To construct and validate an atlas based template of the optic radiation. 34 healthy participants, 24 female and 9 male, mean age 37.18 (SD=14.62), were divided into two groups: Group A (19 subjects) for constructing optic radiation (OR) template and Group B (15 Subjects) for validating the OR template construction. MRI data was acquired for all participants from a 3.0T GE MR750 scanner (General Electric, Milwaukee, WI, USA), using an 8 channel head coil. For each exam, whole brain 64-directions diffusion weighted imaging was acquired with 2 mm isotropic acquisition matrix (TR/TE=8325/86 ms, b = 1000 s/mm2, number of b0s = 2); Additionally, Group A was acquired with Sagittal IR-FSPGR (TR/TE/TI =7.2/2.7/450 ms, 1 mm isotropic acquisition matrix, FOV = 256mm), and Group B was acquired with Axial IR-FSPGR (TR/TE/TI =8.2/3.2/900 ms, 1 mm isotropic acquisition matrix, FOV = 256mm). All diffusion weighted imaging (DWI) was corrected for motion, eddy-current and EPI susceptibility distortion, prior to the tensor reconstruction and T1 structure imaging co-registration in MrDiffusion package (MrVista, Stanford University, http://web.stanford.edu/group/vista/cgi-bin/wiki/index.php/MrVista). Seeding points at the lateral geniculate nucleus (LGN) and calcarine sulcus were used to reconstruct OR using probabilistic tractography with Contrack¹ (Figure 1). Bilateral OR was reconstructed for each participant, and constrained by individual white matter masks derived from T1 using SIENAX (FMRIB Software Library; www.fmrib.ox.ac.uk/fsl). ICBM 2009a nonlinear Asymmetric template (http://www.bic.mni.mcgill.ca/ServicesAtlases/ICBM152NLin2009) was used for mapping the final OR template. After non-brain tissue was removed (Brain Extraction Tool, FMRIB software Library), all T1 images were co-registered non-linearly to ICBM 2009a template using ANTS (Advanced Normalization Tools, http://picsl.upenn.edu/software/ants) to obtain the transformation maps. Individual fibre-based ORs were converted to binary ROI, mapped into ICBM space and averaged to construct the OR template. The initial OR template derived from group A was presented as a probability map. Group B were used to validate the OR template and determine the optimal probability threshold for obtaining the binarized OR template. Briefly, the OR probability map was thresholded from 10% to 90% with 1% increments; all thresholded OR templates were then mapped on to Group B individually using the inversed deformation maps, and compared with probabilistic tractography based individual ORs using the Dice Similarity Coefficient (DSC). Finally the threshold that produced the largest mean DSC in group B was used to obtain the final OR template. Reconstruction of the OR using probabilistic tractography was successful in all subjects, except for a unilateral OR from one subject whose imaging was degraded by imaging artefact. The OR template, constructed from Group A, is presented as a probability map in Figure 2a. Topographical conformity with a histologically determined atlas-based OR² was confirmed visually (Figure 2c). Meyer loop, in particular, was well defined. When the OR probability map was applied to a validating dataset (Group B), the highest mean DSC (SD), 0.703 (0.047) and 0.716 (0.067), corresponded with a probability threshold level of 40% and 44% for the left and right OR respectively (table 1). The OR template, determined by these thresholds, was compared to individual probabilistic tract-based OR reconstructions (Figure 3). While excellent correspondence between the two methods was found, a relative under-estimation of the Meyer loop component of the OR was noted in tract-based reconstructions (figure 4).Nonlinear registration of a tractography generated OR template facilitates accurate estimation of the tract in individual cases without DWI. In our study, a threshold of ~40% provides the best approximation of template-based OR to individually defined tract-based OR. In addition, template-based tractography provides better representation of Meyer loop in some subject; however, care should be taken in interpreting the size and position of the posterior part of the OR.We propose a new pipeline to generate and validate an OR template using probabilistic tractography. We have demonstrated topographic concordance with a histological atlas and robust estimation when compared to individual tractography-based reconstruction. The variance between template based and tractography based OR reconstructions is most apparent in the vicinity of the Meyer loops and posterior part of the OR. Future work should include validation of our work in a larger dataset derived from multiple sites.