaches and the need for methodological progression.

    “The large number of invalid tract reconstructions we observed in our study, in line with a recent study on human tract reconstructions (Maier-Hein et al. 2016), stresses the relatively ill-posed nature of current tractography approaches and the need for methodological progression.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Maier-Hein et al. 2016 source

  • May combine two claims.

    Premisenone given

    Most neuroanatomical connections are at least to a certain degree bidirectional and contain nested sub-regions with their own connections and corresponding weights.

    “While this was recently done to compare measures of connectivity strength between diffusion tractography and tract tracing data from macaque brain (van den Heuvel et al. 2015; Donahue et al. 2016), most neuroanatomical connections are at least to a certain degree bidirectional and contain nested sub-regions with their own connections and corresponding weights.”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites van den Heuvel et al. 2015 source
    • cites Donahue et al. 2016 source

  • May combine two claims.

    Findingown data

    CSD-based tractography led to the highest sensitivity for both inter- and intrahemispheric connections, but the lowest specificity, whereas global tractography resulted in the lowest sensitivity for both inter- and intrahemispheric connections, but the highest specificity.

    “CSD-based tractography led to the highest sensitivity for both inter- and intrahemispheric connections, but the lowest specificity, whereas global tractography resulted in the lowest sensitivity for both inter- and intrahemispheric connections, but the highest specificity.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    For DT-based tractography, at connectome level and over all sets of parameter settings, sensitivity varied from 0.22 ± 0.04 to 0.61 ± 0.02, and specificity from 0.60 ± 0.02 to 0.88 ± 0.02 for intrahemispheric connections.

    “At connectome level and over all sets of parameter settings, sensitivity varied from 0.22 ± 0.04 to 0.61 ± 0.02, and specificity from 0.60 ± 0.02 to 0.88 ± 0.02 for intrahemispheric connections [mean ± SD (at group level)].”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    Over all sets of parameter settings, an increase in number of streamlines led to higher sensitivity and lower specificity values.

    “Over all sets of parameter settings, an increase in number of streamlines led to higher sensitivity and lower specificity values.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    DT-based and global tractography-based streamlines show mostly smoothly delineated fibers, whereas CSD-based streamlines show a more irregular pattern with many crossing fibers.

    “DT-based and global tractography-based streamlines show mostly smoothly delineated fibers, whereas CSD-based streamlines show a more irregular pattern with many crossing fibers.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    There was a parameter range trade-off between sensitivity and specificity: sensitivity 0.06-0.63 interhemispherically and 0.22-0.86 intrahemispherically; specificity 0.99-0.60 interhemispherically and 0.99-0.23 intrahemispherically.

    “There was a parameter range trade-off between sensitivity: 0.06–0.63 interhemispherically and 0.22–0.86 intrahemispherically; and specificity: 0.99–0.60 interhemispherically and 0.99–0.23 intrahemispherically.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    Increasing FA thresholds, e.g., at an angle threshold of 40° and default step size (15 μm), decreased sensitivity from 0.58 to 0.35 for intrahemispheric and from 0.36 to 0.16 for interhemispheric connections, whereas specificity increased from 0.63 to 0.80 for intrahemispheric and from 0.80 to 0.90 for interhemispheric connections.

    “Increasing FA thresholds (i.e., more strict fiber tracking), e.g., at an angle threshold of 40° and default step size (15 µm), decreased sensitivity from 0.58 to 0.35 for intrahemispheric and from 0.36 to 0.16 for interhemispheric connections, whereas specificity increased from 0.63 to 0.80 for intrahemispheric and from 0.80 to 0.90 for interhemispheric connections.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    For DT-based tractography, for interhemispheric connections, sensitivity ranged from 0.09 ± 0.02 to 0.42 ± 0.04, and specificity from 0.75 ± 0.02 to 0.96 ± 0.01.

    “For interhemispheric connections, sensitivity ranged from 0.09 ± 0.02 to 0.42 ± 0.04, and specificity from 0.75 ± 0.02 to 0.96 ± 0.01.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    For CSD-based tractography, at connectome level and over all sets of parameter settings, sensitivity varied from 0.44 ± 0.04 to 0.86 ± 0.02, and specificity from 0.23 ± 0.02 to 0.78 ± 0.02 for intrahemispheric connections.

    “At connectome level and over all sets of parameter settings, sensitivity varied from 0.44 ± 0.04 to 0.86 ± 0.02, and specificity from 0.23 ± 0.02 to 0.78 ± 0.02 for intrahemispheric connections [mean ± SD (at group level)].”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    Similarity (i.e., overlap) decreased as indicated by the Jaccard index from 0.32 to 0.24 and from 0.16 to 0.10 for, respectively, intra- and interhemispheric connections.

    “Similarity (i.e., overlap) decreased as indicated by the Jaccard index from 0.32 to 0.24 and from 0.16 to 0.10 for, respectively, intra- and interhemispheric connections.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    The tractography maps displayed various neuroanatomical pathways with high detail.

    “The tractography maps displayed various neuroanatomical pathways with high detail.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    At an FA threshold of 0.15, step size of 15 μm and angle threshold of 50°, from 25,000 to 250,000 streamlines, intrahemispheric sensitivity increased from 0.29 to 0.44, whereas specificity decreased from 0.85 to 0.72.

    “For example, at an FA threshold of 0.15, step size of 15 µm and angle threshold of 50°, from 25,000 to 250,000 streamlines, intrahemispheric sensitivity increased from 0.29 to 0.44, whereas specificity decreased from 0.85 to 0.72.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    In general, the Jaccard index was higher for both CSD-based and global tractography (specifically at higher connection potentials), but lower for DT-based tractography.

    “In general, the Jaccard index was higher for both CSD-based and global tractography (specifically at higher connection potentials), but lower for DT-based tractography.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Similar patterns were seen for interhemispheric connections.

    “Similar patterns were seen for interhemispheric connections (data not shown).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    For CSD-based tractography, for interhemispheric connections, sensitivity ranged from 0.15 ± 0.02 to 0.63 ± 0.05, and specificity from 0.60 ± 0.05 to 0.95 ± 0.01.

    “For interhemispheric connections, sensitivity ranged from 0.15 ± 0.02 to 0.63 ± 0.05, and specificity from 0.60 ± 0.05 to 0.95 ± 0.01.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • The quote only partly states this claim.

    Findingown data

    Similar patterns and trade-offs were found when applying SIFT, streamline thresholding and group-based average network thresholding.

    “Similar patterns and trade-offs were found, when we applied spherical deconvolution informed filtering of tractograms, streamline thresholding and group-based average network thresholding.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    Increase in network density was associated with increase in sensitivity (more true positives), as well as with decrease in specificity, due to more false positives.

    “Increase in network density was associated with increase in sensitivity (more true positives), as well as with decrease in specificity, due to more false positives (data not shown).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingreasoning

    Improvements in sensitivity were invariably at the expense of specificity and vice versa.

    “Improvements in sensitivity were invariably at the expense of specificity and vice versa.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Intrahemispheric sensitivity decreased from 0.85 to 0.78, whereas specificity increased from 0.24 to 0.37.

    “Intrahemispheric sensitivity decreased from 0.85 to 0.78, whereas specificity increased from 0.24 to 0.37.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    For CSD-based tractography, the Jaccard index of interhemispheric and intrahemispheric connections showed decreasing patterns with increasing group incidence threshold, although there were peaks depending on streamline number and location.

    “Similarly, the Jaccard index of interhemispheric and intrahemispheric connections based on CSD-based tractography showed decreasing patterns, although there were peaks depending on streamline number and location (e.g., interhemispheric connections at 250,000 streamlines and an FOD threshold of 0.175 led to the highest Jaccard index at an incidence threshold of 0.4) (Figure S5).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Over the same range, the Jaccard index slightly decreased from 0.18 to 0.17 for interhemispheric connections, while it slightly increased from 0.32 to 0.34 for intrahemispheric connections.

    “Over the same range, the Jaccard index slightly decreased from 0.18 to 0.17 for interhemispheric connections, while it slightly increased from 0.32 to 0.34 for intrahemispheric connections.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims. The quote only partly states this claim.

    Findingown data

    For global tractography, the Jaccard index of inter- and intrahemispheric connections decreased with increasing group incidence threshold, with higher particle potentials leading to lower Jaccard indices.

    “Also, the Jaccard index of inter- and intrahemispheric connections decreased for global tractography, although different patterns can be seen for different particle potentials (i.e., higher particle potentials led to lower Jaccard indices) (Figure S6).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    For higher particle potentials (0.2 and 0.5), sensitivity was lower and specificity tended to be higher.

    “Particle potential (i.e., amount and distribution of particles) showed less influence on sensitivity and specificity, although for higher particle potentials, (i.e., 0.2 and 0.5) sensitivity was lower and specificity tended to be higher.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Sensitivity and specificity varied to some extent over different angle thresholds, with a generally declining pattern for sensitivity and an increasing pattern for specificity at increasing angle thresholds.

    “However, sensitivity and specificity varied to some extent over different angle thresholds, with a generally declining pattern for sensitivity and an increasing pattern for specificity at increasing angle thresholds.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    FA threshold (for DT-based tractography), particle potential (GT), group-based average network thresholding, streamline thresholding and filtering (SIFT) were negatively related to sensitivity.

    “Conversely, FA threshold (for DT-based tractography), particle potential (GT), group-based average network thresholding, streamline thresholding and filtering (SIFT) were negatively related to sensitivity.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Although the amount of false positive connections decreased with increasing streamline threshold, the number of true positives decreased as well, specifically at higher thresholds.

    “Although the amount of false positive connections decreased with increasing streamline threshold, the number of true positives decreased as well, specifically at higher thresholds (Figures S12 and S13).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The number of true positives increased together with the number of false positives with increasing connection potentials, although there were generally more false positives.

    “Again, the number of true positives increased together with the number of false positives, although there were generally more false positives (data not shown).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    DT and GT led to low sensitivity and high specificity, while CSD resulted in high sensitivity together with low specificity.

    “While DT and GT led to low sensitivity and high specificity, CSD resulted in high sensitivity together with low specificity.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The neuroanatomical tracer data used as reference accurately represents the true connectivity of the rat brain.

  • May combine two claims.

    Findingown data

    With increasing angle thresholds (from 20 to 80°), at an FOD threshold of 0.1, interhemispheric sensitivity decreased from 0.63 to 0.53, while specificity increased from 0.59 to 0.69.

    “With increasing angle thresholds (i.e., from 20 to 80°), at an FOD threshold of 0.1, interhemispheric sensitivity decreased from 0.63 to 0.53, while specificity increased from 0.59 to 0.69.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    There were more false positives and true positives at an angle of 80° for both intra- and interhemispheric connections.

    “However, there were more false positives and true positives at an angle of 80° for both intra- and interhemispheric connections (data not shown).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    The study detected a considerable number of false positive and false negative connectivity reconstructions, which increased with increasing fiber lengths.

    “Despite our ability to track a large variety of brain-wide neuronal fiber connections, we detected a considerable number of false positive and false negative connectivity reconstructions, which increased with increasing fiber lengths.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The neuroanatomical tracer data used as reference accurately represents the true connectivity of the rat brain.

  • May combine two claims.

    Findingown data

    The Jaccard index generally increased with higher number of streamlines for intra- and interhemispheric tract reconstructions, despite an increase in false positive tract reconstructions.

    “The Jaccard index generally increased with higher number of streamlines for intra- and interhemispheric tract reconstructions, despite an increase in false positive tract reconstructions.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Over all sets of parameter settings, we measured slightly higher sensitivity, lower specificity and higher Jaccard index values with increasing number of streamlines, explained by an increase in true positives and a decrease in false negatives.

    “Over all sets of parameter settings, we measured slightly higher sensitivity, lower specificity and higher Jaccard index values with increasing number of streamlines, explained by an increase in true positives and a decrease in false negatives.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • May combine two claims.

    Findingown data

    DT-based and CSD-based tractography algorithms were affected by connection distance, showing a non-linear decrease in sensitivity and Jaccard index, and a non-linear increase in specificity, with increasing connection distance.

    “Our tested DT-based (i.e., deterministic) and CSD-based (i.e., probabilistic) tractography algorithms were affected by connection distance, showing a non-linear decrease in sensitivity and Jaccard index, and a non-linear increase in specificity, with increasing connection distance.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    CSD-based and global tractography yielded highest Jaccard indices for short-distance connections, which decreased for connections of longer distance.

    “CSD-based and global tractography yielded highest Jaccard indices for short-distance connections, which decreased for connections of longer distance.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Conventional deterministic DT-based and more advanced probabilistic CSD-based and global tractography approaches reconstruct the whole-brain connectome with comparable accuracy, with considerable amounts of false positive and false negative connections, which strongly depends on the algorithm and parameter settings.

    “In conclusion, our study shows that conventional deterministic DT-based and more advanced probabilistic CSD-based and global tractography approaches reconstruct the whole-brain connectome with comparable accuracy, be it with considerable amounts of false positive and false negative connections, which strongly depends on the algorithm and parameter settings.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The neuroanatomical tracer data used as reference accurately represents the true connectivity of the rat brain.

  • May combine two claims.

    Findingnone given

    Tractography conditions differ between intra- and interhemispheric connections as well as over different connection distances.

    “Because tractography conditions differ between intra- and interhemispheric connections as well as over different connection distances, algorithms should be chosen purposefully and, if possible, combined to enable most reliable reconstruction of connectomes at large-scale network or whole-brain level.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Diffusion MRI-based tractography is a useful and powerful technique to map structural connections non-invasively in animal and human brain at meso- and macro-level.

    “Diffusion MRI-based tractography has offered exclusive means to map structural connections non-invasively in animal and human brain, and is considered a very useful and powerful technique to unravel the human connectome at meso- and macro-level (Le Bihan et al. 1986; Basser et al. 1994b; Mori et al. 1999; Jones 2008; Jbabdi and Johansen-Berg 2011).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Le Bihan et al. 1986 source
    • cites Basser et al. 1994b source
    • cites Mori et al. 1999

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    SIFT reduces local streamline density bias by filtering the tractogram to improve the fit between the overall distribution of streamlines and the FOD in every voxel, which may improve accuracy of connectome reconstruction.

    “Recently, a new method, called spherical-deconvolution informed filtering of tractograms (SIFT) has been proposed to reduce local streamline density bias by filtering the tractogram (i.e., selectively removing streamlines) to improve the fit between the overall distribution of streamlines and the FOD in every voxel, estimated from the diffusion-weighted image, which may improve accuracy of connect”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Smith et al. 2013 source
    • cites Smith et al. 2015a source

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    CSD estimates multiple fiber directions within a voxel, resulting in a fiber orientation distribution (FOD) that may have multiple peaks.

    “Instead of estimating one principal fiber direction, CSD estimates multiple fiber directions within a voxel, resulting in a fiber orientation distribution (FOD) that may have multiple peaks (e.g., four in case of two crossing fibers in one voxel) (Tournier et al. 2010, 2012).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Tournier et al. 2010

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Tournier et al. 2012 source

  • May combine two claims.

    Interpretationown data

    The increase of the Jaccard index is explained by a considerable increase in true positives (TP) and a decrease in false negatives (FN).

    “This increase of the Jaccard index is again explained by a considerable increase in true positives (TP) and a decrease in false negatives (FN) (data not shown).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Global tractography aims to find the full track configuration that most optimally fits the acquired data.

    “Global tractography aims to find the full track (i.e., whole brain) configuration that most optimally fits the acquired data (Reisert et al. 2011; Mangin et al. 2013; Christiaens et al. 2015b).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Reisert et al. 2011 source
    • cites Mangin et al. 2013 source
    • cites Christiaens et al. 2015b source

  • May combine two claims.

    Interpretationown data

    The angle thresholds and step sizes as applied in the study were within plausible neuroanatomical ranges, thereby not affecting the overall connectome reconstructions.

    “Apparently, the angle thresholds and step sizes—both important factors related to tract curvature—as applied in our study were within plausible neuroanatomical ranges, thereby not affecting the overall connectome reconstructions.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    To influence connectome reconstruction sensitivity and specificity, FOD thresholds should be increased to much higher values (far beyond the default 0.1 threshold), but this would probably lead to undesirable exclusion of many (truly) crossing fibers.

    “Therefore, to influence the connectome reconstruction sensitivity and specificity, we should apparently increase the FOD thresholds to much higher values (i.e., far beyond the default 0.1 threshold), but this would probably lead to undesirable exclusion of many (truly) crossing fibers.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    Correction for eddy currents or bias field was not necessary because a special volume coil and MRI pulse sequence allowed homogenous imaging of post mortem brain tissue.

    “Correction for eddy currents or bias field was not necessary, because we used a special volume coil and MRI pulse sequence (with extra delays after sinus-shaped gradients) that allowed homogenous imaging of post mortem brain tissue.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    The lack of FOD threshold impact may be explained by the different natures of DT-based (deterministic) and CSD-based (probabilistic) tractography or by the different natures of the FA and FOD thresholds.

    “This may be explained by the different natures of DT-based tractography (i.e., deterministic) and CSD-based tractography (i.e., probabilistic) or by the different natures of the FA and FOD thresholds.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    The effects of FOD threshold adjustments may be different across species and datasets.

    “Also, it should be noted that the effects of FOD threshold adjustments may be different across species and datasets.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Both number of streamlines and connection potential increased the total number of reconstructed connections, leading to a higher true positive rate.

    “Both factors increased the total number of reconstructed connections, leading to a higher true positive rate.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductreasoning

    The study did not assess connectivity of subcortical regions, which biases conclusions towards cortical connectomics.

    “We also did not assess connectivity of subcortical regions, which are less easily co-registerable to the rat brain anatomical atlas, which biases our conclusions towards cortical connectomics.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The weights of neuronal connections were not considered in the study.

    “Lastly, the weights of neuronal connections were not considered in our study.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductnone given

    The study only considered streamline endpoints in determining connectivity between two regions of interest, thereby ignoring the three-dimensional structure of anatomical pathways.

    “Another point is that we only considered streamline endpoints in determining connectivity between two regions of interest, thereby ignoring the three-dimensional structure of anatomical pathways.”
    Quote in the paperSubstantial claim
56 of 56

  • The paper relies on this without support.

    The Euclidean distance between region centers of gravity is a valid proxy for fiber length.

    “We determined sensitivity, specificity and the Jaccard index of DT-based, CSD-based and global tractography for each distance category (i.e., as a proxy of fiber length).”

  • The paper relies on this without support.

    The tracer-based connectome from the rat connectome database accurately represents true neuroanatomical connectivity (ground truth).

    “Diffusion tractography-based connectivity networks were compared against the neuronal tracer-based network as ground truth, which yielded true positives (green lines and squares), false positives (red lines and squares), false negatives (dotted red lines and squares), and true negatives (no line and color-coding) (right column).”

  • The paper relies on this without support.

    The 106 cortical regions from the Paxinos and Watson atlas correspond to the regions used in the tracer database.

    “We extracted connectivity information for the same 106 cortical regions (Table 1) as used for diffusion-based connectome reconstructions.”

  • The paper relies on this without support.

    Group-based average network thresholding improves signal-to-noise by including consistently found connections and controlling for particularities of individual rat networks.

    “To assess the effect of group average-based connectome reconstruction, which may improve signal-to-noise by including consistently found connections and controlling for particularities of individual rat networks, we included connections based on incidence over all rats (N = 10) using different thresholds.”

  • The paper relies on this without support.

    The neuroanatomical tracer data used as reference accurately represents the true connectivity of the rat brain.

    “The connectome reconstructions were compared against neuroanatomical tracer data.”

  • The paper relies on this without support.

    The rat brain anatomical atlas and region parcellation used are valid.

    “We also did not assess connectivity of subcortical regions, which are less easily co-registerable to the rat brain anatomical atlas”

  • The paper relies on this without support.

    The streamline endpoint criterion for connectivity is adequate for the study's conclusions.

    “Another point is that we only considered streamline endpoints in determining connectivity between two regions of interest, thereby ignoring the three-dimensional structure of anatomical pathways.”

  • The paper relies on this without support.

    The parameter settings tested are representative of the tractography algorithms' behavior.

    “for a specific set of parameter settings for DT-based, CSD-based and global tractography (as representative examples)”

  • The paper relies on this without support.

    The Jaccard index is a valid measure of overall connectome reconstruction performance.

    “overall performance (i.e., as expressed by the Jaccard index)”

  • The paper relies on this without support.

    The group-based incidence thresholding is a valid method to increase robustness against inter-subject variation.

    “group-based analysis (i.e., incidence thresholds, which might increase robustness against inter-subject variation)”
10 of 10

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Figure 6 group-threshold analysis has conflicting FA thresholds

    Minor

    The results text says the illustrated DT-based group-threshold analysis uses an FA threshold of 0.2, while the Figure 6 caption specifies 0.15.

  • Specificity is mislabeled as the false positive rate

    Minor

    The methods call specificity the false positive rate, but the figure caption correctly identifies the false positive rate as 1 minus specificity.

  • Reported particle potential was not among the tested values

    Minor

    The global-tractography results describe a particle potential of 0.01, whereas the methods list 0.1 as the lowest tested particle potential.

  • Reported global-tractography connection potential was not among the tested values

    Minor

    The distance analysis reports a connection potential of 1, although the methods list tested connection potentials only up to 0.5.

  • CSD angle comparison reverses the stated positive counts

    Minor

    From 20° to 80°, the reported sensitivity falls and specificity rises for both connection types. With the same ground-truth connections, that means fewer true positives and fewer false positives at 80°, contrary to the subsequent statement.

5 of 5

Cites

Flagged first, then by citations

4 of the 79 references couldn’t be matched to a known paper.

Research works printed but not linked: 1

  1. Tournier JD, Calamante F, Connelly A (2010) Improved probabilistic streamlines tractography by 2nd order integration over fibre orientation distributions. Ismrm 88:1670

Cited by

Sorted by citations

Related

Authors
Sinke, M. R. T., Otte, W. M., Christiaens, D., Schmitt, O., Leemans, A., van der Toorn, A., Sarabdjitsingh, R. A., Joëls, M., & Dijkhuizen, R. M.
Affiliations
  • Utrecht University (NL)
  • King's College London (GB)
  • University of Rostock (DE)
  • University Medical Center Groningen (NL)
Corresponding author
Sinke, Michel R. T.
ORCID
Venue
Brain Structure and Function
Publisher
Springer Science and Business Media LLC
Published
2018-02-20
Volume
223
Issue
5
Pages
2269-2285
ISSN
  • 1863-2653 (ISSN-L)
  • 1863-2661
DOI
10.1007/s00429-018-1628-y
PMC
PMC5968063
OpenAlex
W2792108748
Cited by
70
References
79
Citations by year
2026: 1 · 2025: 5 · 2024: 5 · 2023: 3 · 2022: 4 · 2021: 19 · 2020: 15 · 2019: 13 · 2018: 5
FWCI
8.20
Type
journal article
Language
en
Open access
Licence
Locations
  • Brain Structure and Function (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Lirias (submittedVersion, cc-by)
  • Europe PMC (PubMed Central) (submittedVersion, cc-by)
  • PubMed Central (submittedVersion, cc-by)
  • Utrecht University Repository (Utrecht University) (submittedVersion, other-oa)
  • University of Groningen research database (University of Groningen / Centre for Information Technology) (submittedVersion, cc-by)
Funders
  • Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO-VICI 016.130.662, NWO-VENI 016.168.038)
  • Hersenstichting ([F2014(1)-06])
Topics
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Advanced Neuroimaging Techniques and Applications
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Functional Brain Connectivity Studies
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Advanced MRI Techniques and Applications
Keywords
cortical connectome, diffusion MRI tractography, constrained spherical deconvolution, diffusion tensor imaging, rat brain
MeSH terms
Algorithms, Animals, Brain Mapping, Cerebral Cortex, Image Processing, Computer-Assisted, Male, Neural Pathways, Neurons, Sensitivity and Specificity, Rats, Wistar, Diffusion Magnetic Resonance Imaging, Rats, Connectome, White Matter

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-11.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-11.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review