posted on 2023-12-12, 04:40authored byAlaaddin Ibrahimy, Jocelyn Hoye, Hao Wu, Bart de Laat, Su Jin Kim, David L. Wilson, Evan D. Morris
Additional file 1. Figure S1. AICc images for phantom data showing 1- and 2-parameter selection for both voxel-level and cluster-level in coronal, axial, and sagittal views. Figure S2. Occmax images for phantom data showing the maximum occupancy estimated for both voxel-level and cluster-level in coronal, axial, and sagittal views. Figure S3. AICc images for human occupancy data showing 1- and 2-parameter selection for both voxel-level and cluster-level in coronal, axial, and sagittal views. Figure S4. Occmax images for human occupancy data showing the maximum occupancy estimated for both voxel-level and cluster-level in coronal, axial, and sagittal views. Figure S5. a Precision and b accuracy (Eq. 10) of EC50 for a background cluster region with true EC50 = 10 ng/mL using different m and K combinations. K is the number of initial clusters, and m is weighting coefficient spatial distance over the temporal distance. Vertical dash line represents the selected value of m for the reported case. Note: m and K were selected for analysis of both phantom and human data based on similar calculations for multiple regions on the phantom data. Figure S6. a Precision and b accuracy (Eq. 10) of EC50 for putamen region with true EC50 = 50 ng/mL using different m and K combinations. K is the number of initial clusters, and m is weighting coefficient spatial distance over the temporal distance.Vertical dash line represents the selected value of m for the reported case. Note: m and K were selected for analysis of both phantom and human data based on similar calculations for multiple regions on the phantom data.
Funding
National Institute of Biomedical Imaging and Bioengineering