Limits...
The benefits of skull stripping in the normalization of clinical fMRI data.

Fischmeister FP, Höllinger I, Klinger N, Geissler A, Wurnig MC, Matt E, Rath J, Robinson SD, Trattnig S, Beisteiner R - Neuroimage Clin (2013)

Bottom Line: The optimum procedure has not been conclusively established, and a critical dichotomy is whether to use input data sets which contain skull signal, or whether skull signal should be removed.Brain activation changes related to deskulled/not-deskulled input data are determined in the context of very recently developed (New Segment, Unified Segmentation) and standard normalization approaches.Analysis of structural and functional data demonstrates that skull stripping improves language localization in MNI space - particularly when used in combination with the New Segment normalization technique.

View Article: PubMed Central - PubMed

Affiliation: Study Group Clinical fMRI, Department of Neurology, Medical University of Vienna, Austria ; High Field MR Center, Medical University of Vienna, Austria.

ABSTRACT
Establishing a reliable correspondence between lesioned brains and a template is challenging using current normalization techniques. The optimum procedure has not been conclusively established, and a critical dichotomy is whether to use input data sets which contain skull signal, or whether skull signal should be removed. Here we provide a first investigation into whether clinical fMRI benefits from skull stripping, based on data from a presurgical language localization task. Brain activation changes related to deskulled/not-deskulled input data are determined in the context of very recently developed (New Segment, Unified Segmentation) and standard normalization approaches. Analysis of structural and functional data demonstrates that skull stripping improves language localization in MNI space - particularly when used in combination with the New Segment normalization technique.

No MeSH data available.


Related in: MedlinePlus

One-sample t-test group results. Significant activation above a threshold of p < 0.001 uncorrected is overlaid on the brain extracted or the standard MNI152 templates provided by FSL. Note that the position of the activation cluster differs (c.f. slice 18 showing almost no activation for the Unified Segmentation Model with skull-stripping as indicated with a red circle) and the Wernicke peak-voxel is shifted between normalization pipelines > 1 cm (indicated with an arrow, locations are given in MNI coordinates). Only slices covering the Wernicke area are shown. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3814956&req=5

f0030: One-sample t-test group results. Significant activation above a threshold of p < 0.001 uncorrected is overlaid on the brain extracted or the standard MNI152 templates provided by FSL. Note that the position of the activation cluster differs (c.f. slice 18 showing almost no activation for the Unified Segmentation Model with skull-stripping as indicated with a red circle) and the Wernicke peak-voxel is shifted between normalization pipelines > 1 cm (indicated with an arrow, locations are given in MNI coordinates). Only slices covering the Wernicke area are shown. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Mentions: In more detail, we found that the skulled brains, for which normalization quality was inferior (see Fig. 5) lead to misleading activation. This is shown by a significant skulled > deskulled activation in the left inferior parietal cortex (− 49 − 32 18 in Table 4, red circles in Fig. 6) and left anterior superior temporal gyrus (− 55 − 6 4 in Table 4) — both clearly outside the classical Wernicke core. The term “misleading” seems justified for 3 reasons: (1) the remote parietal and temporal activations were not seen with standard clinical thresholds (clinical patient reports), (2) the larger the mismatch between template and brain (which was largest with skulled data), the greater the change in the location of Wernicke activation, and (3) no pipeline changed Wernicke activations significantly in relation to local neuroanatomy (Fig. 8), but neuroanatomy changed in relation to the MNI coordinates (i.e. a brain–template mismatch occurred with skulled data). Therefore, the conclusion must be that normalization of skulled brains shifts part of the “correct” Wernicke activations to “wrong” MNI coordinates in the temporo-parietal cortex. A similar temporo-parietal effect was found for the standard normalization technique. Standard normalization generated a misleading activation increase in left supramarginal gyrus and left middle temporal gyrus outside the Wernicke core.


The benefits of skull stripping in the normalization of clinical fMRI data.

Fischmeister FP, Höllinger I, Klinger N, Geissler A, Wurnig MC, Matt E, Rath J, Robinson SD, Trattnig S, Beisteiner R - Neuroimage Clin (2013)

One-sample t-test group results. Significant activation above a threshold of p < 0.001 uncorrected is overlaid on the brain extracted or the standard MNI152 templates provided by FSL. Note that the position of the activation cluster differs (c.f. slice 18 showing almost no activation for the Unified Segmentation Model with skull-stripping as indicated with a red circle) and the Wernicke peak-voxel is shifted between normalization pipelines > 1 cm (indicated with an arrow, locations are given in MNI coordinates). Only slices covering the Wernicke area are shown. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3814956&req=5

f0030: One-sample t-test group results. Significant activation above a threshold of p < 0.001 uncorrected is overlaid on the brain extracted or the standard MNI152 templates provided by FSL. Note that the position of the activation cluster differs (c.f. slice 18 showing almost no activation for the Unified Segmentation Model with skull-stripping as indicated with a red circle) and the Wernicke peak-voxel is shifted between normalization pipelines > 1 cm (indicated with an arrow, locations are given in MNI coordinates). Only slices covering the Wernicke area are shown. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Mentions: In more detail, we found that the skulled brains, for which normalization quality was inferior (see Fig. 5) lead to misleading activation. This is shown by a significant skulled > deskulled activation in the left inferior parietal cortex (− 49 − 32 18 in Table 4, red circles in Fig. 6) and left anterior superior temporal gyrus (− 55 − 6 4 in Table 4) — both clearly outside the classical Wernicke core. The term “misleading” seems justified for 3 reasons: (1) the remote parietal and temporal activations were not seen with standard clinical thresholds (clinical patient reports), (2) the larger the mismatch between template and brain (which was largest with skulled data), the greater the change in the location of Wernicke activation, and (3) no pipeline changed Wernicke activations significantly in relation to local neuroanatomy (Fig. 8), but neuroanatomy changed in relation to the MNI coordinates (i.e. a brain–template mismatch occurred with skulled data). Therefore, the conclusion must be that normalization of skulled brains shifts part of the “correct” Wernicke activations to “wrong” MNI coordinates in the temporo-parietal cortex. A similar temporo-parietal effect was found for the standard normalization technique. Standard normalization generated a misleading activation increase in left supramarginal gyrus and left middle temporal gyrus outside the Wernicke core.

Bottom Line: The optimum procedure has not been conclusively established, and a critical dichotomy is whether to use input data sets which contain skull signal, or whether skull signal should be removed.Brain activation changes related to deskulled/not-deskulled input data are determined in the context of very recently developed (New Segment, Unified Segmentation) and standard normalization approaches.Analysis of structural and functional data demonstrates that skull stripping improves language localization in MNI space - particularly when used in combination with the New Segment normalization technique.

View Article: PubMed Central - PubMed

Affiliation: Study Group Clinical fMRI, Department of Neurology, Medical University of Vienna, Austria ; High Field MR Center, Medical University of Vienna, Austria.

ABSTRACT
Establishing a reliable correspondence between lesioned brains and a template is challenging using current normalization techniques. The optimum procedure has not been conclusively established, and a critical dichotomy is whether to use input data sets which contain skull signal, or whether skull signal should be removed. Here we provide a first investigation into whether clinical fMRI benefits from skull stripping, based on data from a presurgical language localization task. Brain activation changes related to deskulled/not-deskulled input data are determined in the context of very recently developed (New Segment, Unified Segmentation) and standard normalization approaches. Analysis of structural and functional data demonstrates that skull stripping improves language localization in MNI space - particularly when used in combination with the New Segment normalization technique.

No MeSH data available.


Related in: MedlinePlus