Spherical deconvolution of multichannel diffusion MRI data with non-Gaussian noise models and total variation spatial regularization
Due to a higher capability in resolving white matter fiber crossings, Spherical Deconvolution (SD) methods have become very popular in brain fiber-tracking applications. However, while some of these estimation algorithms assume a central Gaussian distribution for the MRI noise, its real distribution...
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Journal article |
| Published: |
2014
|
| _version_ | 1797085284688461824 |
|---|---|
| author | Canales-Rodríguez, E Daducci, A Sotiropoulos, SN Caruyer, E Aja-Fernández, S Radua, J Mendizabal, Y Iturria-Medina, Y Melie-García, L Alemán-Gómez, Y Thiran, J Sarró, S Pomarol-Clotet, E Salvador, R |
| author_facet | Canales-Rodríguez, E Daducci, A Sotiropoulos, SN Caruyer, E Aja-Fernández, S Radua, J Mendizabal, Y Iturria-Medina, Y Melie-García, L Alemán-Gómez, Y Thiran, J Sarró, S Pomarol-Clotet, E Salvador, R |
| author_sort | Canales-Rodríguez, E |
| collection | OXFORD |
| description | Due to a higher capability in resolving white matter fiber crossings, Spherical Deconvolution (SD) methods have become very popular in brain fiber-tracking applications. However, while some of these estimation algorithms assume a central Gaussian distribution for the MRI noise, its real distribution is known to be non-Gaussian and to depend on many factors such as the number of coils and the methodology used to combine multichannel signals. Indeed, the two prevailing methods for multichannel signal combination lead to noise patterns better described by Rician and noncentral Chi distributions. Here we develop a Robust and Unbiased Model-BAsed Spherical Deconvolution (RUMBA-SD) technique intended to deal with realistic MRI noise. The algorithm relies on a maximum a posteriori formulation based on Rician and noncentral Chi likelihood models and includes a total variation (TV) spatial regularization term. By means of a synthetic phantom contaminated with noise mimicking patterns generated by data processing in multichannel scanners, the performance of RUMBA-SD is compared to that of other well-established SD methods (i.e., CSD and dRL-SD). The inclusion of proper likelihood models and TV regularization in RUMBA-SD leads to an increased ability to resolve fiber crossings with smaller inter-fiber angles and an increased robustness to noise. Finally, the proposed method is also validated in human brain data, producing the most stable fiber reconstructions in front of differing noise types and diffusion schemes based on a small number of gradient directions. |
| first_indexed | 2024-03-07T02:06:44Z |
| format | Journal article |
| id | oxford-uuid:9f4185ae-4432-47c7-abc1-d8a3a48837aa |
| institution | University of Oxford |
| last_indexed | 2024-03-07T02:06:44Z |
| publishDate | 2014 |
| record_format | dspace |
| spelling | oxford-uuid:9f4185ae-4432-47c7-abc1-d8a3a48837aa2022-03-27T00:56:00ZSpherical deconvolution of multichannel diffusion MRI data with non-Gaussian noise models and total variation spatial regularizationJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:9f4185ae-4432-47c7-abc1-d8a3a48837aaSymplectic Elements at Oxford2014Canales-Rodríguez, EDaducci, ASotiropoulos, SNCaruyer, EAja-Fernández, SRadua, JMendizabal, YIturria-Medina, YMelie-García, LAlemán-Gómez, YThiran, JSarró, SPomarol-Clotet, ESalvador, RDue to a higher capability in resolving white matter fiber crossings, Spherical Deconvolution (SD) methods have become very popular in brain fiber-tracking applications. However, while some of these estimation algorithms assume a central Gaussian distribution for the MRI noise, its real distribution is known to be non-Gaussian and to depend on many factors such as the number of coils and the methodology used to combine multichannel signals. Indeed, the two prevailing methods for multichannel signal combination lead to noise patterns better described by Rician and noncentral Chi distributions. Here we develop a Robust and Unbiased Model-BAsed Spherical Deconvolution (RUMBA-SD) technique intended to deal with realistic MRI noise. The algorithm relies on a maximum a posteriori formulation based on Rician and noncentral Chi likelihood models and includes a total variation (TV) spatial regularization term. By means of a synthetic phantom contaminated with noise mimicking patterns generated by data processing in multichannel scanners, the performance of RUMBA-SD is compared to that of other well-established SD methods (i.e., CSD and dRL-SD). The inclusion of proper likelihood models and TV regularization in RUMBA-SD leads to an increased ability to resolve fiber crossings with smaller inter-fiber angles and an increased robustness to noise. Finally, the proposed method is also validated in human brain data, producing the most stable fiber reconstructions in front of differing noise types and diffusion schemes based on a small number of gradient directions. |
| spellingShingle | Canales-Rodríguez, E Daducci, A Sotiropoulos, SN Caruyer, E Aja-Fernández, S Radua, J Mendizabal, Y Iturria-Medina, Y Melie-García, L Alemán-Gómez, Y Thiran, J Sarró, S Pomarol-Clotet, E Salvador, R Spherical deconvolution of multichannel diffusion MRI data with non-Gaussian noise models and total variation spatial regularization |
| title | Spherical deconvolution of multichannel diffusion MRI data with
non-Gaussian noise models and total variation spatial regularization |
| title_full | Spherical deconvolution of multichannel diffusion MRI data with
non-Gaussian noise models and total variation spatial regularization |
| title_fullStr | Spherical deconvolution of multichannel diffusion MRI data with
non-Gaussian noise models and total variation spatial regularization |
| title_full_unstemmed | Spherical deconvolution of multichannel diffusion MRI data with
non-Gaussian noise models and total variation spatial regularization |
| title_short | Spherical deconvolution of multichannel diffusion MRI data with
non-Gaussian noise models and total variation spatial regularization |
| title_sort | spherical deconvolution of multichannel diffusion mri data with non gaussian noise models and total variation spatial regularization |
| work_keys_str_mv | AT canalesrodrigueze sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT daduccia sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT sotiropoulossn sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT caruyere sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT ajafernandezs sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT raduaj sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT mendizabaly sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT iturriamedinay sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT meliegarcial sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT alemangomezy sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT thiranj sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT sarros sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT pomarolclotete sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization AT salvadorr sphericaldeconvolutionofmultichanneldiffusionmridatawithnongaussiannoisemodelsandtotalvariationspatialregularization |