Quantifying white matter tract diffusion parameters in the presence of increased extra-fiber cellularity and vasogenic edema

Inclusion of the isotropic component of free diffusion is insufficient to deal with coexisting CNS inflammation (increased cellularity and vasogenic edema) or tissue loss in diseased tissues.

“However, such modification is insufficient to deal with coexisting CNS inflammation (increased cellularity and vasogenic edema) or tissue loss that may be seen in diseased tissues (Horsfield and Jones, 2002).”

• cites Horsfield and Jones, 2002 sourcePartly supported

  The Reduced diffusion anisotropy and Ischaemic leukoaraiosis sections describe how axonal loss and replacement or proliferation of glial cells affect diffusion measurements. The review does not assess whether adding an isotropic free-diffusion component is sufficient to account for inflammation, vasogenic edema, or tissue loss; that central modeling claim is unsupported. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

DTI-derived FA is not a marker specific to myelin damage because other pathological components may also contribute to diffusion anisotropy change.

“Although decreased DTI-derived fractional anisotropic (FA) has been demonstrated to reflect myelin damage in multiple sclerosis (MS) (Schmierer et al., 2007), it is not a marker specific to myelin damage since other pathological components may also contribute to diffusion anisotropy change (Werring et al., 1999, Assaf et al., 2002).”

• cites Werring et al., 1999 sourceNot checked — no full text of the source
• cites Assaf et al., 2002 sourcePartly supported

  The “High b-Value q-Space DWI vs. Conventional DTI” section (page unavailable) says FA can be affected by an incoherent arrangement of normal fibers, supporting the point that FA is not specific to myelin damage. The paper discusses demyelination, axonal loss, and inflammation in MS, but the passages do not establish that those other pathological components specifically cause changes in FA. (LLM, conf=0.83)

  Experimental — accuracy not yet measured

DBSI resolved and quantified the confounding effects of increased cellularity and vasogenic edema on DTI-derived λ∥, λ⊥, and FA.

“These confounding effects were resolved and quantified by the recently developed DBSI analysis (Wang et al., 2011), as demonstrated herein using data generated by both Monte-Carlo simulation and tissue phantoms.”

• Also stated in: § Discussion · boosted

  “In summary, we have demonstrated that DBSI removes confounding effects of increased cellularity and vasogenic edema on DTI derived λǁ, λ⊥, and FA.”

• cites Wang et al., 2011 sourceSupported

Inflammation-associated vasogenic edema increases apparent diffusion coefficient (ADC).

“For example, inflammation associated vasogenic edema has been recognized to increase apparent diffusion coefficient (ADC) and underestimate the DTI-derived FA of fiber tracts (Pasternak et al., 2009, Naismith et al., 2010).”

• cites Pasternak et al., 2009 sourceNot checked — no full text of the source
• cites Naismith et al., 2010 sourcePartly supported

  The paper’s discussion (page unavailable in our copy) says diffusivity normalized in some acute MS lesions, perhaps as edema associated with an acute inflammatory infiltrate resolved. This supports a link between inflammation-associated edema and elevated diffusivity, but the paper does not specifically identify vasogenic edema or show that it increases ADC. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

Inflammation-associated vasogenic edema underestimates DTI-derived FA of fiber tracts.

“For example, inflammation associated vasogenic edema has been recognized to increase apparent diffusion coefficient (ADC) and underestimate the DTI-derived FA of fiber tracts (Pasternak et al., 2009, Naismith et al., 2010).”

• cites Pasternak et al., 2009 sourceNot checked — no full text of the source
• cites Naismith et al., 2010 sourcePartly supported

  The “DTI parameters became altered during Gd enhancement” section (page unavailable) reports that FA decreased by 35% and mean diffusivity increased by 21% in acute MS lesions. The discussion suggests that normalization of diffusivity may reflect resolution of edema associated with inflammation, but the paper does not establish that vasogenic edema caused the lower FA or higher diffusivity in fiber tracts generally. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

In vivo DBSI-derived restricted isotropic diffusion fraction map of optic nerve from EAE mice revealed a heterogeneous increase in cellularity, closely matching the pattern of DAPI-positive cell nuclei staining of the same nerve.

“Similar to previous results on corpus callosum from cuprizone treated mice (Wang et al., 2011), in vivo DBSI derived restricted isotropic diffusion fraction map of optic nerve from EAE mice revealed a heterogeneous increase in cellularity, closely matching the pattern of DAPI-positive cell nuclei staining of the same nerve (Fig. 6e, k, and l).”

• cites Wang et al., 2011 sourceSupported

• AssumptionRests on an assumption: Current MRI methods do not assess specific elements of CNS pathology.

DBSI allowed for an accurate estimation of white matter tract diffusion properties in the presence of increased cellularity and vasogenic edema.

“The current results from fixed normal tissues and gel phantoms, and our previous in vivo mouse brain data (Wang et al., 2011) suggest that the DBSI approach allowed for an accurate estimation of white matter tract diffusion properties in the presence of increased cellularity and vasogenic edema.”

• cites Wang et al., 2011 sourceSupported

In EAE vs sham optic nerves, DTI-derived λ⊥ was significantly increased (0.21 ± 0.04 vs. 0.11 ± 0.01 μm2/ms, p = 0.0046).

“DTI λ⊥ = 0.21 ± 0.04 vs. 0.11 ± 0.01 μm2/ms (p = 0.0046).”

25-direction DBSI-derived radial diffusivity was well-estimated at SNR ≥ 20.

“DBSI-radial diffusivity was well-estimated at SNR ≥ 20 (Fig. 5B);”

In EAE vs sham optic nerves, DTI-derived λ∥ was significantly decreased (1.34 ± 0.37 vs. 1.89 ± 0.09 μm2/ms, p = 0.028).

“DTI λ∥ = 1.34 ± 0.37 vs. 1.89 ± 0.09 μm2/ms (p = 0.028);”

DBSI-derived non-restricted isotropic diffusion fraction (putative vasogenic edema) was significantly higher in EAE optic nerves than in controls (0.15 ± 0.03 vs. 0.03 ± 0.01, p < 0.0001).

“The non-restricted isotropic diffusion fraction (i.e., the putative vasogenic edema) was also significantly higher in EAE optic nerves than that of the control (0.15 ± 0.03 vs. 0.03 ± 0.01, p < 0.0001).”

• AssumptionRests on an assumption: The Monte-Carlo simulation models using spheres and extra-axonal space accurately represent cellularity and vasogenic edema in tissue.

DBSI-estimated extent of vasogenic edema correlated with experimentally added gel and Monte-Carlo simulations.

“Importantly, the extent of extra-fiber cellularity and edema estimated by DBSI correlated with experimentally added gel and Monte-Carlo simulations.”

Increased cellularity, mimicked by graded increase of the restricted isotropic diffusion tensor component in Monte-Carlo simulations, significantly decreased DTI-derived FA.

“Increased cellularity, mimicked by graded increase of the restricted isotropic diffusion tensor component in Monte-Carlo simulations, significantly decreased FA and axial diffusivity with limited impact on radial diffusivity derived by DTI.”

Increased extent of vasogenic edema, induced by adding gel to fixed normal mouse trigeminal nerves or by increasing non-restricted isotropic diffusion tensor components in Monte-Carlo simulations, significantly increased DTI-derived radial diffusivity.

“Increased extent of vasogenic edema, by addition of various amount of gel to fixed normal mouse trigeminal nerves or by increasing non-restricted isotropic diffusion tensor components in Monte-Carlo simulations, significantly decreased fractional anisotropy (FA), increased radial diffusivity, while less significantly increased axial diffusivity derived by DTI.”

In vivo DBSI using a 25-direction diffusion scheme correctly reflects inflammation of optic nerves in EAE mice.

“Results support that in vivo DBSI using 25-direction diffusion scheme correctly reflect the underlying axonal injury, demyelination, and inflammation of optic nerves in EAE mice.”

With decreasing SNR, both bias and precision of 25-direction DBSI-derived parameters worsened.

“With decreasing SNR, both bias and precision worsened (Fig. 5).”

DBSI-estimated extent of extra-fiber cellularity correlated with experimentally added gel and Monte-Carlo simulations.

“Importantly, the extent of extra-fiber cellularity and edema estimated by DBSI correlated with experimentally added gel and Monte-Carlo simulations.”

MBP-positive staining area was decreased in EAE optic nerves compared with controls (37% ± 5% vs. 45% ± 3%; p = 0.01).

“and MBP-positive staining area (37% ± 5% vs. 45% ± 3%; p = 0.01) exhibiting statistically significant correlations (Figs. 7B and D).”

DBSI-derived non-restricted and restricted isotropic diffusion fractions using 25- and 99-direction diffusion schemes were not different from the true values for the simulation.

“DBSI derived non-restricted and restricted isotropic diffusion fractions using 25- and 99-direction diffusion schemes (Fig. 4A and B) were not different from the true values for the simulation.”

Increased cellularity, mimicked by graded increase of the restricted isotropic diffusion tensor component in Monte-Carlo simulations, had limited impact on DTI-derived radial diffusivity.

“Increased cellularity, mimicked by graded increase of the restricted isotropic diffusion tensor component in Monte-Carlo simulations, significantly decreased FA and axial diffusivity with limited impact on radial diffusivity derived by DTI.”

In vivo DBSI using a 25-direction diffusion scheme correctly reflects demyelination of optic nerves in EAE mice.

“Results support that in vivo DBSI using 25-direction diffusion scheme correctly reflect the underlying axonal injury, demyelination, and inflammation of optic nerves in EAE mice.”

The 25-direction diffusion scheme provided comparable DBSI estimation of fiber diffusion parameters as the 99-direction scheme in phantom experiments and simulations.

“Results from both phantom experiments and simulations suggested that the 25-direction diffusion scheme provided comparable DBSI estimation of both fiber diffusion parameters and extra-fiber cellularity/edema extent as those by 99-direction scheme.”

Increased cellularity, mimicked by graded increase of the restricted isotropic diffusion tensor component in Monte-Carlo simulations, significantly decreased DTI-derived axial diffusivity.

“Increased cellularity, mimicked by graded increase of the restricted isotropic diffusion tensor component in Monte-Carlo simulations, significantly decreased FA and axial diffusivity with limited impact on radial diffusivity derived by DTI.”

Increased cellularity did not affect DBSI-derived fiber FA.

“Results showed that increased cellularity or vasogenic edema did not affect the DBSI-derived fiber FA, axial or radial diffusivity.”

Quantitative immunohistochemical analyses revealed decreased SMI-31-positive axon counts in EAE optic nerves compared with controls (5.3 × 105 ± 1.8 × 105 vs. 7.4 × 105 ± 0.2 × 105 counts/mm2; p = 0.03).

“Quantitative immunohistochemical analyses revealed that DTI/DBSI detected axonal injury and demyelination was also seen in SMI-31-positive axon counts (5.3 × 105 ± 1.8 × 105 vs. 7.4 × 105 ± 0.2 × 105 counts/mm2; p = 0.03; Figs. 7 A and C)”

In EAE vs sham optic nerves, DBSI-derived λ∥ was significantly decreased (1.73 ± 0.22 vs. 2.10 ± 0.04 μm2/ms, p = 0.020).

“DBSI λ∥= 1.73 ± 0.22 vs. 2.10 ± 0.04 μm2/ms (p = 0.020);”

In vivo DBSI using a 25-direction diffusion scheme correctly reflects underlying axonal injury of optic nerves in EAE mice.

“Results support that in vivo DBSI using 25-direction diffusion scheme correctly reflect the underlying axonal injury, demyelination, and inflammation of optic nerves in EAE mice.”

Increased vasogenic edema did not affect DBSI-derived fiber FA.

“Results showed that increased cellularity or vasogenic edema did not affect the DBSI-derived fiber FA, axial or radial diffusivity.”

Increased vasogenic edema did not affect DBSI-derived fiber radial diffusivity.

“Results showed that increased cellularity or vasogenic edema did not affect the DBSI-derived fiber FA, axial or radial diffusivity.”

All diffusion MRI-detected pathologies in the EAE optic nerve exhibited a heterogeneous distribution in the cross-section of the nerve.

“All diffusion MRI detected pathologies in the EAE optic nerve exhibited a heterogeneous distribution in the cross-section of the nerve (Figs. 6a – 6f).”

Increased cellularity did not affect DBSI-derived fiber axial diffusivity.

“Results showed that increased cellularity or vasogenic edema did not affect the DBSI-derived fiber FA, axial or radial diffusivity.”

Increased extent of vasogenic edema, induced by adding gel to fixed normal mouse trigeminal nerves or by increasing non-restricted isotropic diffusion tensor components in Monte-Carlo simulations, significantly decreased DTI-derived fractional anisotropy (FA).

“Increased extent of vasogenic edema, by addition of various amount of gel to fixed normal mouse trigeminal nerves or by increasing non-restricted isotropic diffusion tensor components in Monte-Carlo simulations, significantly decreased fractional anisotropy (FA), increased radial diffusivity, while less significantly increased axial diffusivity derived by DTI.”

The 25-direction diffusion scheme provided comparable DBSI estimation of extra-fiber cellularity/edema extent as the 99-direction scheme in phantom experiments and simulations.

“Results from both phantom experiments and simulations suggested that the 25-direction diffusion scheme provided comparable DBSI estimation of both fiber diffusion parameters and extra-fiber cellularity/edema extent as those by 99-direction scheme.”

DBSI-derived restricted isotropic diffusion fraction (putative cellularity) in EAE optic nerves was significantly higher than in controls (0.10 ± 0.09 vs. 0.03 ± 0.01, p = 0.022).

“DBSI derived restricted isotropic diffusion fraction (i.e., the putative cellularity) in EAE optic nerves was significantly higher than that of the control (0.10 ± 0.09 vs. 0.03 ± 0.01, p = 0.022).”

Increased extent of vasogenic edema, induced by adding gel to fixed normal mouse trigeminal nerves or by increasing non-restricted isotropic diffusion tensor components in Monte-Carlo simulations, less significantly increased DTI-derived axial diffusivity.

“Increased extent of vasogenic edema, by addition of various amount of gel to fixed normal mouse trigeminal nerves or by increasing non-restricted isotropic diffusion tensor components in Monte-Carlo simulations, significantly decreased fractional anisotropy (FA), increased radial diffusivity, while less significantly increased axial diffusivity derived by DTI.”

Increased vasogenic edema did not affect DBSI-derived fiber axial diffusivity.

“Results showed that increased cellularity or vasogenic edema did not affect the DBSI-derived fiber FA, axial or radial diffusivity.”

In Monte-Carlo simulation data, there was no discernible difference between 25-direction and 99-direction diffusion encoding schemes in DBSI-derived axial diffusivity, radial diffusivity, and FA.

“There was no discernible difference between 25- vs. 99-direction diffusion encoding scheme in DBSI derived diffusion parameters on data generated using Monte-Carlo simulation (25-dir vs. 99-dir): axial diffusivity (1.80 ± 0.03 μm2/ms vs. 1.81 ± 0.05 μm2/ms), radial diffusivity (0.05 ± 0.01 μm2/ms vs. 0.04 ± 0.01 μm2/ms) and FA (0.96 ± 0.01 vs. 0.97 ± 0.01 μm2/ms).”

Noise affected the bias and precision of 25-direction DBSI-derived axial diffusivity, radial diffusivity, restricted isotropic diffusion tensor fraction, and non-restricted isotropic diffusion tensor fraction.

“The noise affected the bias and precision of the 25-direction DBSI-derived axial diffusivity (Fig. 5A), radial diffusivity (Fig. 5B), restricted isotropic diffusion tensor fraction (Fig. 5C), and non-restricted isotropic diffusion tensor fraction (Fig. 5D).”

In EAE vs sham optic nerves, DBSI-derived λ⊥ was significantly increased (0.15 ± 0.03 vs. 0.11 ± 0.01 μm2/ms, p = 0.047).

“DBSI λ⊥ = 0.15 ± 0.03 vs. 0.11 ± 0.01 μm2/ms (p = 0.047);”

With the addition of 2% agar gel to fixed trigeminal nerve phantoms, DTI-derived λ⊥ or FA was significantly affected while DBSI-derived parameters remained unchanged.

“With the addition of 2% agar gel, DTI derived λ⊥ or FA was significantly affected while DBSI derived parameters remained unchanged (Fig. 2A).”

• AssumptionRests on an assumption: The Monte-Carlo simulation models using spheres and extra-axonal space accurately represent cellularity and vasogenic edema in tissue.

Increased cellularity did not affect DBSI-derived fiber radial diffusivity.

“Results showed that increased cellularity or vasogenic edema did not affect the DBSI-derived fiber FA, axial or radial diffusivity.”

DAPI-positive nuclear counts correlated with DBSI-estimated cellularity (r = 0.65, p = 0.042).

“DAPI-positive nuclear counts correlated with DBSI estimated cellularity (r = 0.65, p = 0.042; Fig. 7E).”

• AssumptionRests on an assumption: Current MRI methods do not assess specific elements of CNS pathology.

In Monte-Carlo simulation with increasing extra-fiber cellularity, DTI-derived λ⊥ was minimally affected.

“DTI derived λ⊥ was minimally affected in contrast to the more profound changes seen in λ∥.”

• AssumptionRests on an assumption: The fixed trigeminal nerve phantom with 2% agar gel is a valid model of vasogenic edema.

Apparent heterogeneity of cell infiltration was consistent with a heterogeneous increase in DAPI-positive cell nuclei staining.

“Apparent heterogeneity of cell infiltration was also seen, consistent with a heterogeneous increase in DAPI-positive cell nuclei staining (Figs. 6K and k).”

• AssumptionRests on an assumption: Current MRI methods do not assess specific elements of CNS pathology.

The 25-direction diffusion scheme was adequate for DBSI analysis of coherent white matter tracts to quantify the extent of cellularity and edema.

“In the present study, the 25-direction diffusion scheme was assessed and proved to be adequate for DBSI analysis of coherent white matter tracts to quantify the extent of cellularity and edema.”

The region of increased cellularity did not always correspond to visible nerve damage.

“Interestingly, the region of increased cellularity did not always correspond to visible nerve damage (Fig. 6a, b, and e).”

Phantom and animal studies demonstrated a linear correlation between the restricted isotropic diffusion fraction and DAPI nuclei counts.

“Restricted isotropic diffusion fractions have been assigned as “cellularity” fractions because our phantom and animal studies demonstrated a linear correlation between this restricted isotropic diffusion fraction and DAPI nuclei counts.”

• AssumptionRests on an assumption: Current MRI methods do not assess specific elements of CNS pathology.

Employing a 99-direction diffusion-encoding scheme, DBSI has shown promise to accurately detect and quantify crossing fibers, axonal injury, demyelination, and inflammation-associated cell infiltration and edema in both ex vivo phantom and in vivo mouse brain.

“Employing a 99-direction diffusion-encoding scheme, DBSI has shown promise to accurately detect and quantify crossing fibers, axonal injury, demyelination, and inflammation-associated cell infiltration and edema in both ex vivo phantom and in vivo mouse brain.”

DTI successfully detects axon and myelin injury through decreased axial diffusivity and increased radial diffusivity in animal models of CNS diseases and injuries.

“Diffusion tensor imaging (DTI) successfully detects axon and myelin injury through decreased axial diffusivity (λ∥, parallel to white matter tract) and increased radial diffusivity (λ⊥, perpendicular to white matter tract) in animal models of central nervous system (CNS) diseases and injuries (Song et al., 2002, Kim et al., 2006, Sun et al., 2006, DeBoy et al., 2007).”

• cites Song et al., 2002 sourceNot checked — no full text of the source
• cites Kim et al., 2006 sourceNot checked — no full text of the source
• cites Sun et al., 2006 sourceNot checked — no full text of the source

Diffusion anisotropy of white matter tracts in the presence of multiple structural and pathological compartments poses significant challenges in DTI analysis because non-Gaussian models or multiple diffusion tensors are needed to reflect tissue and pathological complexity.

“Thus, diffusion anisotropy of white matter tracts in the presence of multiple structural and pathological compartments poses significant challenges in DTI analysis of white matter tracts since non-Gaussian models or multiple diffusion tensors are needed to reflect the tissue and pathological complexity.”

A reduced scanning time can be achieved by significantly reducing the number of diffusion weighted images.

“A reduced scanning time can be achieved by significantly reducing the number of diffusion weighted images.”

Increased cellularity decreases DTI-derived ADC.

“Increased cellularity has been demonstrated to decrease DTI-derived ADC (Anderson et al., 2000).”

• cites Anderson et al., 2000 sourceNot checked — no full text of the source

Both axon/myelin and extra-fiber pathological changes can impact DTI-derived metrics.

“It is clear that both axon/myelin and extra-fiber pathological changes can impact DTI-derived metrics.”

The increased sum of restricted and non-restricted isotropic diffusion fractions may be a good indicator of inflammation (0.24 ± 0.12 vs. 0.06 ± 0.02, EAE vs. Sham, p = 0.024).

“The increased sum of restricted and non-restricted isotropic diffusion fractions may be a good indicator of inflammation, 0.24 ± 0.12 vs. 0.06 ± 0.02 (EAE vs. Sham, p = 0.024).”

DBSI-derived indices hold promise to reflect the various specific components of CNS pathology.

“In addition, DBSI-derived indices hold the promise to reflect the various specific components of the CNS pathology to identify new beneficial therapies that affect more than just inflammation.”

DBSI quantified the extent of increased cellularity and vasogenic edema resulting from inflammation.

“In addition, DBSI quantified the extent of increased cellularity and vasogenic edema resulting from inflammation.”

In cases where crossing fibers are not of concern, such as optic nerve and spinal cord white matter tracts, a reduced data acquisition scheme may be sufficient for DBSI analysis.

“In the case where crossing fibers are not of concern such as when studying optic nerve and spinal cord white matter tracts, a reduced data acquisition scheme may be sufficient for DBSI analysis.”

Such pathological profile can now be detected using DBSI noninvasively.

“Such pathological profile can now be detected using DBSI noninvasively.”

The close agreement between experimental measurements of fixed trigeminal nerve plus gel phantoms and Monte-Carlo simulations using restricted and non-restricted isotropic diffusion components supported the approach of including a linear combination of isotropic diffusion spectra to account for CNS inflammation and tissue loss.

“The close agreement between the experimental measurements of fixed trigeminal nerve plus gel phantoms and the results from Monte-Carlo simulations using restricted and non-restricted isotropic diffusion components to mimic cells and vasogenic edema supported our approach of including a linear combination of isotropic diffusion spectra to account for the effect of CNS inflammation and tissue loss.”

Both DTI- and DBSI-derived axial and radial diffusivity correctly detected axonal injury and demyelination in the optic nerve at the onset of optic neuritis.

“In the present study, both DTI and DBSI derived axial and radial diffusivity correctly detected axonal injury and demyelination in the optic nerve at the onset of optic neuritis.”

To properly model the effect of CNS inflammation and tissue loss and allow more accurate evaluation of white matter fiber tract diffusion properties, multiple extra-fiber diffusion components (both restricted and non-restricted isotropic diffusion components) must be identified and quantified.

“To properly model the effect of CNS inflammation and tissue loss, allowing a more accurate evaluation of the diffusion properties of white matter fiber tracts, multiple extra-fiber diffusion components (both restricted and non-restricted isotropic diffusion components) must be identified and quantified.”

Existing diffusion weighted data acquired using 25- or other multiple-direction encoding schemes with multiple b-values may be suitable for DBSI analysis without reacquiring new data to examine single coherent fiber tracts.

“Thus, existing diffusion weighted data acquired using 25- or other multiple-direction encoding schemes with multiple b-values may be suitable for DBSI analysis without reacquiring new data to examine the single coherent fiber tracts.”

DTI works well on these mice probably due to the counteraction of increased restricted and hindered/free diffusion tensor components.

“This suggests that DTI works well on these mice probably due to the counteraction of increased restricted and hindered/free diffusion tensor components.”

Various MR techniques used to assess inflammation of optic neuritis have rarely demonstrated the ability to image cellularity changes non-invasively.

“Various MR techniques have been used to assess inflammation of optic neuritis, but rarely demonstrated the ability to image cellularity changes non-invasively.”

DBSI provides a noninvasive imaging biomarker of inflammation without injecting gadolinium contrast agent.

“Without injecting Gadolinium contrast agent, DBSI provides a noninvasive imaging biomarker of inflammation, which could potentially enhance our capability to more accurately evaluate neuro-inflammation in CNS disease.”

In coherent white matter tracts such as optic nerves, the 25-direction diffusion weighting scheme is sufficient to estimate fiber λ∥, λ⊥, FA, and restricted (correlating with cellularity) and non-restricted diffusion fractions.

“We have also demonstrated using in vivo DBSI and post-MRI immunohistochemistry on EAE mice that in the coherent white matter tracts such as optic nerves the 25-direction diffusion weighting scheme is sufficient to estimate fiber λǁ, λ⊥, FA, and restricted (correlating with cellularity) and non-restricted diffusion fractions.”