dke-questions > kmean vs mkt
Showing 1-4 of 4 posts
Feb 22, 2018 03:02 PM | Laura Mancini - UCLH NHS FT
kmean vs mkt
Hello,
sorry, what is the difference between the kmean and the mkt maps created by the dke software?
Many thanks
Laura
sorry, what is the difference between the kmean and the mkt maps created by the dke software?
Many thanks
Laura
Feb 23, 2018 07:02 PM | Hunter Moss
RE: kmean vs mkt
Hi Laura,
Great question.
Kmean (MK) is the average kurtosis over all directions (in theory an infinite amount). MK requires knowledge of both the diffusion and kurtosis tensors. Practically speaking, we advise gathering at least 30 directions at both b = 1000 and 2000 s/mm^2 (more often we gather 64 or 128 directions).
MKT is the mean kurtosis tensor which is a bit different from MK. The kurtosis tensor, W, can be calculated (with the omission of a scaling factor) and then averaged. So, MKT is really the mean (of the) kurtosis tensor. Explicitly, MKT only approximates MK. However, if diffusion tensor were to be completely isotropic then MKT and MK would be equal.
The paper for MKT is: Hansen B, Lund TE, Sangill R, Jespersen SN. Experimentally and computationally fast method for estimation of mean kurtosis. Magn Reson. Med 2013; 69:1754-1760.
MK and MKT have shown to have similar information in this publication: Glenn GR, Helpern JA, Tabesh A, Jensen JH. Quantitative assessment of diffusional kurtosis anisotropy. NMR Biomed 2015;28:448–459.
I hope this answers your question but if not, please feel free to ask for more clarification.
Thanks for using DKE!
-Hunter
Great question.
Kmean (MK) is the average kurtosis over all directions (in theory an infinite amount). MK requires knowledge of both the diffusion and kurtosis tensors. Practically speaking, we advise gathering at least 30 directions at both b = 1000 and 2000 s/mm^2 (more often we gather 64 or 128 directions).
MKT is the mean kurtosis tensor which is a bit different from MK. The kurtosis tensor, W, can be calculated (with the omission of a scaling factor) and then averaged. So, MKT is really the mean (of the) kurtosis tensor. Explicitly, MKT only approximates MK. However, if diffusion tensor were to be completely isotropic then MKT and MK would be equal.
The paper for MKT is: Hansen B, Lund TE, Sangill R, Jespersen SN. Experimentally and computationally fast method for estimation of mean kurtosis. Magn Reson. Med 2013; 69:1754-1760.
MK and MKT have shown to have similar information in this publication: Glenn GR, Helpern JA, Tabesh A, Jensen JH. Quantitative assessment of diffusional kurtosis anisotropy. NMR Biomed 2015;28:448–459.
I hope this answers your question but if not, please feel free to ask for more clarification.
Thanks for using DKE!
-Hunter
Feb 27, 2018 09:02 AM | Live Eikenes
RE: kmean vs mkt
Hi,
I have a follow up question regarding MKT. Is MKT the same as GFA?
Kind regards,
Live
I have a follow up question regarding MKT. Is MKT the same as GFA?
Kind regards,
Live
Feb 27, 2018 08:02 PM | Hunter Moss
RE: kmean vs mkt
Hi Live,
No, they are not equivalent by any means. MKT as was explained in the earlier question is the mean of the kurtosis tensor without any information from the diffusion tensor. MK and MKT are equivalent if and only if the diffusion tensor is fully isotropic.
The generalized FA (gFA), is analogous to the conventional FA, but unlike FA, gFA uses information from the diffusion orientation distribution function (dODF) derived from DKI using both the full diffusion and kurtosis tensors. The dODF is calculated by taking the radial projection of the diffusion displacement probability density function (dPDF) and has a weighting factor to increase sensitivity to longer diffusion displacements.
More information regarding this topic is laid out in great detail in Glenn, G. R., Helpern, J. A., Tabesh, A., & Jensen, J. H. (2015). Quantitative Assessment of Diffusional Kurtosis Anisotropy. NMR in Biomedicine, 28(4), 448–459. http://doi.org/10.1002/nbm.3271.
Thanks for the question!
-Hunter
No, they are not equivalent by any means. MKT as was explained in the earlier question is the mean of the kurtosis tensor without any information from the diffusion tensor. MK and MKT are equivalent if and only if the diffusion tensor is fully isotropic.
The generalized FA (gFA), is analogous to the conventional FA, but unlike FA, gFA uses information from the diffusion orientation distribution function (dODF) derived from DKI using both the full diffusion and kurtosis tensors. The dODF is calculated by taking the radial projection of the diffusion displacement probability density function (dPDF) and has a weighting factor to increase sensitivity to longer diffusion displacements.
More information regarding this topic is laid out in great detail in Glenn, G. R., Helpern, J. A., Tabesh, A., & Jensen, J. H. (2015). Quantitative Assessment of Diffusional Kurtosis Anisotropy. NMR in Biomedicine, 28(4), 448–459. http://doi.org/10.1002/nbm.3271.
Thanks for the question!
-Hunter