This web site provides files associated with the following paper: > Kay, K., Jamison, K.W., Vizioli, L., Zhang, R., Margalit, E., & Ugurbil, K. A critical assessment of data quality and venous effects in sub-millimeter fMRI. NeuroImage (2019). The paper is available at [https://doi.org/10.1016/j.neuroimage.2019.02.006][1]. [1]: https://doi.org/10.1016/j.neuroimage.2019.02.006 The files on this OSF site include: (1) an archive of the code used for the paper and (2) Supplementary Movies 1-6. Raw data for many of the datasets from this paper are available at OpenNeuro at https://doi.org/10.18112/openneuro.ds002702.v1.0.1. See also https://osf.io/j2wsc/ for additional information. ---------- ## Figure captions: ## **Supplementary Movie 1. Inspection of pre-processing results.** Movie available online (https://osf.io/s5kw7/). Each row corresponds to a distinct subject (S1–S5), and each column corresponds to a distinct slice of the GE-EPI acquisition (every 10th slice is shown, yielding a slice-to-slice distance of 0.8 mm × 10 slices = 8 mm). The movie cycles between the mean EPI volume, the T2-weighted anatomical volume, and the T1-weighted anatomical volume. For each volume, contours depicting the white-matter and pial surfaces are toggled on and off (green and cyan indicate left and right hemispheres, respectively). The results demonstrate that EPI undistortion, co-registration between functional and anatomical volumes, and cortical surface reconstruction all performed well. **Supplementary Movies 2–4. Functional volumes after pre-processing.** Movies available online (https://osf.io/s26yz/, https://osf.io/jbpv5/, https://osf.io/axfvt/). Row and column format same as Supplementary Movie 1. These movies show a sequence of 50 EPI volumes chosen randomly from all volumes acquired within a given scan session (which lasted ~80 min). The volumes are raw volumes aside from the temporal resampling and spatial resampling operations that comprised pre-processing (see Methods). Visualizing randomly chosen volumes (as opposed to volumes in chronological order) is a stringent test of data quality, as it accentuates instabilities over time. Three movies are provided: Supplementary Movie 2 shows results using the pre-processing described in the Methods, involving time-varying fieldmap estimates (multiple-fieldmap approach); Supplementary Movie 3 shows results using identical pre-processing procedures except that the first fieldmap acquired in each scan session is used as a static fieldmap estimate that is applied to each EPI volume before subsequent motion estimation and other processing steps (single-fieldmap approach); and Supplementary Movie 4 shows results using identical pre-processing procedures except that no fieldmap-based undistortion is applied (no-fieldmap approach). Inspection of Supplementary Movie 2 reveals the existence of some low spatial frequency artifacts. However, overall stability over time is high in most parts of the brain, indicating that data acquisition was stable and that motion correction and fieldmap-based EPI undistortion performed well. Inspection of Supplementary Movies 3 and 4 indicates that temporal stability is relatively high for the multiple-fieldmap and no-fieldmap approaches but is somewhat low for the single-fieldmap approach. We speculate that this temporal instability is the result of inaccurate undistortion of EPI volumes acquired distant in time from the single fieldmap and that the use of multiple fieldmaps helps mitigate this issue. **Supplementary Movie 5. Static susceptibility effects as a function of cortical depth.** Movie available online (https://osf.io/2b469/). This movie shows bias-corrected EPI intensities (posterior view, spherical surface), progressing from inner to outer cortical depths (Depth 6 through Depth 1). Rows indicate left and right hemispheres; columns indicate distinct subjects (S1–S5); and the colormap for each image ranges from 0–2, as in Figure 5 and Figure 9B. The large influence of cortical depth on static susceptibility effects is visible in this movie. **Supplementary Movie 6. Impact of registration on EPI intensities.** A, Full movie showing registration parameters and EPI intensities. Movie available online (https://osf.io/snb2h/). This movie shows changes in registration quality and surface-mapped EPI intensities as the registration between the EPI and T2 volumes progresses for an example subject (Subject S3). At the top left is a slice through the T2 volume that corresponds to a specific EPI slice. At the bottom left are the registration parameters. At the right are surface visualizations of the outermost depth (Depth 1, top row) and the innermost depth (Depth 6, bottom row) (posterior view, spherical surface). The surface visualizations show raw intensities sampled from the EPI volume (cubic interpolation). At each iteration, registration parameters are updated, followed by updates to the T2 slice (with a rapid alternation against the EPI slice to assess the match) and updates to the surface visualizations. We see that initially, the co-registration between EPI and T2 is poor and there are large swaths of darkness in the surface visualizations. Over time, the co-registration improves and the swaths of darkness are reduced. Overall, this movie provides intuition for how imperfections in registration may lead to apparent dark intensities, and suggests that the final registration solution and its corresponding patterns of EPI intensities are robust and accurate. B, Movie showing only surface visualizations. Movie available online (https://osf.io/f28mc/). This movie progresses much more rapidly than the movie from panel A. We see that over the course of the optimization, large changes occur rapidly in the first several iterations and then small refinements take place until the search settles to a local minimum.