3D multi-shot EPI for high resolution fMRI

Thermal noise is a major limitation of high resolution fMRI (∼1.5mm³) to study functional microstructure. Multi-shot 3D EPI techniques yield larger image signal-to-noise (SNR) values than standard 2D EPI methods but are also more exposed to physiological instabilities. The relative contribution of thermal and physiological noise to fMRI instabilities depends on factors such as image resolution, field strength…(1). Here we developed and optimized 2D EPI and multi-shot 3D EPI MR sequences. The potential of both methods for high resolution fMRI (∼1.5mm³)fMRI at 3T was assessed using measures of temporal stability (tSNR) and statistical significance in an fMRI experiment conducted using a visual task (left and right hemifield flickering checkerboard interspersed with rest periods) (2, 3). The analysis was conducted on a group of 6 subjects.

 

Timing of the physiological regressors

The full physiological model described in (4) was used to extract a set of 20 regressors including cardiac and respiratory phase (5), changes in respiration and hear rate (6-8) and motion parameters (9). These regressors were included in the tSNR calculations and the fMRI analysis. Figure 1 shows the increase in tSNR obtained from the 3D EPI sequence due to physiological correction when different partition numbers are used to sample the physiological regressors (reference partitions). The highest tSNR increase takes place for the central partition, consistent with the linear relationship between physiological noise and signal intensity (10).

Fig 1

Figure 1. Increase in tSNR due to physiological correction using different reference partition (3D EPI sequence).

 

Effect of physiological correction on 2D and 3D EPI

Figure 2a shows tSNR values before and after physiological correction for 2D and 3D EPI in Grey Matter (GM), Visual Cortex (VC) and Lateral Geniculate Nucleus (LGN). Physiological correction yielded an average increase in tSNR of 12, 13 and 6% for 2D EPI and 19, 18 and 11% for 3D EPI. ANOVA of the tSNR values yielded significant main effects of the sequence type (3D/2D), physiological correction (on/off) and scanned region (GM, VC and LGN) with all p<0.001 (all F>276). All two-ways and the three-way interactions between sequence type, physiological correction and region were significant (p<0.05 and F>4.2). After physiological correction, tSNR values from 3D EPI were 128% and 164% of the tSNR values from 2D EPI in VC and LGN respectively. Figure 2b shows the average t-score due to visual stimulation calculated from the 5% most highly activated voxels in the VC and LGN ROIs. 3D EPI yielded t-scores higher by 27% and 28% in VC and LGN respectively after physiological correction.

 

Fig 2

Figure 2. tSNR before and after physiological correction for 2D EPI and multi-shot 3D EPI in Grey Matter (GM), Visual Cortex (VC) and Lateral Geniculate Nucleus (LGN) (a). Average t-score due to visual activation calculated from the 5% most highly activated voxels in VC and LGN.

 

Comparison of 2D EPI and 3D EPI

Figure 3 shows group-averaged maps of increase in tSNR of 3D EPI over 2D EPI before (a) and after (b) physiological correction. Larger tSNR values were obtained using 3D EPI except for brain regions above the Circle of Willis due to pulsating cardiac effect along the (slow) partition direction. Figure 3 (c-f) shows representative activation maps following visual stimulation for 3D EPI (c, d) and 2D EPI (e, f) before (c, e) and after (d, f) physiological correction. Larger effects of physiological correction take place for 3D EPI, leading to improved detection of visual activation.

 

Fig 3

Figure 3. Group-averaged maps of increase in tSNR of 3D EPI over 2EPI before (a) and after (b) physiological correction. Representative activation maps following visual stimulation obtained using 3D EPI (c and d) and 2D EPI (e and f) before (c and e) and after (d and f) physiological correction.

 

Work in progress

Current developments include:

  • Use of water-selective RF pulses.
  • Development of a technique for online reconstruction of fMRI time-series acquired with parallel imaging.
  • Multi-echo 3D EPI acquisitions.
  • Ultra-fast 3D EPI acquisitions (volume TR<1s).

 

Primary contact

Nikolaus Weiskopf (n.weiskopf «at» ucl.ac.uk)

 

References

[1]        Triantafyllou C, Hoge RD, Krueger G, Wiggins CJ, Potthast A, Wiggins GC, Wald LL. Comparison of physiological noise at 1.5 T, 3 T and 7 T and optimization of fMRI acquisition parameters. Neuroimage 2005;26(1):243-250.

[2]       Lutti A, Thomas DL, Hutton C, Weiskopf N. High resolution functional MRI at 3T: 3D/2D/ echo-planar imaging (EPI) with optimized physiological noise correction. Magnetic Resonance in Medicine. In press.

[3]       Lutti A, Josephs O, Thomas DL, Lawson R, Roiser J.P., Hutton C, Weiskopf N. Optimized physiological noise corrrection for 3D EPI time series. Proceedings of the 19th Annual Meeting of ISMRM 2011;3635.

[4]       Hutton C, Josephs O, Stadler J, Featherstone E, Reid A, Speck O, Bernarding J, Weiskopf N. The impact of physiological noise correction on fMRI at 7 T. Neuroimage 2011;57(1):101-112.

[5]       Glover GH, Li TQ, Ress D. Image-based method for retrospective correction of physiological motion effects in fMRI: RETROICOR. Magn Reson Med 2000;44(1):162-167.

[6]       Birn RM, Diamond JB, Smith MA, Bandettini PA. Separating respiratory-variation-related fluctuations from neuronal-activity-related fluctuations in fMRI. Neuroimage 2006;31(4):1536-1548.

[6]       Shmueli K, van GP, de Zwart JA, Horovitz SG, Fukunaga M, Jansma JM, Duyn JH. Low-frequency fluctuations in the cardiac rate as a source of variance in the resting-state fMRI BOLD signal. Neuroimage 2007;38(2):306-320.

[6]       Chang C, Glover GH. Effects of model-based physiological noise correction on default mode network anti-correlations and correlations. Neuroimage 2009;47(4):1448-1459.

[6]       Friston KJ, Williams S, Howard R, Frackowiak RSJ, Turner R. Movement-related effects in fMRI time-series. Magnetic Resonance in Medicine 1996;35(3):346-355.

[6]       Kruger G, Glover GH. Physiological noise in oxygenation-sensitive magnetic resonance imaging. Magn Reson Med 2001;46(4):631-637.

back to physics pages