The brain stem is an essential component in the central nervous system involved in the regulation of sleep-wake cycle, cardiac functions, and respiratory functions¹. DWI and tractography provide an ideal tool to understand changes in the brain stem structure associated with neurological disorders² since DWI and tractography can provide a comprehensive look into the neural connections of the brain in vivo and with little risk to subject³. One of the biggest challenges facing current methods of tractography is false positives resulting from noise and low spatial resolution of in vivo measurements. Thresholding tractography maps to limit the amount of false positives in the data set may be extremely useful in the diagnosis of neurological diseases. The goal of the protocol was to optimize the acquisition sequence for the brain stem, which can to be performed in less than 20 minutes, without sacrificing the quality of the data. Then limit the number of false positive through the application of a threshold filtering of the calculated streamline tracks. Eleven healthy control participants, between the ages of 20 and 40 years, were recruited with no prior history of neurological disorders. Participants were scanned using a 3.0 T Phillips system with a 32-channel head coil. A high-angular-resolution diffusion imaging (HARDI) sequence was acquired. A total 71 directions were obtained; 1 with b = 0, 6 with b = 100 s/mm² and 64 with b = 1000 s/mm² using a model electrostatic repulsion to determine the diffusion gradient directions⁴. The acquired image resolution was 1.7 x 1.7 x 1.7 mm³, TR/TE = 5000/86 ms, and Δ/δ = 42.1/9.1 ms. These images were then interpolated by a factor of 2 using cubic convolution with the CONGRID function in IDL [MH1] for a final image resolution of 0.85 mm isotropic. The fractional anisotropy (FA) and mean diffusivity (MD) were reconstructed from the HARDI data with in-house software written in IDL (Exelis Visual Information Systems, Boulder, CO). The thalamus and the pons were hand drawn using ITKSNAP using the FA and MD maps. The mixture of Wisharts (MOW) model was utilized for reconstruction of the fiber orientations⁵; this model characterizes the displacement probability function allowing the estimation of multiple fibers per voxel. MOW was calculated four separate times per data set each with a different probability displacement radius of: 13, 16, 18, and 20 microns. Utilizing the maxima of the displacement probability function as fiber orientations, deterministic tractography was performed with 125 seeds per voxel, fiber step size of one-half voxel, and no step-to-step angular deviations greater than 50°. The streamlines connecting the pons and thalamus ROIs was generated using in-house C based software. The streamlines connecting the 2 node-network (i.e. pons and thalamus) was converted into a tract density mask with each voxel containing an intensity value corresponding to the number of streamlines passing through that voxel. Finally a threshold was applied to the density maps for voxels containing less than 1%, 3%, 5%, and 10% of the maximum number of streamline tracts.Eleven healthy control participants were successfully scanned and the connectivity between the pons and thalamus was estimated with deterministic tractography. The total scan time was 16 minutes and the isotropic resolution of 1.7 x 1.7 x 1.7 mm³ represents a 38.5% reduction in voxel volume compared to traditional in vivo acquisition of 2 x 2 x 2 mm³ with comparable acquisition time. Figure 1 shows the effect of applying a threshold to the density map images. Without a threshold, a large number of voxels exhibit streamlines making to discrimination of any apparent structure difficult. As the threshold increases, small bundles emerge displaying a large number of streamlines. Figure 2 shows a structure that lies within the expected area associated as the locus coeruleus and represents the first time the locus coeruleus has been segmented in vivo with tractography. One of the greatest problems with tractography is the number of false positives that result from the inherent low spatial resolution in DWI data⁶. These false positives make data interpretation difficult. By optimizing the spatial resolution and thresholding the number of fibers in a voxel, we were able to eliminate extraneous fibers. With this approach, tractography can add in the study anatomical abnormalities due to neurological diseases, such as PTSD, Alzheimer’s, or Parkinson’s.