Phase vs. magnitude information in functional magnetic resonance imaging time series: toward understanding the noise
Phase vs. magnitude information in functional magnetic resonance imaging time series: toward understanding the noise
复制标题
DOI:
10.1016/j.mri.2009.02.006
复制
发表时间:
2009-10-01
影响因子:
2.5
通讯作者:
Bowtell, Richard
中科院分区:
文献类型:
--
作者:
Petridou, Natalia;Schaefer, Andreas;Bowtell, Richard
Although it has been shown that the phase of the MR signal from the brain is particularly prone to variation due to respiration, the overall physiological information contained in phase time series is not well understood. Here. we explore the different physiological processes contributing to the phase time series noise, identify their spatiotemporal characteristics and examine their relationship to BOLD-related and non-BOLD-related physiological noise in the magnitude time series. This was performed by manipulating the contribution of physiological noise to the total signal variance by modulating the TE and voxel volume, and using a short TR in order to adequately sample physiological signal fluctuations. The phase and magnitude signals were compared both before and after removal of signal fluctuations at the primary respiratory and cardiac frequencies with RETROICOR. We found that the temporal phase noise increased with TE at a faster rate than predicted by I/TSNR as a result of physiological noise. As suggested by previous studies, the primary contributor to phase physiological noise was respiration-related effects which were manifested at a large scale (>1 cm). Notably, RETROICOR removed respiration-related large-scale artifacts and this resulted in considerable improvements in the temporal phase stability (7-90%). Physiological noise in the magnitude time series after RETROICOR consisted of low-frequency BOLD-related fluctuations (