Artikel
Comparison of functional MRI data between SPM12 and FreeSurfer using a customised anatomical atlas
Vergleich funktioneller MRT-Daten zwischen SPM12 und FreeSurfer mittels eines angepassten anatomischen Atlasses
Suche in Medline nach
Autoren
-
Katharina Hense
- Universitätsklinikum Regensburg, Klinik und Poliklinik für Neurochirurgie, Regensburg, Deutschland
-
Christina Wendl
- Universitätsklinikum Regensburg, Institut für Röntgendiagnostik, Regensburg, Deutschland
-
Nils-Ole Schmidt
- Universitätsklinikum Regensburg, Klinik und Poliklinik für Neurochirurgie, Regensburg, Deutschland
-
Katharina Rosengarth
- Universitätsklinikum Regensburg, Klinik und Poliklinik für Neurochirurgie, Regensburg, Deutschland
| Veröffentlicht: | 25. Mai 2022 |
|---|
Gliederung
Text
Objective: Many studies comparing the performance of different softwares for analyzing MRI data are limited to the examination of structural data. In doing so, they show that analysis tools such as SPM12 and FreeSurfer do not or only slightly differ in analyses such as the calculation of gray and white matter volume or cortical thickness. In contrast, a closer examination of the results of the analysis of functional MRI data cannot be found. One reason for this could be a lack of an objective evaluation system, since both methods do not have a common atlas of brain anatomy. Therefore, in this study, we would like to present a way to use a shared atlas in both methods and thus compare the results of functional MRI analyses.
Methods: Here, we retrospectively analyzed functional MRI data from 12 healthy subjects who participated in a study examining a language paradigm. Since SPM12 and FreeSurfer use different atlases for anatomical parcellation, we looked for a possibility to compare the results of the two analysis methods. For this purpose, we used the established AAL3v1 atlas (1x1x1mm resolution) to obtain a binarized Region of Interest (ROI) files for each included brain area. Some ROIs like the subdivided areas of the thalamus were reassembled due to small region size. We registered the ROIs to the individual brain in native space, a normalization to standard MNI space was not performed. After 1st level analysis of the functional data in FreeSurfer and SPM12, the cluster size of actived voxels were compared between the two analysis tools.
Results: We analyzed the individual results calculated with both methods using paired t-tests, which were Benjamini-Hochberg (False Discovery Rate; FDR) corrected for multiple comparisons. The analysis of the cluster size of activated voxels of each Region of Interest showed no significant differences in none of the 94 ROIs.
Conclusion: We conclude from the results that creating an atlas that works in SPM12 and FreeSurfer is possible. This also allows the comparison of functional as well as structural results of both softwares. Since we have not found any study comparing functional MRI data between the two analysis software, we believe that this is the first successful attempt to objectively compare functional results between FreeSurfer and SPM12.
