This DATSETNAMEreadme.txt file was generated on 2019-03-07 by Philippa J. Johnson. Updated: 2019-06-28 ------------------- GENERAL INFORMATION ------------------- 1. Title of Dataset Equine Stereotactic Population Average Brain Atlas with Neuroanatomic Correlation 2. Author Information Philippa J. Johnson, Valentin Janvier, Wenming Luh, Marnie Fitz-Maurice, Theresa Southard, Erica Barry Principal Investigator Contact Information Name:Philippa J. Johnson Institution:Cornell College of Veterinary Medicine Address:930 Campus Rd, Ithaca, NY 13081 USA Email:pjj43@cornell.edu 3. Date of data collection 2017-2018 4. Geographic location of data collection (where was data collected?): Cornell Magnetic Resonance Imaging Facility, Martha Van Rensselaer Hall, Cornell University, Ithaca, NY. 5. Information about funding sources that supported the collection of the data: This research was funded by the Harry M. Zweig memorial fund for equine research. -------------------------- SHARING/ACCESS INFORMATION -------------------------- 1. Licenses/restrictions placed on the data: Data to be used for research and teaching purposes. Data are shared under a CC0 declaration, but the authors request proper attribution. If you use this data, please cite it as: Johnson, Philippa J., Valentin Janvier, Wen-Ming Luh, Marnie Fitz-Maurice, Teresa Southard, Erica F. Barry. (2019) Equine Stereotactic Population Average Brain Atlas with Neuroanatomic Correlation [datset]. Cornell University Library eCommons Repository. https://doi.org/10.7298/cyrs-7b51.2 2. Was data derived from another source? No --------------------- DATA & FILE OVERVIEW --------------------- File List Tissue Masks 1. Filename: Amygdala_pub.nii.gz a. Short description: Manual segmentation of both Amygdala. 2. Filename: Caudate_Nuclei_pub.nii.gz a. Short description: Manual segmentation of both Caudate Nuclei. 3. Filename: Corpus_Callosum_pub.nii.gz a. Short description: Manual segmentation of the corpus callosum. 4. Filename: Fornix_pub.nii.gz a. Short description: Manual segmentation of the fornix. 5. Filename: Globus_Palidus_pub.nii.gz a. Short description: Manual segmentation of both globus palidus. 6. Filename: Hypothalamus_pub.nii.gz a. Short description: Manual segmentation of the hypothalamus. 7. Filename: Left_Hippocampus_pub.nii.gz a. Short description: Manual segmentation of the left hippocampus. 8. Filename: Medulla_Oblongata_pub.nii.gz a. Short description: Manual segmentation of the medulla oblongata. 9. Filename: Mesencephalon_pub.nii.gz a. Short description: Manual segmentation of the mesencephalon. 10. Filename: Olfactory_Bulbs_pub.nii.gz a. Short description: Manual segmentation of the olfactory bulbs. 11. Filename: Optic_pub.nii.gz a. Short description: Manual segmentation of the optic pathway including caudal nerve, chiasm and rostral tracts. 12. Filename: Pineal_Gland_pub.nii.gz a. Short description: Manual segmentation of the pineal gland. 13. Filename: Pons_pub.nii.gz a. Short description: Manual segmentation of the pons. 14. Filename: Right_Hippocampus_pub.nii.gz a. Short description: Manual segmentation of the right hippocampus. 15. Filename: Rostral_Commissure_pub.nii.gz a. Short description: Manual segmentation of the rostral commissure. 16. Filename: Thalamus_pub.nii.gz a. Short description: Manual segmentation of the thalamus 17. Filename: Atlas_pub.nii.gz a. Short description: All masks combined to an atlas with each structure given a grey value; amygdala = 1, Caudate Nuclei = 2, Corpus Callosum = 3, Fornix = 4, Globus Palidus = 5, Hypothalamus = 6, Left Hippocampus = 7, Medulla Olongata = 8, Mesencephalon = 9, Olfactory Bulb = 10, Optic Chiasm = 11, Pineal Gland = 12, Pons = 13, Right Hippocampus = 14, Rostral Commissure = 15 and Thalamus = 16 Tissue Probability Maps 1. Filename: template_prob_csf.nii.gz 2. Filename: template_prob_gm.nii.gz 3. Filename: template_prob_wm.nii.gz 4. Filename: template_seg_all.nii.gz 5. Filename: template_seg_csf.nii.gz 6. Filename: template_seg_gm.nii.gz 7. Filename: template_seg_wm.nii.gz Final Template 1. Filename: template_pub_nii.gz 2. Relationship between files: All segmentations were performed on and created from the Final template. 3. Are there multiple versions of the dataset? Yes Datafilesin the Zip folder, and this readme were updated 2019-06-28 to add a few additional files, rename a few files and update the tissue probability map. -------------------------- METHODOLOGICAL INFORMATION -------------------------- 1. Description of methods used for collection/generation of data: Brains were imaged ex-vivo and in situ within the cranium within 4 hours of euthanasia. Imaging was performed in a GE Discovery MR750 3.0T MRI scanner with the use of a 16 channel Flex NeoCoil that was wrapped around the dorsal aspect of the cranium. A T1-weighted 3-dimensional sequence (magnetization-prepared 180 degrees radio-frequency pulses and rapid gradient-echo (MP RAGE)) was performed with the following parameters; repetition time 7.364, echo time 3.468, inversion time 425, averages 1, matrix 256 x 256, spatial resolution 1mm3. 2. Methods for processing the data: Template Creation: The MRI data were corrected for low frequency intensity inhomogeneity (Tustison et al., 2010). A combined approach of automated (Smith, 2002) and manual removal of non-brain tissues was applied prior to the images being affine registered (Smith et al., 2004) and spatially normalized (Friston et al., 1995). The origin of the images was set to the anterior commissure, then the data was reoriented to a standard FMRI Software Library (FSL) orientation for consistency (Jenkinson et al., 2012). The data was then trimmed (removal of non-brain space from the image) and set to standard dimension for all images. The data was then transformed into a common space population template (MeanNonlinear) using Advanced Normalization Tools (ANTs) which uses diffeomorphism and affine transformation to produce a group mean size and shape (Avants et al., 2011). This template was generated with a stereotaxic coordinate system according to MNI template specifications and in line with other animal templates (Nitzsche et al., 2015). Tissue Probability Maps (TPMs): TPMs were created for each subject and template using FMRIB’s Automated Segmentation Tool (FAST) which segments brain matter into cerebral spinal fluid, grey matter, and white matter while correcting for spatial intensity variations (Zhang, Brady & Smith, 2001). FAST was used to create partial volume maps, TPMs of each tissue type, binary segmentation masks and bias field maps. The partial volume masks were used to calculate the tissue volume to account for partial volume effects and increase sensitivity. Anatomic Prior Creation: From the population template anatomically significant regions were manually delineated from histological references using mask creating software FSLeyes (Dewey and Da Costa; Furr and Reed). Segmented regions were only those that were discernable on the T1-weighted imaging sequences and included; olfactory bulbs, rostral commissure, caudate nuclei, Globus pallidus, thalamus, hypothalamus, optic chiasm, pineal gland, corpus callosum, fornix, hippocampi, amygdala, mesencephalon, pons, medulla oblongata and cerebellum. Gross whole brain and transverse and sagittal slices were photographed and correlated to the associated priors and atlas surface. 3. Instrument- or software-specific information needed to interpret the data: Data able to be visualized using multiple software types including FASLeyes and MRIcro. 4. Describe any quality-assurance procedures performed on the data: Template Quality Assessment: For quality assessment purposes templates using both linear rigid and affine registration were created for comparison. After the preprocessing procedure described above, a single subject was chosen at random to serve as a reference. For the creation of a rigid registration average FMRIB’s Linear Registration Tool (FLIRT) was used to register each subject to the reference image with six degrees of freedom (Jenkinson & Smith, 2001; Jenkinson et al., 2002). These registered images were concatenated and then divided by number of subjects to create a linear rigid mean template (MeanRigid). Similarly, FLIRT was used to create a linear affine mean template (MeanAffine) by registering each subject to the reference image with twelve degrees of freedom the concatenating and dividing in the same fashion (Jenkinson & Smith, 2001; Jenkinson et al., 2002). The standard deviation for each voxel was calculated from each of the templates MeanRigid, MeanAffine, and MeanNonlinear. The quality of each template and subject volumes were assessed subjectively and quantitively. Signal-to-noise (SNR) and contrast-to-noise (CNR) ratios were calculated using the widely accepted equations (Allen, Damasio, & Grabowski, 2002).