40 Years of diffusion MRI in the brain: From history to emerging frontiers

As diffusion time increases, water molecules probe larger portions of the tissue environment and encounter more barriers or opportunities for exchange.

“As diffusion time increases, water molecules probe larger portions of the tissue environment and encounter more barriers or opportunities for exchange (Le Bihan, 1995; Novikov et al., 2019).”

• cites Le Bihan, 1995 sourceNot checked — no full text of the source
• cites Novikov et al., 2019 sourceSupported

In imaging practice, the ideal narrow-pulse Stejskal-Tanner equation is insufficient because imaging gradient pulses, crusher gradient pulses, and their cross-terms with diffusion-encoding gradients can contribute substantially to diffusion signal attenuation.

“In imaging practice, however, the ideal narrow pulse Stejskal–Tanner equation is insufficient because imaging gradient pulses, crusher gradient pulses, and their cross-terms with the diffusion-encoding gradient pulses can contribute substantially to the diffusion signal attenuation (Cohen et al., 2015; Le Bihan & Breton, 1985; Rashid et al., 2025).”

• cites Cohen et al., 2015 sourceCould not verify

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

  Experimental — accuracy not yet measured
• cites Le Bihan & Breton, 1985Not matched

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

• cites Rashid et al., 2025 sourceSupported

Brain diffusion MRI must be interpreted simultaneously through physics, biology, image contrast, and inverse problem theory.

“Brain diffusion MRI must, therefore, be interpreted simultaneously through physics, biology, image contrast, and inverse problem theory, making it a rich and powerful imaging modality (Jelescu & Budde, 2017; Le Bihan, 1995; Novikov et al., 2019).”

• cites Jelescu & Budde, 2017Not matched

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

• cites Le Bihan, 1995 sourceNot checked — no full text of the source
• cites Novikov et al., 2019 sourceSupported

DTI suggested that macroscopic structural connectivity could be charted in vivo, even if only approximately.

“DTI suggested that macroscopic structural connectivity could be charted in vivo, even if only approximately (Assaf et al., 2019; Hagmann et al., 2008).”

• cites Assaf et al., 2019 sourceSupported
• cites Hagmann et al., 2008 sourceSupported

The first tractography demonstrations offered a three-dimensional, noninvasive picture of large-scale anatomical connections in living humans.

“The first demonstrations (Basser et al., 2000; Conturo et al., 1999; Jones et al., 1999; Mori et al., 1999; Poupon et al., 2000) immediately captured the imagination of the field because they offered something previously unattainable in living humans: a three-dimensional, noninvasive picture of large-scale anatomical connections (Fig. 4).”

• cites Basser et al., 2000 sourceNot checked — no full text of the source
• cites Conturo et al., 1999 sourceSupported
• cites Jones et al., 1999 sourceNot checked — no full text of the source

IVIM broadened diffusion MRI from a method of tissue microstructure toward a more general method capable of quantitative imaging of perfusion parameters (e.g., blood volume and blood flow through fIVIM and D*) without exogenous tracers or contrast agents.

“The conceptual importance of IVIM for brain neuroimaging has been substantial, broadening diffusion MRI from a method of tissue microstructure toward a more general method capable of quantitative imaging of perfusion parameters (e.g., blood volume and blood flow through fIVIM and D*) without the need for exogenous tracers or contrast agents (Iima & Le Bihan, 2016; Le Bihan, 2019).”

• cites Iima & Le Bihan, 2016 sourceNot checked — no full text of the source
• cites Le Bihan, 2019 sourcePartly supported

  Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured

Noise floor effects, Gibbs ringing, eddy currents, susceptibility distortions, and free-water partial volume can bias tensor estimates unless explicitly controlled.

“Noise floor effects, Gibbs ringing, eddy currents, susceptibility distortions, and free-water partial volume can all bias tensor estimates unless explicitly controlled (Beaulieu, 2002; Le Bihan et al., 2001; Tournier et al., 2011).”

• cites Beaulieu, 2002 sourceNot checked — no full text of the source
• cites Le Bihan et al., 2001 sourceNot checked — no full text of the source
• cites Tournier et al., 2011 sourcePartly supported

  The Data Processing and Quality Control sections describe eddy-current distortions that alter the effective b-value and identify Gibbs ringing as an artifact associated with implausible tensor-derived FA values. The acquisition section also discusses susceptibility artifacts, and Mixture Models discusses accounting for CSF contamination. Our copy does not establish the specific noise-floor effect or explicitly say that free-water partial volume biases tensor estimates, so the full list is not supported. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

Structural connectome analyses are most convincing when integrated with neuroanatomical knowledge, functional data, and careful control of tractography biases.

“Structural connectome analyses are most convincing when integrated with neuroanatomical knowledge, functional data, and careful control of tractography biases (Assaf et al., 2019; Daducci et al., 2016).”

• cites Assaf et al., 2019 sourceSupported
• cites Daducci et al., 2016 sourcePartly supported

  Section 3.2 (Biological plausibility; page unavailable) recommends comparing tractography estimates with known anatomy, while the Introduction and Section 5 warn that overlooked tractography pitfalls can bias connectivity analyses. The passages do not recommend integrating structural analyses with functional data; a search for functional MRI, functional data, and functional connectivity found no supporting passage. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

The Standard Model framework underlies, or is closely related to, several influential models and approaches, including CHARMED, AxCaliber, NODDI, WMTI, SMT, and related multi-compartment formulations.

“This framework underlies, or is closely related to, several influential models and approaches, including CHARMED, AxCaliber, NODDI, WMTI, SMT, and related multi-compartment formulations (Assaf & Basser, 2005; Assaf et al., 2008; Zhang et al., 2012).”

• cites Assaf & Basser, 2005 sourceNot checked — no full text of the source
• cites Assaf et al., 2008 sourceSupported
• cites Zhang et al., 2012 sourceNot checked — no full text of the source

Tractography visualizes a model of anatomical connectivity constrained by diffusion data, not anatomy itself.

“Tractography, therefore, visualizes a model of anatomical connectivity constrained by diffusion data, not anatomy itself (Daducci et al., 2016; Tournier et al., 2011).”

• cites Daducci et al., 2016 sourceSupported
• cites Tournier et al., 2011 sourceSupported

Q-ball imaging and diffusion spectrum imaging use richer q-space sampling to estimate an orientation distribution function representing the relative prominence of water diffusion along different spatial directions within a voxel.

“Methods such as Q-ball imaging (Tuch, 2004) and diffusion spectrum imaging (DSI) (Wedeen et al., 2005) use richer q-space sampling to estimate an orientation distribution function (ODF), representing the relative prominence of water diffusion along different spatial directions within a voxel.”

• cites Tuch, 2004 sourceNot checked — no full text of the source
• cites Wedeen et al., 2005 sourceSupported

Probabilistic approaches have been used to create the first atlases of large-scale brain connections.

“Probabilistic approaches have also been used to create the first atlases of large-scale brain connections (Behrens et al., 2003; Hagmann et al., 2003; Parker et al., 2003).”

• cites Behrens et al., 2003 sourceNot checked — no full text of the source
• cites Hagmann et al., 2003 sourceSupported
• cites Parker et al., 2003 sourceNot checked — no full text of the source

Tractography is one of the defining achievements of the brain-centered diffusion MRI story.

“In that sense, tractography is one of the defining achievements of the brain-centered diffusion MRI story (Assaf et al., 2019; Gong, et al., 2009).”

• cites Assaf et al., 2019 sourceSupported
• cites Gong, et al., 2009 sourceSupported

HARDI acquires diffusion-weighted data along many more directions than the minimum six required for tensor estimation.

“To mitigate the single-tensor crossing-fiber limitation, high-angular-resolution diffusion imaging (HARDI) acquires diffusion-weighted data along many more directions than the minimum six required for tensor estimation (Frank, 2002).”

• cites Frank, 2002 sourceSupported

Diffusion MRI became a foundational tool for the Human Connectome era because it offered a scalable route to individual structural connectivity.

“Diffusion MRI became a foundational tool for the Human Connectome era because it offered a scalable route to individual structural connectivity (Assaf et al., 2019; Hagmann et al., 2003).”

• cites Assaf et al., 2019 sourceSupported
• cites Hagmann et al., 2003 sourcePartly supported

  The Abstract and Results (page numbers unavailable) describe using diffusion MRI to map brain connectivity in vivo in two healthy volunteers, supporting a route to individual structural connectivity. The paper does not establish that the approach was scalable or that it became foundational to the later Human Connectome era. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

Diffusion kurtosis imaging was among the earliest MRI approaches developed to describe empirically the curvature of the diffusion-weighted MRI signal decay with b-value, reflecting the underlying non-Gaussian displacement probability distribution.

“Diffusion kurtosis imaging (DKI) was among the earliest MRI approaches developed to describe empirically the curvature of the diffusion-weighted MRI signal decay with b-value, reflecting the underlying non-Gaussian displacement probability distribution (Jensen et al., 2005).”

• cites Jensen et al., 2005 sourceSupported

Diffusion MRI has enabled in vivo mapping of tract-specific maturation and its links with emerging motor, language, and cognitive functions.

“These metrics should not be interpreted as direct measures of myelin or axonal density, but they have enabled in vivo mapping of tract-specific maturation and its links with emerging motor, language, and cognitive functions (Dubois et al., 2021; Lebel et al., 2019).”

• cites Dubois et al., 2021 sourceSupported
• cites Lebel et al., 2019 sourceNot checked — no full text of the source

Multi-shell approaches, including NODDI, constrained NODDI (C-NODDI), and diffusion kurtosis imaging, provide model-dependent indices related to neurite organization, extracellular water, axonal density, and tissue integrity.

“Multi-shell approaches, including NODDI, constrained NODDI (C-NODDI), and diffusion kurtosis imaging, provide model-dependent indices related to neurite organization, extracellular water, axonal density, and tissue integrity (Alsameen et al., 2023; Raghavan et al., 2021; Schilling et al., 2022).”

• cites Alsameen et al., 2023 sourcePartly supported

  The Introduction (page unavailable) describes NODDI and C-NODDI as models yielding neurite-density and orientation indices, with extracellular water included in the diffusion model. The Methods section (page unavailable) reports a multi-shell acquisition. The cited paper does not establish the sentence’s claim about diffusion kurtosis imaging, which is attributed to the citations collectively. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Raghavan et al., 2021 sourcePartly supported

  The Introduction describes multi-shell NODDI and its indices for neurite density, orientation dispersion, and free-water fraction; the Discussion relates these measures to white matter microstructure. The passages support the NODDI portion, but do not address constrained NODDI (C-NODDI) or diffusion kurtosis imaging. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Schilling et al., 2022 sourcePartly supported

  The Discussion (Age-related microstructural decline; page unavailable) says advanced diffusion models can measure neurite density, extracellular diffusivity, axonal dispersion, and diffusion non-Gaussianity, and links advanced modeling to data with multiple diffusion weightings. It does not identify NODDI, C-NODDI, or diffusion kurtosis imaging by name, or establish the specific claim about extracellular water. The Methods (Processing; page unavailable) report that this study extracted DTI microstructural metrics rather than those named approaches. (LLM, conf=0.87)

  Experimental — accuracy not yet measured

In a standardized multicenter study, diffusion kurtosis combined with dynamic contrast-enhanced perfusion MRI improved diagnostic confidence and glioma-subtype prediction.

“in a standardized multicenter study, diffusion kurtosis combined with dynamic contrast-enhanced perfusion MRI improved diagnostic confidence and glioma-subtype prediction (Richter et al., 2024).”

• cites Richter et al., 2024 sourceSupported

White-matter free water has shown associations with cognitive performance and subsequent decline, particularly in limbic tracts.

“White-matter free water has shown associations with cognitive performance and subsequent decline, particularly in limbic tracts (Nakaya et al., 2022; Peter et al., 2025).”

• cites Nakaya et al., 2022 sourcePartly supported

  The Discussion (page not provided) reports an association between free water and cognitive performance, including a prior finding based on whole-brain white-matter regions. Nakaya et al.’s own cognitive correlations concern cortical regions, particularly the temporal lobe; the cited passages do not establish subsequent cognitive decline or an association particularly in limbic white-matter tracts. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Peter et al., 2025 sourceSupported

These measures reflect a composite of axonal organization, myelin, extracellular water, fiber geometry, and cerebrovascular injury.

“However, these measures reflect a composite of axonal organization, myelin, extracellular water, fiber geometry, and cerebrovascular injury (Beck et al., 2021; Jones et al., 2013).”

• cites Beck et al., 2021 sourcePartly supported

  The discussion passage indexed under Table 5 says diffusion measures can be influenced by axonal density, myelination, extracellular conditions, and fibre architecture. It also mentions blood vessels and perivascular spaces, but does not establish that the measures reflect cerebrovascular injury. No page number was supplied for the indexed passage. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Jones et al., 2013 sourceNot checked — no full text of the source

CHARMED and AxCaliber were influential early attempts to estimate restricted fractions and axon caliber distributions from diffusion data.

“CHARMED and AxCaliber were influential early attempts to estimate restricted fractions and axon caliber distributions from diffusion data (Assaf & Basser, 2005; Assaf et al., 2008).”

• cites Assaf & Basser, 2005 sourceNot checked — no full text of the source
• cites Assaf et al., 2008 sourceSupported

The ENIGMA Schizophrenia DTI Working Group identified widespread, although modest, microstructural differences involving association, commissural, and projection pathways in 4,322 individuals.

“Schizophrenia remains the best-established example: the ENIGMA Schizophrenia DTI Working Group identified widespread, although modest, microstructural differences involving association, commissural, and projection pathways in 4,322 individuals (Kelly et al., 2018).”

• cites Kelly et al., 2018 sourceSupported

More focused work has linked long-distance structural connectivity within the global neuronal workspace to impaired conscious access and psychotic symptoms.

“More focused work has linked long-distance structural connectivity within the global neuronal workspace to impaired conscious access and psychotic symptoms (Berkovitch et al., 2021).”

• cites Berkovitch et al., 2021 sourceSupported

Diffusion-based network mapping supports the view of Parkinson's disease as a somato-cognitive action-network disorder.

“More broadly, diffusion-based network mapping supports the view of Parkinson’s disease as a somato-cognitive action-network disorder (Ren et al., 2026).”

• cites Ren et al., 2026 sourcePartly supported

  The ‘PD as a somato-cognitive action disorder’ section proposes that Parkinson’s disease be understood as a SCAN disorder. But the Methods section (‘Identification of the cortical SCAN regions’) describes resting-state functional connectivity and functional network mapping, not diffusion-based network mapping. The proposed view is supported; the attributed method is not. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

Recent longitudinal work suggests that diffusion-derived axonal-density measures may detect early, spatially selective degeneration related to clinical decline in Alzheimer's disease.

“Recent longitudinal work further suggests that diffusion-derived axonal-density measures may detect early, spatially selective degeneration related to clinical decline in Alzheimer’s disease (Gong et al., 2026; Wang et al., 2023).”

• cites Gong et al., 2026 sourcePartly supported

  The Discussion’s summary of main findings reports that longitudinal diffusion-derived axonal-density measures declined more steeply in cognitively impaired participants and were associated with cognitive and functional decline. However, the Strengths and limitations section says the analysis considered only whole-brain white matter; it describes mapping the spatial distribution of degeneration as a capability for future study, not an established spatially selective finding. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites Wang et al., 2023 sourceNot checked — no full text of the source

Similar activation-related cell-size changes were observed in brain slices, where vascular confounds are absent.

“Similar activation-related cell-size changes were also observed in brain slices, where vascular confounds are absent (Buckley et al., 1999; Flint et al., 2009).”

• cites Buckley et al., 1999 sourceNot checked — no full text of the source
• cites Flint et al., 2009 sourceSupported

The BOLD response does not always track underlying neuronal signaling faithfully and can diverge from activation-driven oxygen metabolism in some regions or conditions.

“Because the BOLD response depends on local changes in cerebral blood flow, blood volume, and oxygenation, it does not always track underlying neuronal signaling faithfully and can diverge from activation-driven oxygen metabolism in some regions or conditions (Epp et al., 2026; Logothetis & Wandell, 2004; Sirotin & Das, 2009).”

• cites Epp et al., 2026 sourceSupported
• cites Logothetis & Wandell, 2004 sourceNot checked — no full text of the source
• cites Sirotin & Das, 2009 sourcePartly supported

  The paper’s Abstract reports that trial-related blood-volume and oxygenation signals in the primary visual cortex of two alert monkeys were poorly predicted by concurrent neuronal recordings. This supports the warning that haemodynamic signals need not faithfully track local neuronal activity. The study used optical imaging, however; it did not directly measure BOLD responses or demonstrate divergence from activation-driven oxygen metabolism. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

Because water homeostasis and intra-/extracellular exchange are central to neuronal physiology, especially in neurons with limited volume regulation, activity-related diffusion changes may reflect an active component of signaling rather than merely a by-product.

“Because water homeostasis and intra-/extracellular exchange are central to neuronal physiology, especially in neurons with limited volume regulation (Agre, 2005), activity-related diffusion changes may reflect an active component of signaling rather than merely a by-product.”

• cites Agre, 2005 sourcePartly supported

  Split into 2 atoms — 0 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 1 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured

T2-weighted spin-echo diffusion MRI retains residual vascular effects related to BOLD.

“T2-weighted spin-echo diffusion MRI retains residual vascular effects related to BOLD (Aso et al., 2009; de Riedmatten, Spencer, Olszowy, et al., 2025; Kershaw et al., 2009; Miller et al., 2007).”

• cites Aso et al., 2009 sourceNot checked — no full text of the source
• cites de Riedmatten, Spencer, Olszowy, et al., 2025 sourceSupported
• cites Kershaw et al., 2009 sourceNot checked — no full text of the source

Diffusion MRI raised the possibility of sensing rapid changes in the intrinsic physical state of neural tissue itself.

“Diffusion MRI, therefore, raised the possibility of sensing rapid changes in the intrinsic physical state of neural tissue itself (Fig. 6) (Le Bihan, 2026).”

• cites Le Bihan, 2026 sourceSupported

Diffusion MRI was shown to detect brain activation through transient decreases in ADC, suggesting sensitivity to intrinsic microstructural changes in addition to possible vascular effects.

“A major conceptual shift occurred in the early 2000s when diffusion MRI was shown to detect brain activation through transient decreases in ADC, suggesting sensitivity to intrinsic microstructural changes in addition to possible vascular effects (Fig. 6) (Darquié et al., 2001; Le Bihan et al., 2006).”

• cites Darquié et al., 2001 sourceNot checked — no full text of the source
• cites Le Bihan et al., 2006 sourcePartly supported

  The Results show that diffusion MRI detected activation-related changes in the human visual cortex, and the Conclusion links them to likely biophysical changes in activated cells. But the Results explicitly caution that the signal increase cannot be converted into a drop in a single diffusion coefficient; the authors instead favor a shift from a fast- to a slow-diffusing water phase. Thus, describing this paper's finding as a transient decrease in ADC overstates what it establishes. (LLM, conf=0.89)

  Experimental — accuracy not yet measured

Given the high density of synapses, dendrites, and cellular processes in cortex, even subtle activity-related dendritic-spine "twitching" may alter the diffusion environment of water molecules.

“Given the high density of synapses (8 108/mm3), dendrites (400 m/mm3), and cellular processes in cortex (Shapson-Coe et al., 2024), even subtle activity-related dendritic-spine “twitching” (Crick, 1982) may alter the diffusion environment of water molecules (Halpain, 2000; Le Bihan, 2007).”

• cites Shapson-Coe et al., 2024 sourcePartly supported

  The Results (Cellular and Synaptic Organization; page unavailable) document abundant synapses and dendritic processes in the reconstructed cortex. But the paper reports about 150 million synapses in an approximately cubic-millimeter sample, not the cited 8 × 10^8/mm³, and the passages do not give 400 m/mm³ of dendrites. The spine movement and water-diffusion suggestion is attributed to other citations in the sentence, not to Shapson-Coe et al. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites Crick, 1982 sourceNot checked — no full text of the source
• cites Halpain, 2000 sourceNot checked — no full text of the source

Preclinical work has provided strong evidence that DfMRI responses are mechanistically distinct from BOLD responses.

“Preclinical work has provided strong evidence that DfMRI responses are mechanistically distinct from BOLD responses (Abe, Tsurugizawa, et al., 2017; Tsurugizawa et al., 2013).”

• cites Abe, Tsurugizawa, et al., 2017Not matched

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

• cites Tsurugizawa et al., 2013 sourceSupported

Because part of its nonvascular mechanism, DfMRI may be less blurred by vascular architecture supplying multiple functional units, as observed with BOLD.

“Because part of its nonvascular mechanism, it may be less blurred by vascular architecture supplying multiple functional units, as observed with BOLD (Sirotin & Das, 2009).”

• cites Sirotin & Das, 2009 sourceCould not verify

  The paper’s Abstract reports anticipatory haemodynamic signals in monkey visual cortex and questions whether haemodynamic signals uniformly reflect local neuronal activity. Searches for DfMRI, vascular blurring, spatial resolution, and vessels supplying multiple functional units found no passage addressing the claimed comparison with BOLD. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

Diffusion-based virtual MRE appears as an attractive, scalable noninvasive biomarker to investigate brain stiffness in the aging brain and for longitudinal or interventional studies of neurodegenerative diseases, such as Alzheimer's disease including intervention stratification.

“Diffusion-based virtual MRE also appears as an attractive, scalable noninvasive biomarker to investigate brain stiffness in the aging brain and for longitudinal or interventional studies of neurodegenerative diseases, such as Alzheimer’s disease including intervention stratification (Le Bihan, 2026).”

• cites Le Bihan, 2026 sourceSupported

Weaver and colleagues showed that intrinsic brain motion generated by cardiac pulsatility can also be exploited to estimate mechanical properties.

“Weaver and colleagues showed that intrinsic brain motion generated by cardiac pulsatility can also be exploited to estimate mechanical properties (Weaver et al., 2012; Zorgani et al., 2015).”

• cites Weaver et al., 2012 sourceSupported
• cites Zorgani et al., 2015 sourceSupported

Functional MRI, whether BOLD- or diffusion-based, and tractography provide complementary views of the functional and anatomical networks supporting cognition.

“Functional MRI, whether BOLD- or diffusion-based, and tractography provide complementary views of the functional and anatomical networks supporting cognition (Gong, He, et al., 2009; Koch et al., 2002).”

• cites Gong, He, et al., 2009 sourcePartly supported

  The Introduction describes fMRI as a way to study functional networks and diffusion tractography as a way to study anatomical connectivity; the Conclusion says the anatomical findings are compatible with functional-network studies. Neither passage describes diffusion-based functional MRI or directly connects these methods to cognition. Page numbers are not provided in our copy. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites Koch et al., 2002 sourcePartly supported

  The Introduction and Discussion compare BOLD-based functional connectivity with diffusion-tensor measures of anatomical connectivity, supporting the claim that the methods provide complementary views. But the paper uses diffusion imaging to assess anatomy, not as a form of functional MRI; it does not support the phrase “diffusion-based” functional MRI. Page numbers are not available in our copy. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

Because neural propagation has a finite maximum speed within the connectome, a relativistic formalism using a brain-specific limiting velocity is required.

“However, because neural propagation has a finite maximum speed within the connectome, a relativistic formalism using a brain-specific limiting velocity is required, which is not trivial matter (Le Bihan, 2020).”

• cites Le Bihan, 2020 sourceSupported

This reinforces the central thesis that diffusion MRI of the brain occupies a distinctive middle ground between anatomy and physiology.

“This reinforces the central thesis that diffusion MRI of the brain occupies a distinctive middle ground between anatomy and physiology (Le Bihan, 2007).”

• cites Le Bihan, 2007 sourceSupported

During anesthesia or minimally conscious states, reduced thalamic activity could be interpreted within this framework as a "flattening" of functional geometry and a weakening of effective connectivity, despite residual local activity.

“During anesthesia or minimally conscious states, reduced thalamic activity could be interpreted within this framework as a “flattening” of functional geometry and a weakening of effective connectivity, despite residual local activity, as revealed through DTI and resting-state fMRI (Fig. 7) (Demertzi et al., 2019).”

• cites Demertzi et al., 2019 sourcePartly supported

  The Results and Discussion (page not available in our copy) report that anesthesia favors a low-coordination pattern and that patients in minimally conscious states show greater dependence on anatomical connectivity despite persistent spontaneous activity. They do not establish reduced thalamic activity, a flattening of functional geometry, or weakened effective connectivity. The Methods describe diffusion spectrum imaging (DSI), not DTI, and specify that the anatomical-connectivity regions exclude subcortical regions. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

Mechanical stiffness is recognized by neurosurgeons as relevant to approach selection, resection strategy, and complication risk.

“Mechanical stiffness is recognized by neurosurgeons as relevant to approach selection, resection strategy, and complication risk (Lagerstrand et al., 2021).”

• cites Lagerstrand et al., 2021 sourceSupported

IVIM-based vMRE further expands this concept by generating virtually synthesized rather than physically induced shear-wave phenomena within tissues using the principles of IntraVoxel Incoherent Motion (IVIM) established earlier for blood microcirculation to model the intravoxel phase-shift distribution that would be induced by virtual tissue motion.

“IVIM-based vMRE further expands this concept by generating virtually synthesized rather than physically induced shear-wave phenomena within tissues using the principles of IntraVoxel Incoherent Motion (IVIM) established earlier for blood microcirculation (Le Bihan, 2019; Le Bihan et al., 2017) to model the intravoxel phase-shift distribution that would be induced by virtual tissue motion.”

• cites Le Bihan, 2019 sourceSupported
• cites Le Bihan et al., 2017 sourceNot checked — no full text of the source

In humans, visual stimulation increased stiffness in visual cortex by about 6-11%.

“In humans, visual stimulation increased stiffness in visual cortex by about 6–11% (Lan et al., 2020)”

• cites Lan et al., 2020 sourceSupported

Free-water correction, multi-shell acquisition, and joint structural-functional modeling are already making diffusion MRI less reliant on a few oversimplified scalar summaries.

“Free-water correction, multi-shell acquisition, and joint structural–functional modeling are already making diffusion MRI less reliant on a few oversimplified scalar summaries (Assaf et al., 2019; Fieremans et al., 2016).”

• cites Assaf et al., 2019 sourcePartly supported

  The 'Local measures of white matter microstructure' section says that common DTI indices such as mean diffusivity and FA are nonspecific, and describes compartment models that yield more detailed tissue parameters. The 'Part 2/ Functional models' section relates structural connectivity to functional organization. Our copy does not establish the specific combination of free-water correction, multi-shell acquisition, and joint structural–functional modeling claimed here (searched those terms and close variants). (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Fieremans et al., 2016 sourcePartly supported

  The Methods and Results describe varying diffusion time and fitting biophysical models to white-matter measurements, supporting the broader move beyond a few scalar summaries. But the Methods > In vivo measurements reports b = 0 and b = 500 s/mm² acquisitions, not multi-shell acquisition; the paper does not describe free-water correction or joint structural–functional modeling. Page numbers are unavailable in our copy. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

Visual stimulation-induced stiffness increase mirrors activation-related ADC and sADC decreases observed with DfMRI.

“mirroring activation-related ADC and sADC decreases observed with DfMRI (Le Bihan et al., 2006).”

• cites Le Bihan et al., 2006 sourcePartly supported

  The Results report a decrease in water diffusion during visual activation, supporting the general DfMRI comparison. But they attribute the signal change to expansion of the slow-diffusion pool and state that changes in the pools’ diffusion coefficients could not explain it; our copy does not establish the claimed sADC decrease. The stiffness comparison is the citing authors’ framing, not a finding attributed to this source. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

In a standard T2-weighted spin-echo sequence, imaging gradients contribute a nonzero baseline b-value typically around 10-50 s/mm2.

“In a standard T2-weighted spin-echo sequence, imaging gradients contribute a nonzero baseline b-value (which, hence, cannot be 0, as often reported)., typically around 10–50 s/mm².”

Early diffusion MRI used weak gradients and therefore required long diffusion times (>70 ms) for sufficient encoding.

“Early diffusion MRI used weak gradients and, therefore, required long diffusion times (>70 ms) for sufficient encoding.”

Acute stroke gave diffusion MRI its first compelling clinical indication by enabling objective diagnosis at the earliest stage through detecting reduced water mobility in hypoxic tissue associated with cytotoxic edema.

“Acute stroke, however, gave diffusion MRI its first compelling clinical indication: it enabled objective diagnosis at the earliest stage by detecting reduced water mobility in hypoxic tissue associated with cytotoxic edema, thereby opening the way to clinical trials for emergency thrombolytic therapies (Chien et al., 1992; Moseley, Cohen, Mintorovitch, et al., 1990; Warach et al., 1992).”

• cites Chien et al., 1992Not checked — no full text of the source
• cites Moseley, Cohen, Mintorovitch, et al., 1990 sourceNot checked — no full text of the source
• cites Warach et al., 1992 sourceNot checked — no full text of the source

Faster acquisitions together with stronger gradients supporting EPI made it possible to acquire multiple diffusion-weighted images up to b = 1,000 s/mm2 in approximately 1 minute with substantially reduced motion sensitivity.

“Faster acquisitions, together with stronger gradients supporting EPI, made it possible to acquire multiple diffusion-weighted images up to b = 1,000 s/mm² in approximately 1 minute with substantially reduced motion sensitivity.”

Echo-planar imaging reduced vulnerability to patient motion and physiological pulsatility by collapsing acquisition into very short readouts.

“Echo-planar imaging (EPI) reduced vulnerability to patient motion and physiological pulsatility by collapsing acquisition into very short readouts.”

Diffusion-weighted MRI enabled early detection of acute ischemia by revealing tissue injury before it became apparent on conventional structural imaging.

“First, diffusion-weighted MRI enabled the early detection of acute ischemia by revealing tissue injury before it became apparent on conventional structural imaging, thereby supporting timely treatment and improving outcomes for millions of patients worldwide.”

• Also stated in: § Introduction: Why the Brain Became the Natural Organ for Diffusion MRI · plain

  “Acute stroke, however, gave diffusion MRI its first compelling clinical indication: it enabled objective diagnosis at the earliest stage by detecting reduced water mobility in hypoxic tissue associated with cytotoxic edema, thereby opening the way to clinical trials for emergency thrombolytic therapies (Chien et al., 1992; Moseley, Cohen, Mintorovitch, et al., 1990; Warach et al., 1992).”

• cites Chien et al., 1992Not checked — no full text of the source
• cites Moseley, Cohen, Mintorovitch, et al., 1990 sourceNot checked — no full text of the source
• cites Warach et al., 1992 sourceNot checked — no full text of the source

With a head gradient coil capable of 40 mT/m, it became possible to acquire a set of diffusion-weighted images with 16 b-values up to 1,000 s/mm2 in just 1 minute and without motion artifacts.

“Suddenly, with a head gradient coil capable of 40 mT/m, it became possible to acquire a set of diffusion-weighted images 16 b-values up to 1,000 s/mm² in just 1 minute and without motion artifacts.”

The ADC measured at one diffusion time may not match the value measured at another, even when the same b-values are used.

“Another fundamental point is time dependence. The ADC measured at one diffusion time may not match the value measured at another, even when the same b-values are used.”

When diffusion-encoding gradients of 30 mT/m are added to reach a b-value near 1,000 s/mm2, imaging gradients and cross-terms may account for approximately 15% of the total.

“When diffusion-encoding gradients of, for example, 30 mT/m are added to reach a b-value near 1,000 s/mm², imaging gradients and cross-terms may account for approximately 15% of the total.”

No routine clinical sequence separates these contributions explicitly.

“No routine clinical sequence separates these contributions explicitly.”

In coherent fiber bundles, diffusion is greater parallel than perpendicular to the bundle.

“The central observation was simple: in coherent fiber bundles, diffusion is greater parallel than perpendicular to the bundle.”

Color-coded orientation maps of white-matter tracts were produced from diffusion images acquired along different directions, assuming that fibers align with the direction of greatest diffusion.

“Soon after anisotropy was recognized, color-coded orientation maps of white-matter tracts were produced from diffusion images acquired along different directions, assuming that fibers align with the direction of greatest diffusion (Douek et al., 1991).”

• cites Douek et al., 1991 sourceNot checked — no full text of the source

• AssumptionThe authors state this as an assumption: Fibers align with the direction of greatest diffusion.

In standard clinical sequences, the effective diffusion time is only approximate because imaging and diffusion-encoding gradients are mixed.

“Furthermore, in standard clinical sequences, the effective diffusion time is only approximate because imaging and diffusion-encoding gradients are mixed.”

Diffusion MRI became the imaging signature of early infarction.

“Although the exact mechanisms remained debated, the clinical impact was immediate: diffusion MRI became the imaging signature of early infarction (Chien et al., 1992; Warach et al., 1992), demonstrating the translational value of preclinical MRI.”

• cites Chien et al., 1992Not checked — no full text of the source
• cites Warach et al., 1992 sourceNot checked — no full text of the source

Diffusion MRI makes MRI sensitive to micrometer-scale water displacements over diffusion times of a few tens of milliseconds.

“Diffusion MRI transformed brain imaging by making MRI sensitive not only to anatomy and spin-relaxation contrast, but also to micrometer-scale water displacements over diffusion times of a few tens of milliseconds.”

Imaging-gradient contributions vary across shells in multi-shell acquisitions, causing erroneous diffusion weighting if they are not properly accounted for.

“The issue becomes particularly important in multi-shell acquisitions because these contributions vary across shells, causing erroneous diffusion weighting if they are not properly accounted for.”

Without fast acquisition, the hyperacute stroke applications that defined the field would not have become clinically practical when they did.

“Without fast acquisition, the brain may still have been a promising target, but the hyperacute stroke applications that defined the field would not have become clinically practical when they did.”

Diffusion MRI is not a single biomarker but an acquisition-and-model framework.

“From a signal-theory perspective, diffusion MRI is not a single biomarker but an acquisition-and-model framework.”

The birth of brain diffusion MRI was not a single event but a convergence of physical theory, NMR pulse-sequence design, MRI hardware, and neurological need.

“The birth of brain diffusion MRI was, therefore, not a single event but a convergence of physical theory, NMR pulse-sequence design, MRI hardware, and neurological need.”

Two very different tissues can produce similar ADC values if their underlying compartment mixtures differ in one way and converge in another.

“Two very different tissues can produce similar ADC values if their underlying compartment mixtures differ in one way and converge in another.”

The DTI framework ultimately led to reconstruction of large-scale structural pathways and connectomes in vivo, relevant to psychiatric disorders.

“Second, the strong orientational order of white matter produces diffusion anisotropy, which can be explored through a diffusion tensor imaging (DTI) framework, ultimately leading to the reconstruction of large-scale structural pathways and connectomes in vivo, relevant to psychiatric disorders.”

The diffusion MRI signal is sensitive to tissue organization because cell membranes, myelin, axonal packing, orientation dispersion, exchange, and perfusion-related incoherent fluid motion modulate the displacement distribution sampled by diffusing water molecules.

“The resulting signal is exquisitely sensitive to tissue organization, because cell membranes, myelin, axonal packing, orientation dispersion, exchange, as well as perfusion-related incoherent fluid motion all modulate the displacement distribution sampled by diffusing water molecules.”

The IVIM framework generalized diffusion MRI by recognizing that intravoxel microscopic motion includes both molecular diffusion and microvascular flow components.

“Around the same time, the intravoxel incoherent motion (IVIM) framework generalized diffusion MRI by recognizing that intravoxel microscopic motion includes both molecular diffusion and microvascular flow components (Le Bihan et al., 1986, 1988).”

• cites Le Bihan et al., 1986, 1988Not matched

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

The ADC is not an intrinsic or effective diffusion coefficient, but an acquisition-dependent parameter summarizing diffusion-related signal attenuation under specific experimental conditions.

“Importantly, the ADC, most often estimated from diffusion MRI signals acquired at two b-values (e.g., b as close as possible to 0, however, see above, and b = 1,000 s/mm²) and inverting Eq. (4), is not an intrinsic or effective diffusion coefficient, but an acquisition-dependent parameter summarizing diffusion-related signal attenuation under specific experimental conditions.”

The same reduction in apparent diffusivity may arise from cytotoxic edema, increased cellular packing, altered membrane permeability, orientation dispersion, partial volume with CSF, or vascular effects.

“The same reduction in apparent diffusivity may arise from cytotoxic edema, increased cellular packing, altered membrane permeability, orientation dispersion, partial volume with CSF, or vascular effects.”

Using Eq. (2) in diffusion MRI can substantially underestimate the true b-value and hence overestimate diffusion, because the measured signal is more attenuated than predicted by the idealized expression.

“Using Eq. (2) in diffusion MRI can substantially underestimate the true b-value and hence overestimate diffusion, because the measured signal is more attenuated than predicted by the idealized expression.”

The simple scaling argument that water displacement is comparable with cellular dimensions is the conceptual bridge that makes diffusion MRI a sub-voxel probe of microstructure.

“This simple scaling argument is the conceptual bridge that makes diffusion MRI a sub-voxel probe of microstructure (Le Bihan, 1995).”

• cites Le Bihan, 1995 sourceNot checked — no full text of the source

The measured diffusion MRI attenuation depends on the MRI gradient waveform, resulting diffusion sensitivity, diffusion time, and the extent to which the chosen biophysical model matches tissue reality.

“The measured attenuation depends on the MRI gradient waveform, resulting diffusion sensitivity, diffusion time, and the extent to which the chosen biophysical model matches tissue reality.”

The strong orientational order of white matter produces diffusion anisotropy.

“Second, the strong orientational order of white matter produces diffusion anisotropy, which can be explored through a diffusion tensor imaging (DTI) framework, ultimately leading to the reconstruction of large-scale structural pathways and connectomes in vivo, relevant to psychiatric disorders.”

The hyperintense lesion on diffusion-weighted images and the corresponding ADC reduction reflect altered water mobility in metabolically challenged tissue, not simply "more water" or "less water".

“The hyperintense lesion on diffusion-weighted images and the corresponding ADC reduction do not indicate simply “more water” or “less water”; they reflect altered water mobility in metabolically challenged tissue (Fig. 2).”

Even where more sophisticated models are available, DTI often remains the first analytic layer because it provides a stable orientation field and familiar summary indices.

“Even where more sophisticated models are available, DTI often remains the first analytic layer because it provides a stable orientation field and a familiar set of summary indices (Le Bihan et al., 2001).”

• cites Le Bihan et al., 2001 sourceNot checked — no full text of the source

Diffusion-tensor MR imaging of the human brain can obtain tensor-derived metrics and orientations in vivo.

“Pierpaoli and colleagues subsequently demonstrated diffusion-tensor MR imaging of the human brain, confirming that tensor-derived metrics and orientations could be obtained in vivo (Pierpaoli et al., 1996).”

• cites Pierpaoli et al., 1996 sourceNot checked — no full text of the source

Diffusion in the brain is highly anisotropic, especially in white matter.

“Diffusion is highly anisotropic in the brain, especially in white matter.”

Before diffusion MRI, whole-brain white-matter anatomy in living humans was inferred indirectly or reconstructed from postmortem dissection.

“Before diffusion MRI, whole-brain white-matter anatomy in living humans was inferred indirectly or reconstructed from postmortem dissection.”

In regions where multiple fiber populations coexist, tensor-derived metrics may be ambiguous or misleading.

“In such regions, tensor-derived metrics may be ambiguous or misleading.”

Acute stroke remains the clearest example of why diffusion MRI should be understood as a dynamic microstructure imaging method rather than merely a specialized contrast weighting.

“Even decades later, acute stroke remains the clearest example of why diffusion MRI should be understood as a dynamic microstructure imaging method rather than merely a specialized contrast weighting.”

DTI offered the first noninvasive whole-brain maps of major white-matter architecture in living humans.

“For neuroscientists and clinicians alike, DTI offered the first noninvasive whole-brain maps of major white-matter architecture in living humans, turning diffusion MRI from a lesion-sensitive sequence into a platform for investigating normal anatomy, development, aging, and white-matter diseases (see below).”

Under appropriate conditions, blood flow through quasi-randomly oriented capillary segments can attenuate the diffusion-weighted MRI signal within the IVIM framework.

“Under appropriate conditions, blood flow through quasi-randomly oriented capillary segments can attenuate the diffusion-weighted MRI signal within the IVIM framework.”

Tractography extended the logic of DTI from local orientation estimation to long-range pathway reconstruction.

“Tractography extended the logic of DTI from local orientation estimation to long-range pathway reconstruction.”

White-matter anisotropy established the neuroscientific reach of brain diffusion MRI.

“If stroke established the clinical value of brain diffusion MRI, white-matter anisotropy established its neuroscientific reach.”

No single microstructural mechanism explains all observations of ADC drop in ischemia across time and tissue compartments.

“Yet no single microstructural mechanism explains all observations across time and tissue compartments.”

DTI became the common language of brain diffusion MRI.

“Despite those limitations, DTI became the common language of brain diffusion MRI.”

The idea that macroscopic structural connectivity could be charted in vivo led directly to tractography and later to the modern discourse of the brain connectome.

“That idea led directly to tractography, and, later, to the modern discourse of the brain connectome.”

The first application that irreversibly changed the status of diffusion MRI in medical imaging was acute cerebral ischemia.

“The first application that irreversibly changed the status of diffusion MRI in medical imaging was acute cerebral ischemia.”

The introduction of the DTI framework converted the qualitative observation of anisotropy into a set of quantitative descriptors.

“It converted the qualitative observation of anisotropy into a set of quantitative descriptors.”

Streamlines are not axons; they are algorithmic reconstructions based on local orientation fields, propagation rules, and stopping criteria.

“Streamlines are not axons; they are algorithmic reconstructions based on local orientation fields, propagation rules, and stopping criteria.”

Such anisotropy reflects the ordered architecture of axons and their surrounding microenvironment, even though the MRI voxel is orders of magnitude larger than a single fiber.

“Such anisotropy reflects the ordered architecture of axons and their surrounding microenvironment, even though the MRI voxel is orders of magnitude larger than a single fiber (Beaulieu, 2002).”

• cites Beaulieu, 2002 sourceNot checked — no full text of the source

Another contributor to ADC is blood microcirculation, particularly at very low b-values.

“Another contributor to ADC is blood microcirculation, particularly at very low b-values.”

Axial diffusivity λ1 and radial diffusivity (λ2 + λ3)/2 are composite observables in vivo rather than unique markers of axonal integrity or myelin content.

“Similarly, axial diffusivity λ1 and radial diffusivity (λ2 + λ3)/2 are often discussed mechanistically, but in vivo they remain composite observables rather than unique markers of axonal integrity or myelin content.”

Diffusion MRI provided strong empirical evidence that it was sensitive to microscopic pathophysiology, probing tissue injury at a phase when the distinction between salvageable and irreversibly injured tissue mattered most.

“Second, it provided strong empirical evidence that diffusion MRI was sensitive to microscopic pathophysiology, probing tissue injury at a phase when the distinction between salvageable and irreversibly injured tissue mattered most.”

Blood water moving through capillary networks arranged in a quasi-random spatial pattern can be described as a random walk, pseudo-diffusion process, making Einstein's diffusion model conceptually applicable.

“Although it operates on a very different physical scale, blood water moving through capillary networks arranged in a quasi-random spatial pattern can be described as a random walk, pseudo-diffusion process, as blood changes direction repeatedly from one capillary segment to the next, making Einstein’s diffusion model conceptually applicable.”

Tractography is inferential, not photographic.

“However, tractography is inferential, not photographic.”

Biological tissues are dynamic, hierarchically organized systems whose molecular composition and cellular architecture shape water mobility.

“Biological tissues are dynamic, hierarchically organized systems whose molecular composition and cellular architecture shape not only mechanical and electrical properties (see below), but also water mobility.”

A major limitation of the Standard Model is that parameter estimation is often ill-conditioned: different combinations of intra-axonal fraction, extra-axonal diffusivities, orientation dispersion, free water, exchange, and noise can produce very similar signals.

“A major limitation is that parameter estimation is often ill-conditioned: different combinations of intra-axonal fraction, extra-axonal diffusivities, orientation dispersion, free water, exchange, and noise can produce very similar signals.”

Multiple ODF peaks may reveal distinct fiber populations in regions of crossing, kissing, or fanning fibers.

“Multiple ODF peaks may, therefore, reveal distinct fiber populations in regions of crossing, kissing, or fanning fibers.”

Directionally organized structures, such as aligned axons or tissue layers, produce anisotropic diffusion.

“Directionally organized structures, such as aligned axons or tissue layers, produce anisotropic diffusion (Beaulieu, 2002).”

• cites Beaulieu, 2002 sourceNot checked — no full text of the source

Diffusion MRI constrains connectivity rather than measuring it exhaustively.

“The field increasingly recognizes that diffusion MRI constrains connectivity rather than measuring it exhaustively.”

Deviations from the Gaussian assumption are not merely fitting residuals; they encode the interaction of water with complex tissue geometry, exchange, and multiscale organization.

“These deviations are not merely fitting residuals; they encode the interaction of water with complex tissue geometry, exchange, and multiscale organization.”

The shift from localized lesions to distributed circuits owes much to the way diffusion MRI made brain wiring visible, albeit imperfectly.

“The shift from localized lesions to distributed circuits owes much to the way diffusion MRI made brain wiring visible, albeit imperfectly.”

Membranes constrain water motion by separating intracellular and extracellular spaces.

“Membranes constrain water motion by separating intracellular and extracellular spaces.”

Diffusion MRI is most informative when acquisitions are diversified across b-values, gradient directions, diffusion times, and encoding waveforms, rather than reduced to a single shell or single summary metric.

“Diffusion MRI is, therefore, most informative when acquisitions are diversified across b-values, gradient directions, diffusion times, and encoding waveforms, rather than reduced to a single shell or single summary metric.”

Time dependence provides an additional dimension of contrast, because barriers, disorder, and exchange are probed differently as diffusion length changes.

“Time dependence provides an additional dimension of contrast, because barriers, disorder, and exchange are probed differently as diffusion length changes.”

DKI has proved sensitive to pathological changes in ischemia, neurodegeneration, and tumors.

“DKI has proved sensitive to pathological changes in ischemia, neurodegeneration, and tumors (Helpern & Jensen, 2015; Iima & Le Bihan, 2016).”

• cites Helpern & Jensen, 2015Not matched

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

• cites Iima & Le Bihan, 2016 sourceNot checked — no full text of the source

Small changes in acquisition, reconstruction, or filtering can alter connectivity matrices and downstream inferences.

“Small changes in acquisition, reconstruction, or filtering can alter connectivity matrices and downstream inferences.”

The identifiability problem is central to biophysical diffusion modeling.

“This identifiability problem is central.”

Signal decay in both white and gray matter often deviates substantially from the Gaussian assumption.

“Signal decay in both white and gray matter often deviates substantially from the Gaussian assumption.”

As gradient hardware improved and acquisitions extended to higher b-values and multiple diffusion times, it became clear that a mono-exponential model, such as Eq. (3) and the DTI framework, captures only a subset of the information present in brain diffusion data.

“As gradient hardware improved and acquisitions extended to higher b-values and multiple diffusion times, it became increasingly clear that a mono-exponential model, such as Eq. (3) and the DTI framework, captures only a subset of the information present in brain diffusion data.”

The DKI model breaks down at very high b-values (e.g., around 3,000 s/mm2 for typical ADC0 and K values found in tissues), requiring higher-order terms.

“Importantly, because this model is a second-order expansion on b, it breaks down at very high b-values (e.g., around 3,000 s/mm² for typical ADCo and K values found in tissues), requiring higher-order terms.”

Graph metrics are only as meaningful as the tractography from which they are derived.

“At the same time, the field learned that graph metrics are only as meaningful as the tractography from which they are derived.”

In biological tissues, increased kurtosis is generally interpreted as reflecting greater microstructural complexity, restriction, or hindrance.

“In biological tissues, increased kurtosis is generally interpreted as reflecting greater microstructural complexity, restriction, or hindrance.”

Tractography works best as a mesoscale anatomical inference tool.

“Even so, the conceptual caution remains the same: tractography works best as a mesoscale anatomical inference tool.”

Diffusion MRI can detect tissue features at the scale where cellular organization and pathology emerge, without directly imaging them.

“It can detect tissue features at the scale where cellular organization (and pathology) emerge, without directly imaging them.”

Pathological processes-including edema, inflammation, tumor infiltration, fibrosis, necrosis, and treatment effects-modify tissue microarchitecture and therefore diffusion signals.

“Pathological processes—including edema, inflammation, tumor infiltration, fibrosis, necrosis, and treatment effects—modify tissue microarchitecture and, therefore, diffusion signals.”

Kurtosis parameters remain descriptive rather than explanatory because they quantify non-Gaussianity without identifying its specific biological origin.

“however, kurtosis parameters remain descriptive rather than explanatory because they quantify non-Gaussianity without identifying its specific biological origin.”

Diffusion MRI is indispensable for a modern account of how large-scale brain systems are physically scaffolded, provided that tractography is interpreted as an indirect, model-based estimate of anatomical connectivity.

“Diffusion MRI is thus indispensable for a modern account of how large-scale brain systems are physically scaffolded, provided that tractography is interpreted as an indirect, model-based estimate of anatomical connectivity.”

No single diffusion marker can reliably distinguish viable tumor from pseudo-progression or radiation necrosis.

“no single diffusion marker can reliably distinguish viable tumor from pseudo-progression or radiation necrosis (Lawrence et al., 2025).”

• cites Lawrence et al., 2025 sourceNot checked — no full text of the source

Diffusion-weighted imaging and ADC maps contribute to the assessment of tumor cellularity, grade, necrosis, and peritumoral abnormalities.

“Diffusion-weighted imaging and ADC maps contribute to the assessment of tumor cellularity, grade, necrosis, and peritumoral abnormalities.”

Gray matter presents a different and arguably greater challenge for diffusion MRI modeling than white matter.

“Gray matter presents a different and arguably greater challenge.”

IVIM-based fMRI could not compete once alternative methods emerged in the early 1990s.

“IVIM-based fMRI could not compete once alternative methods emerged in the early 1990s.”

These measures are promising candidate biomarkers that complement structural MRI, molecular biomarkers, and clinical assessment.

“These measures are promising candidate biomarkers that complement structural MRI, molecular biomarkers, and clinical assessment.”

Diffusion metrics should not be interpreted as direct measures of myelin or axonal density.

“These metrics should not be interpreted as direct measures of myelin or axonal density, but they have enabled in vivo mapping of tract-specific maturation and its links with emerging motor, language, and cognitive functions (Dubois et al., 2021; Lebel et al., 2019).”

DfMRI remained controversial because in vivo diffusion-weighted signals can still contain vascular contributions.

“At the same time, DfMRI remained controversial because in vivo diffusion-weighted signals can still contain vascular contributions (Miller et al., 2007).”

• cites Miller et al., 2007 sourceNot checked — no full text of the source

Once confounds are carefully controlled, an intrinsic diffusion response function should be used in place of the canonical hemodynamic response function used in BOLD fMRI.

“Once these confounds are carefully controlled, an intrinsic diffusion response function should be used in place of the canonical hemodynamic response function used in BOLD fMRI (Aso et al., 2009).”

• cites Aso et al., 2009 sourceNot checked — no full text of the source

IVIM MRI has been combined with BOLD fMRI in a complementary "negative" approach that improves spatial specificity by suppressing signals from large feeding arteries and draining veins.

“IVIM MRI has also been combined with BOLD fMRI in a complementary “negative” approach that improves spatial specificity by suppressing signals from large feeding arteries and draining veins (Song et al., 2018).”

• cites Song et al., 2018 sourceNot checked — no full text of the source

DfMRI was proposed to reflect neuromechanical coupling: rapid changes in water diffusion within neural tissue caused by biophysical events accompanying neuronal activation.

“Unlike BOLD and IVIM fMRI, which depend primarily on neurovascular coupling, DfMRI was proposed to reflect neuromechanical coupling: rapid changes in water diffusion within neural tissue caused by biophysical events accompanying neuronal activation.”

The neuromechanical-coupling concept rests on a broad body of work showing that physiological activation involves not only electrical and biochemical events but also rapid physical changes in cells and their surroundings.

“The neuromechanical-coupling concept rests on a broad body of work showing that physiological activation involves not only electrical and biochemical events but also rapid physical changes in cells and their surroundings.”

They also motivate a two-stage model of activation in which rapid tissue-microstructural events precede the slower hemodynamic phase measured by conventional fMRI.

“They also motivate a two-stage model of activation in which rapid tissue-microstructural events precede the slower hemodynamic phase measured by conventional fMRI.”

BOLD fMRI remains an indirect measure of neural activity through neurovascular coupling.

“While BOLD fMRI transformed cognitive neuroscience, it remains an indirect measure of neural activity through neurovascular coupling (Roy & Sherrington, 1890).”

• cites Roy & Sherrington, 1890Not matched

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

DfMRI offers a different view of brain function.

“Beyond these potential practical advantages, DfMRI offers a different view of brain function.”

Diffusion-based virtual MRE offers a potential alternative by inferring tissue stiffness from diffusion MRI without requiring physical shear-wave propagation.

“Diffusion-based virtual MRE (vMRE) offers a potential alternative by inferring tissue stiffness from diffusion MRI without requiring physical shear-wave propagation (Le Bihan et al., 2017).”

• cites Le Bihan et al., 2017 sourceNot checked — no full text of the source

In conventional MRE, amplitude and frequency of real shear waves are dictated by hardware and human tolerances.

“In conventional MRE, amplitude and frequency of real shear waves are dictated by hardware and human tolerances.”

For brain clinical applications, diffusion-vMRE may be technically simpler and more deployable than conventional MRE.

“For brain clinical applications, diffusion-vMRE may be technically simpler and more deployable than conventional MRE.”

• AssumptionRests on an assumption: vMRE does not require external mechanical vibrations, which may be difficult to use in selected settings, particularly patients with traumatic brain injury, elderly population or children.

Brain diffusion MRI had already taught investigators that water mobility is sensitive to dynamic cellular states.

“In other words, brain diffusion MRI had already taught investigators that water mobility is sensitive to dynamic cellular states.”

Because tumor consistency directly affects whether a tumor is suctionable or fibrous, vMRE information may improve surgical planning.

“Because consistency directly affects whether a tumor is suctionable or fibrous, this information may improve surgical planning.”

• AssumptionRests on an assumption: Tumor consistency directly affects whether a tumor is suctionable or fibrous.

Meningioma surgery is another potential indication for vMRE because tumor firmness is difficult to predict on conventional MRI.

“Meningioma surgery is another potential indication because tumor firmness is difficult to predict on conventional MRI.”

• AssumptionRests on an assumption: Tumor firmness is difficult to predict on conventional MRI.

DfMRI underscores that diffusion measurements are not limited to static properties of tissue architecture; they can, in principle, reflect dynamic changes in the same architecture.

“DfMRI underscores that diffusion measurements are not limited to static properties of tissue architecture. They can, in principle, reflect dynamic changes in the same architecture, even if those changes are subtle and transient.”

Its rationale is that microstructural features that hinder water diffusion, such as cellularity, axonal packing, myelination, extracellular-matrix organization, and tissue geometry, also contribute to mechanical stiffness.

“Its rationale is that microstructural features that hinder water diffusion, such as cellularity, axonal packing, myelination, extracellular-matrix organization, and tissue geometry, also contribute to mechanical stiffness.”

Noninvasively tracking neural activity along specific pathways remains challenging.

“Noninvasively tracking neural activity along specific pathways, therefore, remains challenging.”

This development enables an entirely new class of elastographic contrast, one not constrained by mechanical systems, vibration frequencies, driver placement, or tissue-wave attenuation physics.

“This development enables an entirely new class of elastographic contrast, one not constrained by mechanical systems, vibration frequencies, driver placement, or tissue-wave attenuation physics.”

This approach was first evaluated and validated for liver-fibrosis staging.

“This approach was first evaluated and validated for liver-fibrosis staging (Kromrey et al., 2020; Le Bihan et al., 2017).”

• cites Kromrey et al., 2020 sourceNot checked — no full text of the source
• cites Le Bihan et al., 2017 sourceNot checked — no full text of the source

Diffusion MRI can contribute to provide model-dependent estimates of axonal microstructural features.

“Diffusion MRI can contribute to provide model-dependent estimates of axonal microstructural features.”

Brain elasticity can be viewed as a mechanobiological marker of tissue microstructure, reflecting the integrated contribution of dendrites, axons, astrocytic processes, extracellular matrix, and interstitial-water balance.

“Brain elasticity can, therefore, be viewed as a mechanobiological marker of tissue microstructure, reflecting the integrated contribution of dendrites, axons, astrocytic processes, extracellular matrix, and interstitial-water balance.”

In IVIM-based virtual elastography, wave propagation is no longer a physical event-it becomes a simulated behavior, inferred from the relationship between tissue stiffness and the attenuation of the amplitude of motion-encoded MR signals induced by a virtual traveling wave.

“In IVIM-based virtual elastography, wave propagation is no longer a physical event—it becomes a simulated behavior, inferred from the relationship between tissue stiffness, as obtained through diffusion-vMRE, and the attenuation of the amplitude of motion-encoded MR signals induced by a virtual traveling wave (Fig. 8).”

Metrics such as ADC, FA, mean diffusivity, kurtosis, or model-derived intracellular fractions are not self-interpreting.

“Metrics such as ADC, FA, mean diffusivity, kurtosis, or model-derived intracellular fractions are not self-interpreting.”

Diffusion MRI is powerful for investigating normal and diseased brain tissue because water motion is strongly constrained by cellular and mesoscale architecture.

“Diffusion MRI is powerful for investigating normal and diseased brain tissue because water motion is strongly constrained by cellular and mesoscale architecture.”

The same change in a scalar diffusion metric can arise from multiple microstructural causes.

“The same change in a scalar metric can arise from multiple microstructural causes.”

Cardiac pulsations may become less efficient to clean the brain through the glymphatic system when elasticity decreases with age.

“so that cardiac pulsations may become less efficient to “clean” the brain (Fig. 8) through the glymphatic system”

• AssumptionRests on an assumption: In the normal brain, cardiac frequency (about 1 Hz) is the nominal frequency that results in heart-induced mechanical waves propagating to about 20 cm/s, so that each heart beat propagates to the whole brain.

When elasticity decreases with age, the propagation speed of cardiac-induced mechanical waves goes down.

“When elasticity decreases with age, the propagation speed goes down”

• AssumptionRests on an assumption: In the normal brain, cardiac frequency (about 1 Hz) is the nominal frequency that results in heart-induced mechanical waves propagating to about 20 cm/s, so that each heart beat propagates to the whole brain.