2812

Inter-site, inter-subject and inter-session variability of B1+ and B0 in the human brain at 7 tesla.

William T Clarke¹, Olivier Mougin², Ian D Driver³, Catarina Rua⁴, Andrew T Morgan⁵, Susan Francis², Richard Bowtell², Richard Wise³, Adrian Carpenter⁴, Christopher T Rodgers⁴, Keith Muir⁵, and Stuart Clare¹

¹Wellcome Centre for Integrative Neuroimaging, FMRIB, NDCN, University of Oxford, Oxford, United Kingdom, ²Sir Peter Mansfield Imaging Centre, School of Physics and Astronomy, University of Nottingham, Nottingham, United Kingdom, ³Cardiff University Brain Research Imaging Centre, School of Psychology, Cardiff University, Cardiff, United Kingdom, ⁴Wolfson Brain Imaging Centre, Department of Clinical Neurosciences, University of Cambridge, Cambridge, United Kingdom, ⁵Imaging Centre of Excellence, University of Glasgow, Glasgow, United Kingdom

Synopsis

At ultra-high field strengths, e.g. 7 tesla, the homogeneity of the transmit B₁⁺-field and the B₀-field decreases compared to lower fields. This can strongly affect the contrast in images especially when imaging large volumes, such as the whole brain.

In this work we measure the mean and variance of B₁⁺- and B₀-fields in whole brain images acquired on four, whole-body human 7T scanners. We demonstrate that inter-subject differences dominate over inter-site differences in explaining the observed B₀- and B₁⁺-field variability. Inter-session differences are not significant.

Introduction

At ultra-high field strengths, e.g. 7 tesla, the homogeneity of the transmit B₁⁺-field and the B₀-field decreases compared to lower fields. In addition, the dependence of the magnitude of the transmit B₁⁺-field on the subject-specific coil-loading increases compared with lower field systems.¹ Both these factors affect the ability of scanners to effectively homogenize and calibrate the fields to a desired target.²

The contrast of many pulse sequences depends on the local B₁⁺- and B₀-fields, so this degradation compared to operation at lower fields can reduce or completely remove contrast from images.³ This is apparent when imaging whole organs (e.g. the brain), which experience significant inhomogeneity of B₁⁺ and B₀ across their volume.⁴

In this work we measure the means and variances of B₁⁺- and B₀-fields in whole brain images acquired on four whole-body human 7T scanners in order to characterise contributions to their variability.

Methods

The data in this work was collected as part of the UK7T Network’s “Travelling Heads Study”.⁵ Ten subjects (7 male, 69.8±12.3 kg, 32.0±5.9 years) underwent structural and functional neuroimaging scans at the five 7T-capable UK sites. In this work a subset of this data from Sites 1-4 is used. Each subject was scanned at each site, and a further four times at their designated home site (nine scans in total per subject). The hardware specifications for each site are given in Table 1.

At each site, as part of the scanning protocol and after applying the vendor’s transmit calibration process, the transmit B₁⁺-field was mapped using a 3D DREAM flip-angle mapping sequence.^6,7 Subsequently, after applying the vendor’s own automatic B₀ shimming process, the B₀-field was mapped using a dual-echo GRE sequence.

The resulting flip-angle and B₀-maps were masked by applying brain masks created by FSL BET from concurrently acquired MPRAGE images.⁸ The B₁⁺-field values were normalized to the target value and are therefore expressed as a unitless ratio. The mean and standard deviation (SD) of the masked B₁⁺- and B₀-fields, and the per-subject and per-site field distributions were plotted in Matlab.

Analysis of variance (ANOVA) was carried out separately on the mean and SD of each measured field for 3 factors: site, subject and session (repeated home site scans).

When implementing the structural and anatomical scans in the travelling heads study, the B₁⁺-field was recalibrated based on the mean value of the measured DREAM flip-angle map. This factor was reapplied to the DREAM maps in MATLAB to scale them to the final scan-time values, and the 3-way ANOVA analysis was repeated.

Results

Figure 1 shows, per-subject, the site-wise distributions of the voxel-wise normalized B₁⁺-fields. Figure 2 shows the same data plot subject-wise, per-site. Figure 3 shows, per-subject, the site-wise distributions of the voxel-wise normalized B₀-field. The results of the ANOVA tests are presented in Table 2.

The mean (± SD) B₀-field standard deviation across all scans was 53±12 Hz. The mean (± SD) B₁⁺-field mean across all scans was 0.75±0.06, after scaling this was 1.00±0.08. The mean (± SD) B₁⁺-field standard deviation across all scans was 0.17±0.02, after scaling this was 0.23±0.03.

Discussion

The ANOVA tests show significant variance (p<<0.05) due to site and subject differences on all measures except the means of the scaled B₁⁺-fields. Furthermore, the tests show that for both B₁⁺- and B₀-fields inter-subject differences account for more variance in means and standard deviations than inter-site (and inter-session) variance. Inter-session variance was not significant in all tests.

After rescaling the B₁⁺-fields so that the mean B₁⁺-field of each subject was 1.00 there was no significant variance of the mean explained by inter-subject, site or session effects. The scaling increased the mean standard deviation of the B₁⁺-field across all scans, as in all cases the scanner’s own calibration underestimated the mean value across the whole brain. ANOVA on the standard deviation of the scaled B₁⁺-field values retained significance for the effect of sites and subjects, though the variance explained (η_p²) by inter-site effects was decreased (from 0.172 to 0.062) to below variance explained by inter-subject effects (from 0.128 to 0.150).

Conclusion

The B₁⁺- and B₀-fields at 7 tesla in the human brain have been characterized in a multi-centre travelling heads study. Analysis of variance shows that inter-subject differences typically dominate the variance of both B₁⁺- and B₀-fields. After applying manual scaling of the transmit gain, inter-subject and inter-site variability of mean B₁⁺-fields are not significant. Inter-session differences at the same site and with the same subject are not statistically significant.

Acknowledgements

The UK7T Network and this work was funded by the UK's Medical Research Council (MRC). [MR/N008537/1]. The Wellcome Centre for Integrative Neuroimaging is supported by core funding from the Wellcome Trust (203139/Z/16/Z). CTR is funded by a Sir Henry Dale Fellowship from the Welcome Trust and the Royal Society [098436/Z/12/B].

We thank David Porter for supporting the UK7T Network in using the scanner and facilities at ICE, University of Glasgow.

References

1. Vaughan JT, Garwood M, Collins CM, Liu W, DelaBarre L, Adriany G, Andersen P, Merkle H, Goebel R, Smith MB, Ugurbil K. 7T vs. 4T: RF power, homogeneity, and signal-to-noise comparison in head images. Magn Reson Med 2001;46(1):24-30.

2. Clarke WT, Mougin O, Driver ID, Rua C, Francis S, Wise R, Carpenter A, Muir K, Bowtell RW, Clare S. Variability of B1+ and B0 fields in the human brain at 7T. Proc Intl Soc Mag Reson Med 26 2018:4199.

3. Balchandani P, Naidich TP. Ultra-High-Field MR Neuroimaging. AJNR Am J Neuroradiol 2015;36(7):1204-1215.

4. Zwanenburg JJ, Hendrikse J, Visser F, Takahara T, Luijten PR. Fluid attenuated inversion recovery (FLAIR) MRI at 7.0 Tesla: comparison with 1.5 and 3.0 Tesla. Eur Radiol 2010;20(4):915-922.

5. Clarke WT, Mougin O, Driver ID, Rua C, Morgan AT, Clare S, Francis S, Wise R, Carpenter A, Muir K, Bowtell RW. The UK7T Network – optimized design of a multi-site, multi-vendor travelling heads study. Proc Intl Soc Mag Reson Med 26 2018:317.

6. Brenner D, Tse DHY, Pracht ED, Feiweier T, Stirnberg R, Stöcker T. 3DREAM – A Three-Dimensional Variant of the DREAM Sequence. Proc Intl Soc Mag Reson Med 22. 2014; Milan, Italy. p 1455.

7.Nehrke K, Bornert P. DREAM--a novel approach for robust, ultrafast, multislice B(1) mapping. Magn Reson Med 2012;68(5):1517-1526.

8. Smith SM. Fast robust automated brain extraction. Hum Brain Mapp 2002;17(3):143-155.

Figures

Table 1. Scanners and hardware used in the UK7T Network travelling heads study. The Network comprises five sites, with three different models of scanners, from two different vendors. All scanning used the Nova Medical Inc. (Wilmington MA, USA) single-Tx-channel head coil. The coils were identical except for the coil-scanner interface. Data from Sites 1-4 was used in this work.

Figure 1. Site-wise, per-subject, box and whisker plots of voxel-wise B₁⁺-field distributions in whole-brain masked field maps. Four distributions are shown in each panel for each site. For all but two subjects, data from one site (red, wide) corresponds to an average of five repeat scan sessions (red, narrow). The median of each plot is indicated by a white target, the box indicates the 25^th and 75^th percentiles; the whiskers extend to the 1^st and 99^th percentile. The data shown here is the B₁⁺-field distributions normalised to the target after vendor calibration, but before manual scaling.

Figure 2. Subject-wise, per-site, box and whisker plots of voxel-wise B₁⁺-field distributions in whole-brain masked field maps. Ten distributions are shown in each panel for each subject. For each site, data for two subjects (red & blue, wide) corresponds to an average of five repeat scan sessions (red & blue, narrow). The median of each plot is indicated by a white target, the box indicates the 25^th and 75^th percentiles; the whiskers extend to the 1^st and 99^th percentile. The data shown here is the B₁⁺-field distributions normalized to the target after vendor calibration, but before manual scaling.

Figure 3. Site-wise, per-subject, box and whisker plot of voxel-wise B₀-field distributions in whole-brain masked field maps. Four distributions are shown in each panel for each site. For each subject one site (red, wide) is an average of five repeat scan sessions (red, narrow). The median of each plot is indicated by a white target, the box indicates the 25^th and 75^th percentiles; the whiskers extend to the 1^st and 99^th percentile.

Table 2. Results of the 3-way analysis of variance (ANOVA) tests. The effect size is calculated as ‘partial-eta-squared’ (η_p²). p-Values that are significant (p<0.05) are in bold font.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)

2812