Characterization of Nyquist ghost in EPI-fMRI acquisition sequences implemented on two clinical 1.5 T MR scanner systems: effect of readout bandwidth and echo spacing

Abstract

In EPI-fMRI acquisitions, various readout bandwidth (BW) values are used as a function of gradients' characteristics of the MR scanner system. Echo spacing (ES) is another fundamental parameter of EPI-fMRI sequences, but the employed ES value is not usually reported in fMRI studies. Nyquist ghost is a typical EPI artifact that can degrade the overall quality of fMRI time series. In this work, the authors assessed the basic effect of BW and ES for two clinical 1.5 T MR scanner systems (scanner-A, scanner-B) on Nyquist ghost of gradient-echo EPI-fMRI sequences. BW range was: scanner-A, 1953-3906 Hz/pixel; scanner-B, 1220-2894 Hz/pixel. ES range was: scanner-A, scanner-B: 0.75-1.33 ms. The ghost-to-signal ratio of time series acquisition (GSRts) and drift of ghost-to-signal ratio (DRGSR) were measured in a water phantom. For both scanner-A (93% of variation) and scanner-B (102% of variation) the mean GSRts significantly increased with increasing BW. GSRts values of scanner-A did not significantly depended on ES. On the other hand, GSRts values of scanner-B significantly varied with ES, showing a downward trend (81% of variation) with increasing ES. In addition, a GSRts spike point at ES = 1.05 ms indicating a potential resonant effect was revealed. For both scanners, no significant effect of ES on DRGSR was revealed. DRGSR values of scanner-B did not significantly vary with BW, whereas DRGSR values of scanner-A significantly depended on BW showing an upward trend from negative to positive values with increasing BW. GSRts and DRGSR can significantly vary with BW and ES, and the specific pattern of variation may depend on gradients performances, EPI sequence calibrations and functional design of radiofrequency coil. Thus, each MR scanner system should be separately characterized. In general, the employment of low BW values seems to reduce the intensity and temporal variation of Nyquist ghost in EPI-fMRI time series. On the other hand, the use of minimum ES value might not be entirely advantageous when the MR scanner is characterized by gradients with low performances and suboptimal EPI sequence calibration.

Figure 1

Graphical diagram of gradient‐echo EPI‐fMRI acquisition sequence. The echo spacing represents the interval between successive echoes.

Figure 2

Set up of water phantom acquisitions (a); measurements of ${\mathit{\text{GSR}}}_{\mathit{\text{ts}}}$ and ${\mathit{\text{DR}}}_{\mathit{\text{GSR}}}$ were performed in slice 0 which is the median slice of the acquired phantom volume. Scanner‐A (b): images with $\text{BW}=1953\text{Hz}/\text{pixel}$ and $\text{ES}=0.9\text{ms}$. Scanner‐B (c): images with $\text{BW}=1906\text{Hz}/\text{pixel}$ and $\text{ES}=0.9\text{ms}$. The ROIs (b, c) used for the measurement of signal intensity within the phantom $({\mathit{\text{SI}}}_i,\mathrm{i}=1,2,3,4)$ (solid line), and the ROIs used for the measurement of signal intensity of Nyquist ghost $({\mathit{\text{NI}}}_i,\mathrm{i}=1,2,3,4)$ (dashed line) and background (BI) (dotted line), are shown. The ROIs employed for the measurement of ${\mathit{\text{NI}}}_i(\mathrm{i}=1,2,3,4)$ are shifted by N/2 voxels (64/2) along the phase encoding direction in respect to the corresponding ROIs employed for the measurement of ${\mathit{\text{SI}}}_i(\mathrm{i}=1,2,3,4)$.

Figure 3

Scanner‐A (solid line): ${\mathit{\text{GSR}}}_{\mathit{\text{ts}}}$ (a) and ${\mathit{\text{DR}}}_{\mathit{\text{GSR}}}$ (b) as a function of BW at fixed $\text{ES}=0.9\text{ms}$. Scanner‐B (dotted line): ${\mathit{\text{GSR}}}_{\mathit{\text{ts}}}$ (a) and ${\mathit{\text{DR}}}_{\mathit{\text{GSR}}}$ (b) as a function of BW at fixed $\text{ES}=0.9\text{ms}$. Graphs report the mean and standard deviation of the three measurements performed for each BW value.

Figure 4

Scanner‐A (solid line): ${\mathit{\text{GSR}}}_{\mathit{\text{ts}}}$ (a) and ${\mathit{\text{DR}}}_{\mathit{\text{GSR}}}$ (b) as a function of ES at fixed $\text{BW}=1953\text{Hz}/\text{pixel}$. Scanner‐B (dotted line): ${\mathit{\text{GSR}}}_{\mathit{\text{ts}}}$ (a) and ${\mathit{\text{DR}}}_{\mathit{\text{GSR}}}$ (b) as a function of BW at fixed $\text{BW}=1502\text{Hz}/\text{pixel}$. Graphs report the mean and standard deviation of the three measurements performed for each ES value.

Figure 5

Temporal variation of ghost‐to‐signal ratio values during EPI‐fMRI acquisitions superimposed with the second order polynomial trend fitted to time series data points: (a) scanner‐A, $\text{BW}=3125\text{Hz}/\text{pixel},\text{ES}=0.9\text{ms}$; (b) scanner‐B, $\text{BW}=1502\text{Hz}/\text{pixel},\text{ES}=0.75\text{ms}$; (c) scanner‐B, $\text{BW}=1502\text{Hz}/\text{pixel},\text{ES}=1.05\text{ms}$; (d) scanner‐A, $\text{BW}=1953\text{Hz}/\text{pixel},\text{ES}=0.75\text{ms}$.