Levels of choline containing compounds (Cho) have been shown to be relevant in the study of brain disorders and brain development. A number of studies have been conducted on rat or mouse models of disease at high field strengths. The Cho peak at about 3.22 ppm in rat and mouse brain spectra contains significant signal contamination from the taurine (Tau) resonance at about 3.25 ppm, even at the high field strength of 9.4 T (1,2). Quantification of Cho has been performed by either relying on spectral fitting software such as LCModel (3,4) or by subtracting the area of the other Tau resonance at ≈ 3.42 ppm from the combined Cho and Tau peak (5,6). The latter method may suffer from subtraction errors and uncertainties from making the assumption that the two Tau resonances are identical in the acquired spectrum. The choline protons are uncoupled; however, those of Tau are involved in J-coupling interactions with each other. The purpose of this work is to investigate the response of the J-coupled Tau protons to a PRESS (Point RESolved Spectroscopy) sequence at 9.4 T as a function of the two echo times, TE₁ and TE₂, to find a TE combination that minimizes Tau signal in the Cho spectral region in order to improve Cho quantification. Experiments were performed with a 21.5 cm diameter MRI 9.4 Tesla scanner. A birdcage radiofrequency (RF) coil was used for phantom experiments and a surface coil was employed for the animal scans. Phantom spectra were acquired from a 10 mM creatine (Cr)/50 mM Tau solution with the shortest TE combination achievable, namely, {TE₁, TE₂} = {12 ms, 9 ms} and with TE₁ and TE₂ values ranging from 15 ms to 100 ms in steps of 5 ms. A TE combination which minimized the Tau signal area in the 3.15 – 3.34 ppm region was selected as the optimal TE combination. Spectra were also obtained from a 10 mM Cr/ 50 mM Tau/ 5 mM Cho solution with the short-TE in addition to with the determined optimal {TE₁, TE₂} combination. Spectra were acquired in 32 averages with a repetition time, TR, of 5 s from 5 × 5 × 5 mm³ voxels placed in the centre of the phantom. A total of 8192 complex data points sampled at 10,000 Hz were collected. In-vivo spectra were acquired from the brains of three Sprague Dawley rats from a 5 × 5 × 5 mm³ voxel. A short-TE and an optimal TE spectrum were obtained from each rat. Rat spectra were measured as 2048 complex data points in 128 averages with a TR of 3 s and LCModel software was used for analysis. An optimal TE combination that minimizes Tau signal as a result of J-coupling evolution was found to be {TE₁, TE₂} = {25 ms, 50 ms}. Figure 1 shows short-TE and optimized-TE spectra acquired from the Tau phantom. The area of the optimized-TE Tau peak is about - 6.7 % of that obtained with the short-TE. Figure 2 demonstrates the efficacy of the optimized sequence on the phantom containing 5 mM Cho. The Tau signal is significantly reduced revealing the complete Cho resonance. Figure 3 shows short-TE and optimized-TE spectra obtained from one of the rats. The Cho and Tau resonances are not resolved in the short-TE spectrum; however, in the optimized-TE spectrum there is negligible Tau signal (low signal to noise of Tau resulted in a concentration estimate of 0 with a CRLB of 999 %). Similar spectra and LCModel results were obtained for all the three rats. For each rat, the Cho CRLB was 3-4% for both the short-TE and optimized-TE spectra. Assuming a T₂ of 178 ms for Cho (7), the T₂-corrected Cho concentration obtained with the optimized-TE is about 78 % of that obtained with the short-TE. The difference indicates that the presence of Tau signal at short-TE affects Cho signal quantification. It has been demonstrated that a PRESS sequence with {TE₁, TE₂} equal to {25 ms, 50 ms} is suitable for resolving the Cho signal (≈ 3.22 ppm) from overlapping Tau signal in vivo at 9.4 T. Further work will be conducted to quantify the improvement in Cho quantification as a result of Tau signal suppression.