Investigation of BOLD contrast in fMRI using multi‐shot EPI

Abstract

https://onlinelibrary.wiley.com/doi/abs/10.1002/(SICI)1099-1492(199706/08)10:4/5%3C179::AID-NBM463%3E3.0.CO;2-X Menon, R.S., Thomas, C.G. and Gati, J.S. (1997), Investigation of BOLD contrast in fMRI using multi-shot EPI. NMR Biomed., 10: 179– 182. https://doi.org/10.1002/(SICI)1099-1492(199706/08)10:4/5<179::AID-NBM463>3.0.CO;2-X The following text in the Results and Discussion section: This procedure has the effect of linearizing the plots, effectively showing what the contrast sensitivity would be if the same number of images were acquired for each trial. Under these conditions, it becomes quite clear that the contrast sensitivity is still strongly dependent on TR. Thus in doing a multi-slice fMRI experiment with EPI, it is better to acquire all the segments for a single slice with the shortest possible TR before going on to the next slice. Note that this is exactly opposite to the strategy that one would use in anatomic imaging, where interleaving the segments and slices would yield superior SNR. Examination of the points for etl = 32 shows that the biggest drop in normalized contrast sensitivity occurs when the total acquisition time of the image exceeds 0.5 s. This is likely because the physiological sources of motion that are effectively frozen at the shorter TR values now start to modulate the intensity of the individual segments. Is replaced by the following: This procedure has the effect of showing what the contrast sensitivity would be if the same number of images were acquired for each trial. Under these conditions, it becomes quite clear that the contrast sensitivity is still strongly dependent on TR. Thus in doing a multi-slice fMRI experiment with EPI, it is better to interleave the segments and slices to yield superior SNR, as would be done in anatomic imaging. Examination of the points for etl = 32 and etl = 16 show that greater CSBOLD is realized at longer TR times which allow for higher flip angles resulting in higher SNR. Use of a short TR allows for collection of a larger number of images over the same period of total imaging time, hence yielding the best statistical detection thresholds, but the cost is paid in decreased SNR. Physiological sources of motion become more important to modulation of the image intensity over time. The following text in the Conclusion section: For a fixed number of slices, the best approach is to perform the acquisition of all the segments in a single slice before continuing on to the rest of the slices. Is replaced by the following: For a fixed number of slices, the best approach is to interleave the acquisition of the segments and slice across the entire imaging volume. The final sentence of the Conclusion should be supplemented with: It is important to note that no navigator echo was used in the data acquisition of this experiment, consequently these results do not necessarily apply if navigator echo correction was implemented.