Lipid plays an important role in brain white matter function, particularly for myelin, and quantitative assessment of myelin with MRI has many applications. However, the T2* of MR signals from the lipid components of myelin are shorter than 1 ms (1), making it very difficult to detect myelin proton signals. Recently, Ultra Short TE (UTE) has been applied to explore the detection of myelin (1, 2). Based on our earlier study of short T2 imaging (3), we developed a new method, which takes advantage of both Zero TE (ZTE) and UTE, to positively visualize myelin protons from mouse brain in vivo.Cholesterol, which constitutes approximately 40 Mol % of myelin lipid and 30 Mol % of grey matter lipid (4), was selected to demonstrate feasibility of ZUS for direct myelin proton detection. Powdered cholesterol (Sigma, St. Louis, MO) was placed into a plastic tube of 1.5 cm diameter to form column with density of 1.83 mMol cm^-3 (Cholesterol tube). A saline filled plastic tube of 1.5 cm diameter (Saline tube) was attached to the Cholesterol tube by Teflon tape (Phantom, Fig.1A). Mice (C3H/HeNCrl, Charles River Laboratories, Wilmington, MA) were used as subjects for in vivo MRI. Animals were anesthetized by inhalation of isoflurane through a nose cone for the duration of the study.The major differences between ZTE and UTE are the timing of the excitation pulses and sampling of FID data. With ZTE, the projection gradient is ramped up and stabilized. Subsequently a short, hard RF pulse is issued with the gradient on. After a receiver dead time delay, the FID is sampled under the constant gradient. With UTE, a non-selective or selective pulse is issued. After a short time delay (TE), the projection gradient is ramped up, and the FID is sampled under the rising gradient ramp and the following constant gradient. With ZTE, signals with T2* as short as a few hundred microseconds, limited by the receiver dead time and signal acquisition time, can be detected. With UTE, signals with T2* as short as those detected by ZTE are not detected, but UTE selectively excites signals within a limited range of resonance frequencies and with longer T2* values. Subtraction of UTE from ZTE thus provides an image from signals with only the shortest T2* values detected by ZTE. MR imaging was performed on a Bruker BioSpec 70/30 scanner using a FOV = 35 mm and a digital resolution = 0.137 mm for the animal study and a 50 mm FOV and 0.195 mm digital resolution for the phantom study. ZTE was acquired with a bandwidth = 250 kHz, excitation pulse = 5 µs (5°), number of average = 1 and TR = 10 ms. UTE was acquired with bandwidth = 250 kHz, TE = 9 µs, excitation pulse 5 µs (5°) (UTE1) or with bandwidth = 100 kHz, TE = 1.5 ms and excitation pulse 1 ms (5°) (UTE2). Scan time for either ZTE or UTE was about 35 min.The ZTE image of the phantom shows both cholesterol and saline (Fig. 1B). Both UTE1 and UTE2 show saline only (UTE2 shown in Fig. 1C). The subtraction shows cholesterol only (Fig. 1D). ZTE, UTE2, and ZUS images of the brain of a live mouse are provided in Fig. 2 A, B, and C, respectively. ZUS allowed visualization of the skull (arrow) and brain. Note that the white matter of the corpus callosum (arrowhead) has higher lipid signal than surrounding grey matter. ZUS images may be optimized to visualize signals with a limited range of short T2* values by adjusting the pulse length of UTE, which excites the unwanted signals to be eliminated in ZUS. UTE2 was selected in this study, eliminating T2* signals longer than 500 μs in the ZUS image. T1 weighting was minimized by the use of a 5° excitation angle and a TR of 10 ms. The direct visualization of only solid state signals from powdered cholesterol sample and the higher signal from the mouse corpus callosum in vivo in the ZUS images support our contention that ZUS provides information from myelin protons with possibly some contamination from water associated myelin in vivo. The intensity of solid skull bone also visualized in ZUS may serve as an index for quantitative assessment of lipids in the brain.