Measuring and reducing the carbon footprint of fMRI preprocessing in fMRIPrep

Description: These maps were generated as part of the paper 'Measuring and reducing the carbon footprint of fMRI preprocessing in fMRIPrep'. This study comprised of fMRI data for a single run of a stop signal task from 257 subjects taken from the publicly available UCLA Consortium for Neuropsychiatric Phenomics LA5c Study (https://openneuro.org/datasets/ds000030/versions/1.0.0). In addition to these group-level maps, raw data used in statistical analysis (including ROI analysis in FSL Featquery) can be found on the Open Science Framework (https://osf.io/4e26k/). Each map type is split by Pipeline ID (according to the preprocessing conducted in fMRIPrep - details of each ID are detailed in Table 2). This collection contains three types of maps. *** Activation Count: These maps reflect the percentage of subjects showing significant individual-level activation (as determined in first-level analysis in FSL FEAT, thresholded at Z > 3.1, p < .05) in each voxel for both the ‘go > successful stop’ and ‘successful stop > go’ contrasts (seperate maps for each contrast). Corresponds to Supplementary Figure 7. *** Group Level: Unthresholded group-level activation maps taken from FSL FEAT. Includes maps associated with the contrasts of both 'go > successful stop' and 'successful stop > go' (these maps are exact inverses). Corresponds to Supplementary Figure 8. *** Timeseries SD Maps: Standard deviation (SD) maps reflect the sample’s mean standard deviation of preprocessed timeseries values for each voxel across all 180 volumes. Exracted following preprocessing in fMRIPrep but prior to smoothing in AFNI. Missing for Pipeline 5 given that smoothing for this pipeline was conducted during preprocessing in fMRIPrep (as part of ICA AROMA), and as such an equivilant map could not be extracted. Corresponds to Supplementary Figure 9. *** Activation SD Maps: Standard deviation (SD) maps reflect the sample’s mean standard deviation of z-statistics from statistical modelling in FSL FEAT, for each voxel. Extracted from each subject's unthresholded contrast of go > successful stop. Corresponds to Supplementary Figure 10.

Private Collection: To share the link to this collection, please use the private url: https://neurovault.org/collections/JNAGJRUI/