Quantitative mapping of the brain’s structural connectivity using diffusion MRI tractography: A review

In the paper

“Despite focusing on different developmental ranges, these studies consistently observe significant age-related increases of FA and MD in widespread fiber tracts during postnatal growth in both cross-sectional and longitudinal population designs (Lebel and Deoni, 2018).”

What the sources say

• Disputed

  Section 3.3 (Developmental Timing; page not provided) reports that FA increases with age, but MD decreases—the opposite of the claimed MD increase. It also discusses longitudinal developmental curves, so the disagreement is specifically about MD's direction of change. (LLM, conf=0.90)

  Checked by a reasoning model — may be wrong.

  cites Lebel and Deoni, 2018 sourceExperimental — accuracy not yet measured

In the paper

“actually none proved effective in discriminating between true and spurious connections in the connectome networks (Maier-Hein et al., 2017; Schiavi et al., 2020a).”

What the sources say

• Partly supported

  The Results (‘Tractograms contained more invalid than valid bundles’; page unavailable) report that tested tractography approaches produced more invalid than valid bundles, supporting the concern about spurious connections. But the Results (‘Statistical analysis of processing steps’; page unavailable) say filtering affected performance and one clustering approach achieved 92% validly connecting streamlines. The paper therefore does not establish that no knowledge- or data-driven filtering method proved effective at discrimination. (LLM, conf=0.85)

  cites Maier-Hein et al., 2017 sourceExperimental — accuracy not yet measured
• ?Not matched

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

  cites Schiavi et al., 2020a

In the paper

“It is important to note that tractography is known to be sensitive to the choice of the underlying fiber tracking algorithms, which can further affect any subsequent quantitative analyses using the tractography data.”

What the sources say

• ?Could not verify

  The Microstructure models > Effect of model fitting passage mentions combining global tractography with microstructure estimation, but does not compare fiber-tracking algorithms or their effects on analyses of tractography data. Searches for tractography and tracking algorithms found no passage in our copy addressing that specific claim; pages were not provided. (LLM, conf=0.72)

  cites Afzali et al., 2020 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Alexander, 2006 source
• ?Not checked — no full text of the sourcecites Panagiotaki et al., 2012 source

In the paper

“the Anatomically-Constrained Tractography (ACT) framework (Smith et al., 2012) and similar methods (Girard and Descoteaux, 2012; Lemkaddem et al., 2014; Yeh et al., 2017) ensure that both streamline propagation and termination are constrained based on knowledge of where neuronal fibers locate, e.g., connecting between gray matter areas via the white matter.”

• Also stated in: § Termination bias · plain

  “they do prevent the generation of streamlines that cannot possibly represent biological connectivity within the brain.”

What the sources say

• ?Not checked — no full text of the sourcecites Smith et al., 2012 source
• ?Not matched

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

  cites Girard and Descoteaux, 2012
• Supported

  The Introduction states that the method places fiber endpoints in gray matter and constrains the remaining spline control points to white matter. The Methods section says it filters out fibers ending prematurely in white matter because they do not represent anatomically valid connections. These passages support the claim about anatomically constrained paths and termination. (LLM, conf=0.88)

  cites Lemkaddem et al., 2014 source

In the paper

“Possibly the earliest bias established for dMRI tractography was the fact that if the streamline algorithm simply takes a step of finite size in the direction of the local fiber orientation at the current vertex (a so-called “first-order” method), streamlines in curved bundles will tend to under-estimate the curvature, resulting in erroneous trajectories (Tournier et al., 2002).”

What the sources say

• Partly supported

  The Tracking Algorithm section describes finite steps taken in the local major-eigenvector direction. Model A reports that larger steps increase endpoint offset and that tracks can depart systematically from an ideal curved path, especially without interpolation. The copy does not establish the claimed direction of the bias—that these tracks tend to underestimate curvature. (LLM, conf=0.85)

  cites Tournier et al., 2002 sourceExperimental — accuracy not yet measured

In the paper

“In particular, it has been shown that tractography techniques suffer from a large number of false positives, i.e., streamlines that are reconstructed but do not correspond to real anatomical bundles”

What the sources say

• Supported

  The Results section (“Tractograms contained more invalid than valid bundles”) reports that 36 ± 17% of reconstructed streamlines connected regions not connected in the ground truth, and that invalid bundles outnumbered valid bundles. The Introduction likewise says most evaluated algorithms produced large amounts of false-positive bundles. (LLM, conf=0.92)

  cites Maier-Hein et al., 2017 source
• Partly supported

  The Discussion (page unavailable) reports that high-sensitivity tractography tends to have a high false-positive rate, supporting the concern about reconstructed pathways without anatomical support. But the paper evaluates false positives by anatomical region, not by counting individual streamlines, and says some methods and settings can achieve much lower false-positive rates. The claim overgeneralizes those findings to tractography techniques as a whole. (LLM, conf=0.85)

  cites Thomas et al., 2014 sourceExperimental — accuracy not yet measured

In the paper

“and that these erroneous connections can severely bias the estimation of connectivity (Drakesmith et al., 2015; Maier-Hein et al., 2017; Zalesky et al., 2016b).”

What the sources say

• Supported

  The Introduction says inferred brain-network measures are sensitive to errors in edges, including false positives. In the Discussion’s “Effects of false positives” section (page not provided), the authors report that even a single false positive significantly affects inferred network metrics and that added false edges produce a strong bias in global efficiency. (LLM, conf=0.91)

  cites Drakesmith et al., 2015 source
• Supported

  The Discussion (page unavailable) reports that 64% of the bundles systematically recovered by most submissions were absent from the ground truth. This supports the claim that false-positive connections can severely distort estimated connectivity, though the paper cautions that its simulation cannot quantify the effect on in vivo connectivity in absolute terms. (LLM, conf=0.90)

  cites Maier-Hein et al., 2017 source
• ?Not matched

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

  cites Zalesky et al., 2016b

In the paper

“On one hand, tractography results may vary across the large set of available diffusion models (Afzali et al., 2020; Alexander, 2006; Panagiotaki et al., 2012; Yablonskiy and Sukstanskii, 2010) and tractography methods”

What the sources say

• ?Not checked — no full text of the sourcecites Basser et al., 2000 source
• Supported

  The Discussion (page not provided) compares single- and multi-fibre tractography and says they can give strikingly different results for non-dominant pathways. It also describes differences among diffusion-reconstruction techniques. This supports the claim that tractography results may vary with the model or method used. (LLM, conf=0.88)

  cites Behrens et al., 2007 source
• Supported

  The Introduction says diffusion-tensor-based tractography can produce false tracts or stop prematurely. The Discussion compares different tracking approaches and reports that they reconstruct different portions of fiber bundles, supporting the claim that results vary with the model and method used. Page numbers are not available in our copy. (LLM, conf=0.90)

  cites Descoteaux et al., 2008 source

In the paper

“despite the fact that white matter fibers synapse in the gray matter (Daducci et al., 2016)”

What the sources say

• Supported

  Section 2.4, “Partial fibers” (page unavailable), states that axons connect neurons located in gray matter and contrasts this with reconstructed tracts that stop prematurely in white matter. This supports the citing sentence’s point that white-matter fibers connect to neurons in gray matter. (LLM, conf=0.87)

  cites Daducci et al., 2016 source

In the paper

“On the other hand, tractography results can also be sensitive to the choice of parameters (such as seeding and stopping thresholds) within a certain algorithm (Chamberland et al., 2014; Côté et al., 2013; Gong et al., 2018; Moldrich et al., 2010; Xie et al., 2020).”

What the sources say

• Supported

  Section 2.2 says subtle variations in tractography parameters can lead to different results and specifically discusses seeding strategies and stopping criteria. The Results (section 3.4) show that fixed parameter choices can produce misleading reconstructions, including spurious streamlines. (LLM, conf=0.94)

  cites Chamberland et al., 2014 source
• ?Not matched

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

  cites Côté et al., 2013
• Supported

  The Introduction states that varying tractography thresholds has a large impact on fiber tracking using UKFt near edema. The Discussion adds that reducing thresholds used to start and stop tractography can increase fiber-tract size, supporting the claim that results are sensitive to parameter choices within an algorithm. (LLM, conf=0.93)

  cites Gong et al., 2018 source

In the paper

“For automated tract identification, one study indicates that streamline labeling has higher test-retest reproducibility than ROI-based segmentation (Zhang et al., 2019)”

What the sources say

• Supported

  The Discussion reports that fiber clustering had significantly higher test–retest reproducibility than the cortical-parcellation-based method for identifying anatomical fiber tracts. This supports the citing sentence’s comparison of streamline-based identification with ROI-based segmentation. (LLM, conf=0.94)

  cites Zhang et al., 2019 source

In the paper

“Methods using manual streamline selection have been shown to be highly reproducible for specific tracts and segmentation protocols using the diffusion tensor tractography (Wakana et al., 2007)”

What the sources say

• Supported

  The Methods describe assigning tracked fibers to specific tracts using manually defined ROIs and testing the resulting protocols across raters. The Conclusions report high reproducibility for the 11 selected tracts; the Discussion cautions that this does not establish high reproducibility for tractography in general. (LLM, conf=0.92)

  cites Wakana et al., 2007 source

In the paper

“It was recently demonstrated that the performance of filtering methods can be significantly boosted by combining knowledge- and data-driven strategies.”

What the sources say

• Supported

  The Discussion (page unavailable) contrasts the limited anatomical improvement from data-driven COMMIT with the substantial performance gain after COMMIT2 added an anatomical prior. The Results (page unavailable) also report further filtering improvements from the additional priors in COMMIT2 tree. (LLM, conf=0.90)

  cites Ocampo-Pineda et al., 2021 source
• ?Not matched

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

  cites Schiavi et al., 2020a

In the paper

“Research is still underway to determine the best method(s) for filtering, with some disagreement on the beneficial versus detrimental effect of these methods on the structural connectome (Buchanan et al., 2020; Civier et al., 2019; Drakesmith et al., 2015; Frigo et al., 2020; Schiavi et al., 2020a; Yeh et al., 2016a).”

What the sources say

• Supported

  The Introduction describes competing approaches: some researchers advocate leaving networks unthresholded, while many studies remove weak connections, which may include plausible connections. The Discussion reports that the utility of thresholding differs by diffusion metric, and the Limitations section says there is no agreed schema for constructing structural networks. Together, these passages support the claim that the value of filtering remains unsettled. (LLM, conf=0.91)

  cites Buchanan et al., 2020 source
• Supported

  The Abstract (page not provided) says there is a lack of consensus on analyzing dense weighted structural connectomes and examines whether removing weak connections is necessary. The Introduction (page not provided) describes both the benefit of reducing the influence of possibly spurious connections and the confounds introduced by threshold choices. This supports the cited paper’s contribution to the broader statement about the debated value of filtering. (LLM, conf=0.88)

  cites Civier et al., 2019 source
• Partly supported

  The Introduction’s “Thresholding” section says removing weak connections can suppress false positives but can also remove genuine connections and bias network metrics; it calls the alternative of leaving graphs unthresholded contentious. This supports a beneficial-versus-detrimental trade-off for network thresholding, but the paper does not establish disagreement about filtering methods for the structural connectome generally. Page numbers were not provided in our copy. (LLM, conf=0.82)

  cites Drakesmith et al., 2015 sourceExperimental — accuracy not yet measured

In the paper

“This prior knowledge helps resolve some of the ambiguities present in the data and, as a consequence, allows significant improvement of the anatomical accuracy of the connectome (Ocampo-Pineda et al., 2021; Schiavi et al., 2020a).”

What the sources say

• Supported

  Section 2.2 describes grouping streamlines hierarchically using Group Lasso. The Discussion reports that this anatomical prior helped distinguish valid from invalid streamlines and significantly improved the accuracy of individual bundles; the Conclusion says it further improved reconstruction quality. These passages support the improvement attributed to this paper in the cited sentence. (LLM, conf=0.90)

  cites Ocampo-Pineda et al., 2021 source
• ?Not matched

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

  cites Schiavi et al., 2020a
• Partly supported

  Section 2 (Group lasso; page not available) says group lasso selects groups of variables and encourages sparsity at the factor level. The paper does not address streamline filtering, bundles, connectomes, or improved anatomical accuracy; those application and outcome claims are unsupported by this source. (LLM, conf=0.84)

  cites Yuan and Lin, 2006 sourceExperimental — accuracy not yet measured

In the paper

“While there are many sophisticated tools to compute these registrations such as FSL (Jenkinson et al., 2012) and ANTs (Avants et al., 2009), the performance is limited by nontrivial factors.”

What the sources say

• Supported

  The citation identifies FSL as an example of a registration tool, and the paper’s “Past to present” section describes FSL’s linear registration and later nonlinear method, FNIRT. That section also discusses factors affecting registration, including scanner data quality, EPI distortions, template detail and artefacts. The separate ANTs citation is not being attributed to this source. (LLM, conf=0.91)

  cites Jenkinson et al., 2012 source
• ?Not checked — no full text of the sourcecites Avants et al., 2009 source

In the paper

“studies have suggested a fine scale white matter parcellation (over 2000 parcels) can be beneficial for machine learning and statistical analysis (Liu et al., 2017; Zhang et al., 2018a) while a coarse scale parcellation (fewer than 200 parcels) can improve the connectome consistency across individuals (Osmanlıoğlu et al., 2020; Rodrigues et al., 2013).”

What the sources say

• Partly supported

  The Discussion reports that finer-grained brain parcellation, at or approaching 2000 nodes, improved classification of consciousness states. The Methods describe statistical comparisons and classification, but specify a maximum of 2000 nodes, not over 2000. The paper describes anatomical-functional brain parcellation, not a white-matter parcellation. (LLM, conf=0.89)

  cites Liu et al., 2017 sourceExperimental — accuracy not yet measured
• ?Not matched

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

  cites Zhang et al., 2018a

In the paper

“For example, many factors (e.g., cell death, edema, gliosis, inflammation, change in myelination, increase in connectivity of crossing fibers, increase in extracellular or intracellular water, etc.) may cause changes in FA (Assaf and Pasternak, 2008; Jones et al., 2013; Le Bihan and Johansen-Berg, 2012).”

What the sources say

• ?Not checked — no full text of the sourcecites Assaf and Pasternak, 2008 source
• ?Not checked — no full text of the sourcecites Jones et al., 2013 source
• Partly supported

  The sections on diffusion anisotropy in white matter and white matter diseases (pages not provided) say that myelin, fiber arrangement, extracellular water and tissue orientation patterns can affect anisotropy, supporting part of the FA claim. The cell-size section links edema and shifts between intracellular and extracellular spaces to changes in ADC, not specifically FA. Our copy does not establish cell death, gliosis, inflammation, or increased connectivity of crossing fibers as causes of FA changes. (LLM, conf=0.86)

  cites Le Bihan and Johansen-Berg, 2012 sourceExperimental — accuracy not yet measured

In the paper

“Several studies conducted on animals using invasive tract-tracing techniques have shown good agreement between the NOS and the biological connectivity as measured using ex vivo tract-tracing data (Delettre et al., 2019; Girard et al., 2020; van den Heuvel et al., 2015a).”

What the sources say

• Supported

  The Tractography-Based Connectivity Matrices section says the MRI connectivity estimates were derived from streamline counts, then normalized to fractions of streamlines. The Results report strong positive correlations between those estimates and invasive tract-tracing data from ferrets (Spearman’s rho 0.67–0.91). Our copy provides section labels but no page numbers. (LLM, conf=0.91)

  cites Delettre et al., 2019 source
• Partly supported

  The Conclusion reports good agreement between ex-vivo macaque tractography and connections identified by histological tracing. But §4.5 says the comparison was of binary connectivity, not histological cell counts or connection strength; it therefore does not establish good quantitative agreement between the number of streamlines (NOS) and biological connectivity. (LLM, conf=0.85)

  cites Girard et al., 2020 sourceExperimental — accuracy not yet measured
• ?Not matched

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

  cites van den Heuvel et al., 2015a

In the paper

“studies have suggested a fine scale white matter parcellation (over 2000 parcels) can be beneficial for machine learning and statistical analysis (Liu et al., 2017; Zhang et al., 2018a) while a coarse scale parcellation (fewer than 200 parcels) can improve the connectome consistency across individuals (Osmanlıoğlu et al., 2020; Rodrigues et al., 2013).”

What the sources say

• Supported

  The Discussion (‘Processing of the data affects consistency’) reports that connectomes with fewer regions were more consistent across subjects. The Methods (‘Atlases and registrations’) identifies the coarsest parcellation as Schaefer 118, which has fewer than 200 regions. (LLM, conf=0.93)

  cites Osmanlıoğlu et al., 2020 source
• ?Not checked — no full text of the sourcecites Rodrigues et al., 2013 source

In the paper

“Various parcellation scales have been applied in different studies, ranging from tens to millions of white matter parcels (Besson et al., 2014; Liu et al., 2017; Osmanlıoğlu et al., 2020; Rodrigues et al., 2013; Zalesky et al., 2010; Zhang et al., 2018a).”

What the sources say

• ?Not checked — no full text of the sourcecites Besson et al., 2014 source
• Partly supported

  The Materials and Methods section (page unavailable) reports parcellation at multiple scales, from 116 to 2,000 brain-network nodes. That supports the use of varying scales, but this paper does not establish the claimed range from tens to millions or describe those nodes as white matter parcels. (LLM, conf=0.88)

  cites Liu et al., 2017 sourceExperimental — accuracy not yet measured
• Partly supported

  The Data and methods section (“Atlases and registrations”; page unavailable) reports structural and functional connectomes at three parcellation scales: 118, 218, and 318 regions. This supports the use of different scales, but the paper does not establish a range from tens to millions of white matter parcels; its reported regions are not described as white matter parcels. (LLM, conf=0.86)

  cites Osmanlıoğlu et al., 2020 sourceExperimental — accuracy not yet measured

In the paper

“The number of streamlines (NOS) is one widely used measure to quantify the connectivity strength (Hagmann et al., 2008b; Ingalhalikar et al., 2014; Roberts et al., 2016; Sporns et al., 2005; van Dellen et al., 2018).”

What the sources say

• ?Not matched

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

  cites Hagmann et al., 2008b
• Partly supported

  In Materials and Methods (page not provided), the paper uses the number of tracked fibers reaching a target region to calculate a connectivity weight and notes that similar measures appear in previous studies. Its measure is a scaled probability, however, not the raw number of streamlines; the passage therefore does not establish that NOS itself is widely used to quantify connectivity strength. (LLM, conf=0.82)

  cites Ingalhalikar et al., 2014 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Roberts et al., 2016 source

In the paper

“However, because of the difference in resolution between the measured dMRI signal and actual axon dimensions (as described in Section 3.3: “Connection density biases”), many studies have emphasized that NOS does not provide a truly quantitative measure of connection strength (Jones, 2010; Jones et al., 2013; Sotiropoulos and Zalesky, 2019; Yeh et al., 2020).”

What the sources say

• ?Not checked — no full text of the sourcecites Jones, 2010 source
• ?Not checked — no full text of the sourcecites Jones et al., 2013 source
• Supported

  In “Mapping edges” (page not provided), the paper says MRI voxels cannot resolve individual axons and the measured signal is far from the scale of interest. In “Quantifying edges,” it says streamline counts are commonly used for connection strength, but dMRI cannot directly measure that strength or the number of connecting axons. (LLM, conf=0.92)

  cites Sotiropoulos and Zalesky, 2019 source

In the paper

“We note that inter-subject connectome consistency may also be improved using other approaches, e.g., consistency-based thresholding (Baum et al., 2018; Roberts et al., 2017).”

What the sources say

• Supported

  The Methods sections on consistency-based thresholding and inter-subject edge consistency describe selecting connections that are consistently reconstructed across subjects and removing potentially spurious edges. That supports the hedged claim that this approach may improve inter-subject connectome consistency. (LLM, conf=0.86)

  cites Baum et al., 2018 source
• ?Not checked — no full text of the sourcecites Roberts et al., 2017 source

In the paper

“More advanced measures include those reflecting inter- and intra-cellular signal fractions derived from Diffusion Kurtosis Imaging (DKI) (Jensen and Helpern, 2010), Free Water modeling (Pasternak et al., 2012; 2009), Neurite Orientation Dispersion and Density Imaging (NODDI) (Zhang et al., 2012), Spherical Mean Technique (SMT) (Kaden et al., 2016), or Apparent Fiber Density (AFD) (Raffelt et al.,”

What the sources say

• Partly supported

  The Simple Models section (page unavailable) discusses intracellular and extracellular compartments and relates kurtosis to a compartment’s water fraction. It says, however, that DKI-derived diffusion and kurtosis measures do not uniquely determine the fractions: kurtosis provides only an upper bound for the fast-diffusing compartment’s fraction. The paper therefore supports DKI as an advanced measure sensitive to tissue structure, but not the claim that it yields inter- and intra-cellular signal fractions. (LLM, conf=0.86)

  cites Jensen and Helpern, 2010 sourceExperimental — accuracy not yet measured
• Partly supported

  The Free-Water Model section describes separate free-water and tissue signal contributions and identifies free water as extracellular. It does not identify the tissue contribution as an intra-cellular fraction, so the citation supports the Free Water example but not that more specific characterization. (LLM, conf=0.86)

  cites Pasternak et al., 2012 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Zhang et al., 2012 source

In the paper

“Usually, given a quantitative measure of interest, a selected summary statistic (e.g. mean of the chosen microstructural measure) is extracted from the tract(s)-of-interest and is used to compute the level of group differences (e.g., the p-value) using statistical group-wise comparison methods such as Student’s t-test, ANOVA, or other more advanced statistical analysis methods (Habeck and Stern, 2”

What the sources say

• Partly supported

  The Toy Example section (page unavailable) describes a standard group comparison using a univariate T-test, supporting that part of the citation. It does not describe extracting mean microstructural measures from tracts of interest or comparing those summaries with ANOVA; its example compares voxels and PCA subject scores instead. (LLM, conf=0.84)

  cites Habeck and Stern, 2010 sourceExperimental — accuracy not yet measured
• Partly supported

  Section 4.5 (p. 16) discusses statistical methods for detecting longitudinal MRI group differences and names ANOVA as a comparator. It does not describe hypothesis-driven comparisons of tracts of interest, extraction of a summary microstructural measure, or use of Student’s t-test or p-values. (LLM, conf=0.79)

  cites Martí-Juan et al., 2020 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Ombao et al., 2016 source

In the paper

“Instead of using a pre-calculated microstructure image, the diffusion or fiber model can be simultaneously estimated when performing tractography so that the microstructural measures are directly computed at points along each streamline (Girard et al., 2017; Gong et al., 2018; Malcolm et al., 2010; Olszewski et al., 2017; Reddy and Rathi, 2016).”

What the sources say

• ?Not checked — no full text of the sourcecites Girard et al., 2017 source
• Supported

  Materials and methods §2.3 says the UKFt fiber model is fitted at each tracking step, using information from the previous step. Results §3.4 describes the resulting free-water fraction as an estimate at each point along the tract, supporting the claim that a microstructural measure is obtained along the streamline. Page numbers are unavailable in our copy. (LLM, conf=0.92)

  cites Gong et al., 2018 source
• Supported

  In II.B, the paper says it estimates model parameters from the measured signal at each position while tracing a fiber. II.C describes repeating that estimation as the fiber advances, and III.C identifies fractional anisotropy as a diffusion measure estimated by the filter. Together, these support computing such measures along a streamline without a pre-calculated microstructure image. (LLM, conf=0.90)

  cites Malcolm et al., 2010 source

In the paper

“Other scalar values that are generally interpreted as measures of connectivity “strength” include microstructural measures, e.g., FA (Bathelt et al., 2017) and R1 (Boshkovski et al., 2020a).”

What the sources say

• Supported

  The Bathelt et al. citation is attached to FA, not R1. The Introduction says the study's network edges represented FA of white-matter connections, and Structural Connectome Analysis says FA determined the connection-matrix weights. This supports FA as a microstructural measure used to represent connectivity strength. (LLM, conf=0.91)

  cites Bathelt et al., 2017 source
• ?Not matched

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

  cites Boshkovski et al., 2020a

In the paper

“The most popular scalar value for connectivity matrix analysis is the NOS (Hagmann et al., 2008b; Ingalhalikar et al., 2014; Roberts et al., 2016; Sporns et al., 2005; van Dellen et al., 2018), where higher values are considered to indicate stronger connectivity between pairs of gray matter regions.”

What the sources say

• ?Not matched

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

  cites Hagmann et al., 2008b
• Partly supported

  Materials and Methods (page unavailable) describes a connectivity matrix derived from streamline counts, with higher connection weights treated as greater connectivity. It does not identify the number of streamlines (NOS) as the most popular scalar value; its own matrix entries are scaled pathway probabilities rather than raw streamline counts. (LLM, conf=0.84)

  cites Ingalhalikar et al., 2014 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Roberts et al., 2016 source

In the paper

“One popular approach is to average data across corresponding points along each streamline in the pathway and leverage a prototype, core, or mean streamline to enable statistical analysis comparing subjects or groups (Colby et al., 2012b; Wang et al., 2016; Yeatman et al., 2012).”

What the sources say

• ?Not matched

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

  cites Colby et al., 2012b
• ?Not checked — no full text of the sourcecites Wang et al., 2016 source
• Supported

  In Methods (AFQ, Fiber Tract Quantification; page unavailable), the paper describes averaging diffusion measurements across corresponding nodes of fibers, weighted by each fiber’s distance from the tract core. The Methods section on Individual and Group Level Inference describes point-wise statistics on those tract profiles for comparisons across subjects or groups. (LLM, conf=0.93)

  cites Yeatman et al., 2012 source

In the paper

“Quantitative measures extracted for individual fiber pathways can also be used to construct a structural connectivity matrix to map the whole-brain network of structural connectivity between all pairs of gray matter regions (Hagmann et al., 2008a; Sporns et al., 2005).”

What the sources say

• ?Not matched

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

  cites Hagmann et al., 2008a
• Partly supported

  In “Steps Toward the Human Connectome,” the paper proposes using probabilistic tractography of fiber pathways to produce an all-to-all structural connectivity matrix for the human brain. It specifies a voxel-wise probabilistic matrix, however, rather than a matrix between all pairs of gray-matter regions, and does not specify quantitative measures extracted for each individual pathway. (LLM, conf=0.84)

  cites Sporns et al., 2005 sourceExperimental — accuracy not yet measured

In the paper

“Instead of using a single scalar value to summarize the entire pathway, many methods perform along-tract analysis to investigate the distribution of the microstructural measures along the fiber pathway (variously called “tractometry,” “profilometry,” and “tract-based morphometry”) (Bells et al., 2011b; O’Donnell et al., 2009; Yeatman et al., 2012).”

What the sources say

• ?Not matched

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

  cites Bells et al., 2011b
• Supported

  Section 2.5 describes calculating FA and other scalar measures at successive arc-length coordinates along fibers, rather than reducing a bundle to one value. The Introduction also describes multiple approaches to measuring diffusion along tracts. This supports the citation’s use of tract-based morphometry as an example of along-tract analysis. (LLM, conf=0.93)

  cites O’Donnell et al., 2009 source
• Supported

  The Introduction contrasts averaging diffusion properties over an entire tract with analyzing measurements at different locations along it. The Methods section describes AFQ resampling fibers into 100 nodes to quantify diffusion properties along the pathway, making this paper an example of the approach cited. (LLM, conf=0.94)

  cites Yeatman et al., 2012 source

In the paper

“Indeed, hubs identified with these measures have been found to entail high metabolic expenditures and be disproportionately implicated in brain disorders (Crossley et al., 2014; Fornito et al., 2015).”

What the sources say

• Supported

  The Discussion section (‘Hub pathology: high value/high cost brain regions’) reports that high-degree and highly efficient hub nodes have higher blood flow and glucose metabolic rates than non-hub nodes. The ‘Hubs and lesions’ section identifies degree centrality as the paper’s primary measure of hubness. Page numbers were not supplied in our copy. (LLM, conf=0.92)

  cites Crossley et al., 2014 source
• ?Not checked — no full text of the sourcecites Fornito et al., 2015 source

In the paper

“In the postnatal stage, in vivo tractography offers a superior tool to chart the spatiotemporal maturation patterns of the reconstructed tracts by employing quantitative dMRI metrics from the infancy period (Dubois et al., 2014; Gilmore et al., 2018; Lebel and Deoni, 2018; Qiu et al., 2015; Vasung et al., 2019) to the adolescent stage (Lebel et al., 2019; Tamnes et al., 2018).”

What the sources say

• ?Not checked — no full text of the sourcecites Dubois et al., 2014 source
• Partly supported

  The Introduction (White matter; page unavailable) says diffusion imaging visualizes major fibre tracts and measures their microstructure using AD, RD and FA. Structural development (White matter; page unavailable) describes how those measures change from infancy through later childhood and adolescence. The paper does not establish that in vivo tractography is a superior tool; the citing sentence assigns adolescence to other references. (LLM, conf=0.82)

  cites Gilmore et al., 2018 sourceExperimental — accuracy not yet measured
• Partly supported

  The Introduction (§2.2; page unavailable) says DTI provides quantitative diffusion metrics and allows reconstruction of fibre architecture through tractography. Developmental Timing (§3.3; page unavailable) uses tractography and FA values to describe maturation across childhood and adolescence, and the review also covers infancy. The copy does not establish that tractography is a superior tool for this purpose; the Conclusions instead describe DTI as powerful but limited. (LLM, conf=0.82)

  cites Lebel and Deoni, 2018 sourceExperimental — accuracy not yet measured

In the paper

“At the mesocale, structural connectivity can be partitioned into modules (also known as communities) that reflect the presence of clusters embedded in the brain’s wiring (Betzel, 2020; Sporns and Betzel, 2016).”

What the sources say

• Supported

  The Introduction describes communities or modules as clusters at the network mesoscale and says community-detection methods apply to structural connectivity. The Background also reports modular structure in structural brain networks, supporting the claim about clusters in the brain’s wiring. (LLM, conf=0.94)

  cites Betzel, 2020 source
• Supported

  The Introduction defines modules, also called network communities, as clusters of densely connected nodes obtained by decomposing a network. The Anatomical Networks section reports community structure in human brain networks, and the Conclusion states that modules exist in structural brain networks. (LLM, conf=0.93)

  cites Sporns and Betzel, 2016 source

In the paper

“In the past decade, researchers have implemented dMRI tractography investigations across broad brain developmental stages ranging from the middle fetal stage to adulthood (Cao et al., 2017; Dubois et al., 2014; Gilmore et al., 2018; Huang, 2010; Lebel and Deoni, 2018; Lebel et al., 2019; Qiu et al., 2015; Zhao et al., 2019b).”

What the sources say

• Partly supported

  The section “Early Development of the Human Brain from a Connectome Perspective” describes studies using diffusion MRI from 20 postmenstrual weeks to age five, and the References list a fetal MR tractography study. The paper therefore supports investigation beginning around mid-gestation, but it does not describe a span of dMRI tractography investigations extending to adulthood; its discussion of adult tractography is a methodological comparison. (LLM, conf=0.87)

  cites Cao et al., 2017 sourceExperimental — accuracy not yet measured
• ?Not checked — no full text of the sourcecites Dubois et al., 2014 source
• Partly supported

  The White matter sections describe diffusion imaging of fibre tracts at birth and DTI studies of development through later childhood and adolescence; the References also list a DTI study spanning infancy to adulthood. Our copy does not document dMRI tractography investigations beginning in the middle fetal stage, so it does not support the full age range claimed. (LLM, conf=0.82)

  cites Gilmore et al., 2018 sourceExperimental — accuracy not yet measured

In the paper

“These results are consistent with classical histological studies (Mitter et al., 2015; Vasung et al., 2010).”

What the sources say

• Supported

  The Conclusion (page unavailable) says histology-based tractography of fetal commissural and internal capsule fibers corresponds remarkably well to many in utero tractography results. This supports the attributed consistency between fetal tractography and histological findings, within the fiber tracts studied. (LLM, conf=0.88)

  cites Mitter et al., 2015 source
• ?Not checked — no full text of the sourcecites Vasung et al., 2010 source

In the paper

“At the local level, node centrality measures provide a useful characterization of the integrative importance of individual brain regions (Sporns et al., 2007).”

What the sources say

• Supported

  The Introduction says centrality measures capture a vertex’s structural importance relative to the rest of the network and uses them to characterize individual brain regions. The Methods section explains that central vertices can serve as connectors or influence many other vertices, while the Discussion links high centrality to measures of information flow and integration. Together, these support the citing sentence’s characterization of integrative importance. (LLM, conf=0.93)

  cites Sporns et al., 2007 source

In the paper

“Indeed, hubs identified with these measures have been found to entail high metabolic expenditures and be disproportionately implicated in brain disorders (Crossley et al., 2014; Fornito et al., 2015).”

What the sources say

• Supported

  The Discussion (‘Hubs and lesions’) says the authors used degree centrality and participation coefficient to identify hubs and found that lesions were more likely in those regions. It also reports that lesions were disproportionately concentrated in rich-club hubs. The finding is general, not a claim that every disorder targets hubs. (LLM, conf=0.94)

  cites Crossley et al., 2014 source
• ?Not checked — no full text of the sourcecites Fornito et al., 2015 source

In the paper

“In the early postnatal stage, ongoing myelination is a clear cause (Paydar et al., 2014)”

What the sources say

• Supported

  The Discussion states that increasing anisotropy in white-matter tracts reflects progressing myelination and that myelination is the dominant contributor to these changes in the first two years of life. This supports the claim about early postnatal tract-related changes. (LLM, conf=0.92)

  cites Paydar et al., 2014 source

In the paper

“After birth, the left anterior cingulate gyrus and left superior occipital gyrus become hubs in toddler brains (Huang et al., 2015).”

What the sources say

• Supported

  The Results section states that the left anterior cingulate gyrus and left superior occipital gyrus become hubs only in the toddler and preadolescent groups, not in neonates. This supports their emergence as hubs after birth, including in toddler brains. (LLM, conf=0.94)

  cites Huang et al., 2015 source

In the paper

“hubs expand into the inferior frontal and insula regions at term age (Ball et al., 2013; Van Den Heuvel et al., 2015b).”

What the sources say

• ?Could not verify

  The Whole-Cortex Survey (page unavailable) mentions inferior frontal cortex and the insula as locations in diffusivity clusters, but does not identify network hubs or track their expansion. The Discussion (page unavailable) discusses cortical microstructural maturation and only speculates about its relationship to functional connectivity. Searches for hubs expanding into these regions at term found no supporting passage. (LLM, conf=0.72)

  cites Ball et al., 2013 sourceExperimental — accuracy not yet measured
• ?Not matched

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

  cites Van Den Heuvel et al., 2015b

In the paper

“The centrality of the pre-cuneus and cuneus still increases before pre-adolescence (Huang et al., 2015; Yap et al., 2011).”

What the sources say

• Supported

  The Results report significantly and monotonically increased normalized nodal efficiency in the bilateral precuneus and cuneus across neonate, toddler, and preadolescent groups. This supports the claim about increasing centrality through the period before preadolescence. (LLM, conf=0.91)

  cites Huang et al., 2015 source
• Partly supported

  The Results section on pediatric network organization reports high centrality for the precuneus and changes in nodal roles across the first two years of life (Fig. 9; no page given). It does not establish that centrality increases for either the precuneus or cuneus, or that such an increase continues up to pre-adolescence. (LLM, conf=0.83)

  cites Yap et al., 2011 sourceExperimental — accuracy not yet measured

In the paper

“In later childhood and adolescence, results seem to be attributed to changes of neurite density, particularly axonal packing (Lebel and Deoni, 2018).”

What the sources say

• Supported

  Section 3.5, “Interpretation of Changes,” directly states that white matter microstructural maturation in later childhood and adolescence appears to be driven by changes in neurite density, particularly axonal packing. The citing sentence preserves the source’s tentative wording. (LLM, conf=0.95)

  cites Lebel and Deoni, 2018 source

In the paper

“Studies also characterize several typical nonlinear trajectories (piecewise, exponential and quadratic) of diffusion metrics for major tracts across large developmental ranges such as infancy to childhood (Pecheva et al., 2017; Reynolds et al., 2019; Stephens et al., 2020) or adolescence to adulthood (Chang et al., 2015; Lebel and Beaulieu, 2011).”

What the sources say

• Partly supported

  The Methods and Results sections examine age-related diffusion metrics in major white-matter tracts of preterm infants. But the paper uses linear regression against post-menstrual age, not piecewise, exponential, or quadratic trajectories, and does not study a range from infancy to childhood. No page numbers are available in our copy. (LLM, conf=0.86)

  cites Pecheva et al., 2017 sourceExperimental — accuracy not yet measured
• Partly supported

  The Results (§3.1; page unavailable) report quadratic trajectories for FA and MD in several major tracts, supporting the nonlinear-trajectory claim. Reynolds et al. studied children aged 2–8 years and fitted linear and quadratic models; our copy does not support attributing piecewise or exponential models, or an infancy-to-childhood age span, to this study. (LLM, conf=0.84)

  cites Reynolds et al., 2019 sourceExperimental — accuracy not yet measured
• Partly supported

  The Tract Trajectories section (page not provided) reports diffusion-metric trajectories for major tracts from birth to age 6, with much faster change in the first year than later years. The Statistical Analysis section (page not provided) says these trajectories were plotted from age-specific averages; our copy does not identify them as piecewise, exponential, or quadratic models. (LLM, conf=0.85)

  cites Stephens et al., 2020 sourceExperimental — accuracy not yet measured

In the paper

“In the fetal brain, it is probably caused by pre-myelination phases such as proliferation and maturation of oligodendrocyte and progenitor cells, the glia that form the myelin sheath (Jakovcevski et al., 2009).”

What the sources say

• Partly supported

  The sections on oligodendrocyte lineage progression and axonal signaling describe proliferating progenitors and the maturation of precursors into pre-myelinating oligodendrocytes in the fetal forebrain before myelination begins (page numbers unavailable). They do not establish that these processes cause the tract-related changes in the citing claim. (LLM, conf=0.84)

  cites Jakovcevski et al., 2009 sourceExperimental — accuracy not yet measured

In the paper

“Early tractography studies usually estimated the aging effect on specific fiber bundles such as the corpus callosum (Sullivan et al., 2006), uncinate fasciculus (Hasan et al., 2009a), and fornix (Stadlbauer et al., 2008b), which show consistent changes including decreased FA, increased MD and decreased tract volume.”

What the sources say

• Partly supported

  The Results report lower FA and higher ADC in older adults’ callosal fiber bundles; higher ADC supports the cited increase in mean diffusivity. They also report fewer identified fibers, but do not report decreased tract volume. The Discussion (p. 1035) says overall corpus callosum size did not differ by age. (LLM, conf=0.88)

  cites Sullivan et al., 2006 sourceExperimental — accuracy not yet measured
• ?Not matched

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

  cites Hasan et al., 2009a
• ?Not matched

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

  cites Stadlbauer et al., 2008b

In the paper

“although neonatal brains already show similar structural hub distribution as adult brains”

What the sources say

• Partly supported

  The Results sections on broad-scale networks and nodal roles report connection hubs in the neonatal network, but they do not compare the neonatal hub distribution with that of adults. The nodal-roles Results instead report substantial changes over the first two years; the Discussion’s reference to adult-like patterns concerns two-year-olds, not neonates. No page numbers were provided in our copy. (LLM, conf=0.83)

  cites Yap et al., 2011 sourceExperimental — accuracy not yet measured

In the paper

“the age when FA reaches its highest peak varies across different white matter tracts.”

What the sources say

• Supported

  The Results (Age effects; page unavailable) describe unique temporal variation among the measured white matter tracts and report that FA follows an inverted U-shaped age curve. The Results also state that the fitted FA results list an age at peak for each tract, supporting the claim that peak age varies across tracts. (LLM, conf=0.88)

  cites Hasan et al., 2010 source
• ?Not matched

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

  cites 2009b
• ?Not checked — no full text of the sourcecites Lebel et al., 2010 source

In the paper

“many refinements occur for the topological roles for each node (Yap et al., 2011).”

What the sources say

• Supported

  The Results section on modularity and local-to-distributed organization reports “relatively large changes in nodal topological roles over age” during the first years of life. The paper compares neonatal, 1-year-old and 2-year-old networks, supporting the paraphrase about refinements after the neonatal period. (LLM, conf=0.92)

  cites Yap et al., 2011 source

In the paper

“By combining tractography atlases with intraoperative electrical stimulation for functional mapping, researchers have constructed probabilistic atlases of white matter pathways and structural hubs that define a minimal common brain connectome with a low inter-individual variability and a low potential of post-lesional compensation (Herbet et al., 2016; Sarubbo et al., 2020).”

What the sources say

• ?Not checked — no full text of the sourcecites Herbet et al., 2016 source
• Partly supported

  Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  cites Sarubbo et al., 2020 sourceExperimental — accuracy not yet measured

In the paper

“and loss of frontal hubs that may be driven by the vulnerable prefrontal tracts (Zhao et al., 2015).”

What the sources say

• Supported

  The Rich-Club Organization discussion (p. 3788) reports a heavy loss of frontal hubs and their connections around age 70 and links that finding to evidence that frontal connections are more susceptible to aging. The discussion of brain regions and connections also describes anterior white matter as especially susceptible to aging. These passages support the citing sentence’s tentative explanation, though the paper says the biological mechanism needs further study. (LLM, conf=0.84)

  cites Zhao et al., 2015 source

In the paper

“In psychiatric surgery, advanced tractography facilitates the combination between brain connectomic analyses and neuromodulation interventions (Davidson et al., 2020; Horn et al., 2017; Riva-Posse et al., 2018)”

What the sources say

• Supported

  The Introduction describes tractography as reconstructing white-matter pathways for psychiatric neurosurgery and says this work has led to circuit-based approaches. The Conclusion states that tractography-based targeting is used in deep brain stimulation for depression. Together, these passages support the claimed combination of connectomic analysis and neuromodulation. (LLM, conf=0.90)

  cites Davidson et al., 2020 source
• Partly supported

  The Abstract > Interpretation and Subjects and Methods > Connectivity Estimation (page numbers unavailable) describe combining diffusion tractography and functional connectome analyses with deep brain stimulation. But the study tests this approach in Parkinson disease, not psychiatric surgery; its references to psychiatric applications do not establish the claim in that setting. (LLM, conf=0.87)

  cites Horn et al., 2017 sourceExperimental — accuracy not yet measured
• Supported

  The Abstract describes using a tractography-based connectome blueprint to plan individualized subcallosal cingulate deep brain stimulation for treatment-resistant depression. The Discussion says tractography-based targeting enables more precise modulation of patient-specific brain circuitry, supporting the claimed combination of connectomic analysis and neuromodulation in this psychiatric-surgery example. (LLM, conf=0.92)

  cites Riva-Posse et al., 2018 source

In the paper

“The brain connectomic concept benefits traditional surgical strategy by a potential shift from focusing on local topography to a network-guided “oncological disconnection surgery” (Duffau, 2021; Hart et al., 2020).”

What the sources say

• ?Not checked — no full text of the sourcecites Duffau, 2021 source
• Partly supported

  Section 3c (p. 9) describes connectomics as a holistic view of brain connectivity and discusses glioblastoma analyses that reveal surgical risks beyond standard structural imaging. The pre-operative glioma example (Figure 6 caption, p. 23) also illustrates network mapping. These passages support a potential move beyond local anatomy, but they do not propose a network-guided “oncological disconnection surgery.” (LLM, conf=0.85)

  cites Hart et al., 2020 sourceExperimental — accuracy not yet measured

In the paper

“a recent large sample study implicates the global dimensions of variation in the human structural connectome in aging-related cognitive decline (Madole et al., 2020).”

What the sources say

• Supported

  The Abstract’s Conclusions explicitly say that the results implicate global dimensions of variation in the human structural connectome in aging-related cognitive decline. The Introduction describes a sample of over 8,000 UK Biobank participants, supporting the description of the study as large-sample. (LLM, conf=0.95)

  cites Madole et al., 2020 source

In the paper

“brain regions within the default mode network, which is involved broadly in high-level cognition (Raichle, 2015; Vatansever et al., 2018; 2017), display the latest maturation age of peak.”

What the sources say

• ?Not checked — no full text of the sourcecites Raichle, 2015 source
• Supported

  The citation is attached to the clause describing the default mode network’s broad involvement in high-level cognition, not to the subsequent claim about maturation age. The Abstract explicitly reports broader DMN involvement in higher cognition and its engagement during visuospatial planning. It does not address maturation age, but that point is not attributed to this citation. (LLM, conf=0.93)

  cites Vatansever et al., 2018 source
• ?Not matched

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

  cites [2017]

In the paper

“Lifespan changes of the connectome also show good consistency with cognitive maturation (Zhao et al., 2015).”

What the sources say

• Supported

  In “Age-Related Alterations in the Global Topology of the WM Network” (page not provided), the authors report that connectome measures peak at ages within the reported range for peak cognitive performance. This supports the claimed consistency between lifespan connectome changes and cognitive maturation, although the paper does not report a direct cognitive assessment. (LLM, conf=0.87)

  cites Zhao et al., 2015 source

In the paper

“Results from a study that employ an advanced diffusion model (NODDI) indicate that the age-related declines in FA are primarily induced by declines in neurite density rather than changes in tract complexity (Cox et al., 2016).”

What the sources say

• Partly supported

  The Results (General factors of white matter microstructure) report that most of the statistical mediation of the age–FA association occurred through neurite density rather than tract complexity. The paper says this suggests that declining FA *may* predominantly be explained by declining neurite density; the citing sentence’s “induced by” implies a stronger causal conclusion than the paper states. (LLM, conf=0.84)

  cites Cox et al., 2016 sourceExperimental — accuracy not yet measured

In the paper

“researchers have found that the location and surrounding damages of lesions contribute more than the number and volume of lesions (Lipp et al., 2020).”

What the sources say

• Supported

  The Introduction states that lesion number and volume correlate poorly with disability scores, while lesion location and diffuse microstructural damage outside lesions better explain disability. This supports the comparison in the citing sentence; the source says damage outside lesions rather than specifically damage immediately surrounding them. (LLM, conf=0.90)

  cites Lipp et al., 2020 source

In the paper

“This limitation is often expressed as the mantra “streamline count is not quantitative.””

• Also stated in: § Connection density biases · boosted

  “highlights the inappropriateness of applying such a strong biological interpretation to raw streamline count.”

What the sources say

• ?Not checked — no full text of the sourcecites Jones, 2010 source
• ?Not checked — no full text of the sourcecites Jones et al., 2013 source

In the paper

“Diffusion magnetic resonance imaging (dMRI) tractography is an advanced imaging technique that enables in vivo reconstruction of the brain’s white matter connections at macro scale.”

• Also stated in: § Introduction · boosted

  “Diffusion magnetic resonance imaging (dMRI) tractography is an imaging method that uniquely enables in vivo reconstruction of the brain’s white matter connections at macro scale (Basser et al., 2000).”

What the sources say

• ?Not checked — no full text of the sourcecites Basser et al., 2000 source

In the paper

“indeed, the diffusion-weighted signal itself does not provide direct evidence of fiber terminations (Smith et al., 2020b).”

What the sources say

• ?Not matched

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

  cites Smith et al., 2020b

In the paper

“For the methodology review, we organize the paper according to three main processing steps that are common across most approaches for quantitative analysis of tractography.”

In the paper

“While many research studies have been performed to compare different tractography algorithms (Bastiani et al., 2012; Fillard et al., 2011; Petrov et al., 2017; Pujol et al., 2015; Sinke et al., 2018; Wilkins et al., 2015; Zhan et al., 2015), there is no consensus on “the best algorithm.””

What the sources say

• ?Not checked — no full text of the sourcecites Bastiani et al., 2012 source
• ?Not checked — no full text of the sourcecites Fillard et al., 2011 source
• ?Not matched

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

  cites Petrov et al., 2017

In the paper

“Where such an issue can be clearly demonstrated in exemplar data, and a tailored mechanism proposed that intrinsically addresses the issue, it is expected that the employment of such a mechanism in the reconstruction of empirical image data will lead to overall improved biological accuracy of the reconstruction (even in the absence of direct validation of such)”

In the paper

“resulting in a substantial proportion of streamlines from whole-brain tractography being unreasonable for quantification of white matter fiber connectivity (Yeh et al., 2016a).”

What the sources say

• ?Not matched

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

  cites Yeh et al., 2016a

In the paper

“The capacity for thresholding to differentiate between true and spurious connections remains a debated topic in the field.”

In the paper

“These common termination criteria are therefore not adequate to ensure biologically plausible streamline generation”

What the sources say

• ?Not checked — no full text of the sourcecites Smith et al., 2012 source

In the paper

“Critically, this step enables cross-subject quantitative comparison of the segmented white matter pathways; see (Smith et al., 2014) for a review of general considerations in multi-subject dMRI studies.”

What the sources say

• ?Not checked — no full text of the sourcecites Smith et al., 2014 source

In the paper

“Many streamline labeling methods use streamline-based registration to directly align individual tractography data to a tractography atlas where a reference tract segmentation is available.”

In the paper

“We conclude that, while there have been considerable advancements in methodological technologies and breadth of applications, there nevertheless remains no consensus about the “best” methodology in quantitative analysis of tractography”

In the paper

“First, given the fact that there is no ground truth in tractography, validation of a tract identification method is difficult.”

In the paper

“The fifth point is that selection of different tract identification methods may require different image registration methods.”

In the paper

“The cortical-parcellation-based methods are more widely used as they enable construction of a connectivity matrix and its subsequent analysis using techniques from graph theory (as described in Section 5.6) (Bassett and Bullmore, 2017; Bullmore and Sporns, 2009a; Gong et al., 2009a; Ingalhalikar et al., 2014; Sporns et al., 2005; Yeh et al., 2016b; Zalesky et al., 2012).”

What the sources say

• ?Not matched

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

  cites Bassett and Bullmore, 2017
• ?Not matched

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

  cites Bullmore and Sporns, 2009a
• ?Not matched

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

  cites Gong et al., 2009a

In the paper

“One factor that can affect the consistency of methods that segment tractography streamline data is the underlying tractography method that is employed.”

In the paper

“The ROI-based methods usually require volumetric registration (e.g., between a population T1-weighted template and individual dMRI data) to align ROIs to the dMRI space.”

In the paper

“Compared to the cortical-parcellation-based methods that focus on streamline terminal regions, fiber clustering methods leverage the full lengths of the streamline trajectories.”

In the paper

“Currently, there is no consensus about which brain parcellation technique could be most useful (de Reus and Van den Heuvel, 2013; Yeh et al., 2020; Zalesky et al., 2012).”

What the sources say

• ?Not checked — no full text of the sourcecites de Reus and Van den Heuvel, 2013 source
• ?Not matched

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

  cites Yeh et al., 2020
• ?Not checked — no full text of the sourcecites Zalesky et al., 2012 source

In the paper

“As a result, the cortical-parcellation-based methods can erroneously group streamlines that follow completely different trajectories through the white matter but nevertheless end up at the same gray matter endpoints, whereas fiber clustering can be insensitive to streamlines diverging at the cortex if they pass through the same deep white matter regions.”

In the paper

“As a result, the cortical-parcellation-based methods can erroneously group streamlines that follow completely different trajectories through the white matter but nevertheless end up at the same gray matter endpoints, whereas fiber clustering can be insensitive to streamlines diverging at the cortex if they pass through the same deep white matter regions.”

In the paper

“In general, fiber clustering methods start with a computation of pairwise streamline geometric similarities, followed by a computational clustering method to group similar streamlines into clusters.”

In the paper

“This can enable the study of the tissue microstructure in local regions along the fiber pathway.”

In the paper

“The most commonly adopted strategy for extraction of quantitative measures is to compute a scalar value as a summary statistic of the fiber pathway.”

In the paper

“Usually, this is done by computing a model in each voxel to derive a 3D microstructural map (e.g., an FA image) and then using tractography to define the locations from which the values of this microstructural measure should be sampled.”

In the paper

“While in its original implementation SIFT (Smith et al., 2013) provides only the NOS adjusted based on voxel-wise spherical deconvolution, its evolution SIFT2 (Smith et al., 2015b) optimizes per-streamline cross-section multipliers to match a whole-brain tractogram to fixel-wise fiber densities and thus provides a value for each streamline that can be used (for example) to define the edge weights”

What the sources say

• ?Not checked — no full text of the sourcecites Smith et al., 2013 source

In the paper

“For microstructural measures, the mean measure within the fiber pathway is the most widely used.”

In the paper

“This approach requires definition of a coordinate system, sampling framework, or surface-based representation to define how to sample the microstructural data of interest (Bells et al., 2011a; Chen et al., 2016b; Colby et al., 2012b; Corouge et al., 2006; Goodlett et al., 2009; Irimia et al., 2020; Jones et al., 2005; O’Donnell et al., 2009; Yeatman et al., 2012; Yeh, 2020).”

What the sources say

• ?Not matched

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

  cites Bells et al., 2011a
• ?Not matched

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

  cites Chen et al., 2016b
• ?Not matched

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

  cites Colby et al., 2012b

In the paper

“Other approaches for tractography quantification leverage information from imaging modalities other than dMRI, such as T1-weighted (T1w) imaging (Boshkovski et al., 2020a; Yeatman et al., 2014a), myelin sensitive maps (Lee et al., 2020; Mancini et al., 2020; Piredda et al., 2021) and g-ratio (Campbell et al., 2018).”

What the sources say

• ?Not matched

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

  cites Boshkovski et al., 2020a
• ?Not matched

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

  cites Yeatman et al., 2014a
• ?Not checked — no full text of the sourcecites Lee et al., 2020 source

In the paper

“The so called filtering methods discussed in Section 3.3 can also be used to perform “microstructure-informed” tractography with the aim to extract quantitative measures of white matter fiber pathways and reduce potential bias in the estimation of connectivity.”

In the paper

“Yeh et al., 2016a), several other factors (such as q-space sampling schemes, fiber orientation models, and tractography algorithms) that are not specifically reviewed in our paper, have also been shown to affect most network measures derived from the connectivity matrix.”

What the sources say

• ?Not matched

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

  cites Yeh et al., 2016a

In the paper

“Myriad components of the processing pipeline can also influence the resulting connectome matrix data, even down to the level of detail of the mechanism by which streamlines are assigned to those parcels (Yeh et al., 2019a).”

What the sources say

• ?Not matched

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

  cites Yeh et al., 2019a

In the paper

“However, it is important to highlight that such multiple comparison correction methods, which assume independence across data samples, can be overly conservative when applied to tests performed on multiple tracts and/or quantitative measurements that are not independent.”

In the paper

“After mapping microstructural measures along each point of the streamlines along the fiber pathway, a widely used strategy is to analyze data along the tract profile.”

In the paper

“For example, MD and FA are not mutually mathematically orthogonal measures (Ennis and Kindlmann, 2006).”

What the sources say

• ?Not checked — no full text of the sourcecites Ennis and Kindlmann, 2006 source

In the paper

“For example, the combination of structural and functional connectivity measures has been shown to increase predictive power (Whitfield-Gabrieli et al., 2016).”

What the sources say

• ?Not checked — no full text of the sourcecites Whitfield-Gabrieli et al., 2016 source

In the paper

“At a global scale, graph measures can be used to describe how the brain’s structural wiring supports the communication of information between distant ROIs and cognitive systems (Avena-Koenigsberger et al., 2018; Mišić et al., 2015).”

What the sources say

• ?Not checked — no full text of the sourcecites Avena-Koenigsberger et al., 2018 source
• ?Not matched

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

  cites Mišić et al., 2015

In the paper

“Due to recent advances in methodological progress such as motion correction algorithms, in vivo tractography on the fetal brain has become available and exhibit highly agreement with ex vivo and neuroanatomical studies (Christiaens et al., 2021; Khan et al., 2019), which indicates strong potential for future prenatal brain investigations.”

In the paper

“Nonetheless, recent work has shown that advanced communication measures can elucidate how the organization of structural connectivity shapes the brains functional dynamics (Goñi et al., 2014; Imms et al., 2021; Mišic et al., 2018; Wang et al., 2019).”

What the sources say

• ?Not matched

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

  cites Goñi et al., 2014
• ?Not checked — no full text of the sourcecites Imms et al., 2021 source
• ?Not matched

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

  cites Mišic et al., 2018

In the paper

“Centrality measures find particular utility in identifying hub regions that may play a central role in integrative brain function (van den Heuvel and Sporns, 2013).”

What the sources say

• ?Not matched

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

  cites van den Heuvel and Sporns, 2013

In the paper

“Therefore, the use of advanced communication measures enables the study of directional and asymmetric functional interactions from structural connectivity mapped with dMRI.”

In the paper

“In general, the anterior and posterior parts of tracts peak earlier than the central part of tracts, around 20 to 40 years old (Lebel et al., 2010), supporting a classical “last-in-first-out” theory.”

What the sources say

• ?Not checked — no full text of the sourcecites Lebel et al., 2010 source