Profitiert multivariate Musteranalyse von fMRT-Daten mit hoher Auflösung und Sensitivität bei hoher Magnetfeldstärke (7T)?
Medizinische Physik, Biomedizinische Technik
Zusammenfassung der Projektergebnisse
A number of open scientific questions related to decoding of visual cortex activation as measured by BOLD fMRI have motivated this project: What is the effect of acquisition resolution on decoding accuracy? - What is the spatial scale of orientation discriminating signals? - What is the effect of spatial filtering of decoding accuracy? - How do tSNR and signal amplitude influence the decoding accuracy? - What is the contribution of draining veins to the decoding signal? - What is the effect of field strength on the above questions? - Are the results specific to the visual system? We have performed extensive and detailed measurements at different spatial resolutions at 7T and 3Ts extension of the already openly available data from the StudyForest project. All data have been made publically available. The data have been analyzed very carefully with various spatial preprocessing options (spatial frequency filters, spatial subsampling and venous and non-venous voxel segmentation). The results can be summarized into these major findings: The optimal resolution for decoding of the orientation of visual grating stimuli is between 2 and 3mm. - The tSNR is relevant for decoding accuracy, while the signal amplitude has little effect. - Spatial filtering of data with 2mm resolution or less does not increase decoding accuracy. For higher resolution data, spatial filtering with 2-3mm FWHM increased decoding accuracy. Spatial downsampling did not improve decoding performance. - The spatial frequency of the decoding-relevant signal modulation lies between 4.5 and 16mm and is consistent with the best decoding performance between 2 and 3mm resolution as predicted by the Shannon sampling theorem. - Venous signal does contribute to the decoding in particular for higher resolution. - The results were confirmed at 3T and at 7T with coarser spatial signal contributions at 3T and lower decoding performance at 3T even for similar tSNR values. The above findings were consistent in an auditory discrimination task leading to the hypothesis that the results may be generalizable across different cortical regions In conclusion, the project has fully completed the proposed experimental work and extended this towards 3T and other brain regions. The results confirmed the hypothesis of higher decoding accuracy at 7T over 3T and of higher spatial frequency contributions at higher field strength. A number of previous hypotheses that were based on either simulations or single measurement points could also be confirmed. The results contribute to the understanding of decoding mechanisms and have resolved some of the discrepancies in the literature regarding the initial questions. All findings can be replicated from the publically available data.
Projektbezogene Publikationen (Auswahl)
- (2016). A studyforrest extension, retinotopic mapping and localization of higher visual areas. Scientific Data, 3, 160093
Sengupta A, Kaule FR, Guntupalli JS, Hoffmann MB, Häusler C, Stadler J, Hanke M
(Siehe online unter https://dx.doi.org/10.1038/sdata.2016.93) - (2016). A studyforrest extension, simultaneous fMRI and eye gaze recordings during prolonged natural stimulation. Scientific Data, 3,160092
Hanke M, Adelhöfer N, Kottke D, Iacovella V, Sengupta A, Kaule FR, Nigbur R, Waite AQ, Baumgartner F, Stadler J
(Siehe online unter https://dx.doi.org/10.1038/sdata.2016.92) - (2016). The effect of acquisition resolution and magnetic field strength on multivariate decoding of fMRI
Sengupta, A.
- (2017). The effect of acquisition resolution on orientation decoding from V1 BOLD fMRI at 7 Tesla. Neuroimage, 148, 64-76
Sengupta, A., Yakupov, R., Speck, O., Pollmann, S. & Hanke, M.
(Siehe online unter https://doi.org/10.1016/j.neuroimage.2016.12.040) - (2017). Ultra high-field (7 Tesla) multi-resolution fMRI data for orientation decoding in visual cortex. Data in Brief
Sengupta, A., Yakupov, R., Speck, O., Pollmann, S. & Hanke, M.
(Siehe online unter https://dx.doi.org/10.1016/j.dib.2017.05.014) - (2018) Spatial band-pass filtering aids decoding musical genres from auditory cortex 7T fMRI [version 2; referees: 2 approved]. F1000Research 2018, 7:142
Sengupta A, Pollmann S and Hanke M.
(Siehe online unter https://dx.doi.org/10.12688/f1000research.13689.2) - (2018b) The effect of acquisition resolution on orientation decoding from V1: comparison of 3T and 7T
Sengupta, A, Speck, O, Yakupov, R, Kanowski, M, Tempelmann, C, Pollmann, S, Hanke, M
(Siehe online unter https://doi.org/10.1101/305417)
