<p>Purpose: The purpose of this study was to investigate the optimal spatial resolution and temporal resolution of dynamic improved motion-sensitized driven-equilibrium steady-state free precession for visualization of respiratory-driven cerebrospinal fluid (CSF) dynamics. Methods: We investigated the differences in the visualization using the midsagittal cross-sections of nine healthy volunteers by three imaging conditions. (A: spatial resolution 0.49×0.49×5 mm, temporal resolution 1000 ms; B: 0.49×0.49×5 mm, 430 ms; and C: 0.78×0.78×5 mm, 200 ms). First, we calculated the CSF of the third and fourth ventricles and the signal-to-noise ratio (SNR) of the pons. Next, we calculated the signal intensity ratio (SIR) of the CSF flowing at 10 cm/s or more and the CSF flowing at 10 cm/s or less due to respiration. We also calculated the difference between the inspiration and expiration SIR. Furthermore, 1) the presence of flow in the third and fourth ventricles centered on the cerebral aqueduct and 2) the change in flow due to respiration was investigated by a three-point scale visual assessment by seven radiological technologists. Results: The SNR was the highest in A, the next highest in B, and the lowest in C in all cases. There were significant differences between A and B, and A and C in CSF of the third and fourth ventricles. However, there was no significant difference between B and C. The CSF signal intensity changed with respiration. The SIR of the third ventricle was higher on inspiration and lower on expiration. Conversely, the SIR of the fourth ventricle was lower on inspiration and higher on expiration. There was a significant difference between A and C and B and C in each SIR (p<0.05). The difference between inspiration and expiration SIR was the highest in B, the next highest in A, and the lowest in C in both the third and fourth ventricles. Significant differences were found between A and C, and between B and C (p<0.05). There was no significant difference in the presence of flow in the third and fourth ventricles centered on the cerebral aqueduct (p=0.264). On the other hand, there was a significant difference between the imaging conditions in the change in flow due to respiration, with B having a higher value than the others (p<0.001). Conclusion: The optimal spatial and temporal resolutions were 0.49×0.49×5 mm and 430 ms, respectively. The results also suggest that it is important to carefully set the imaging conditions for the spatial and temporal resolutions because of the use of phase dispersion in this method.</p>