Background Parallel imaging reconstruction with non-Cartesian k-space trajectories is non-trivial. Image domain reconstruction has been demonstrated for arbitrary k-space trajectories by Pruessmann using iterative reconstruction. [1] More recently, preliminary work has demonstrated direct reconstruction in the k-space for projection and spiral trajectories using segmented GRAPPA. [2-5] Segmented GRAPPA relies on local regularities in the sampling pattern to perform a local interpolation in the k-space. This work demonstrates how parallel imaging can be performed with spiral trajectory using a dual-density approach, in a manner analogous to Cartesian partially parallel acquisition such as GRAPPA. The dual-density approach provides self-calibration, making the reconstruction resistant to changes in coil sensitivity, and results in increased SNR from low k-space over-sampling. Methods Fig.1 Rate 4 Segmented GRAPPA Reconstruction (click see large version) Dual-density spiral [Fig. 1] data are acquired on a 3T Siemens Trio scanner using an 8 channel head coil. The spiral trajectory (128x128) is designed with four segments and a 24x24 navigator region. In the reconstruction the missing spiral k-space points are estimated from two nearest radial neighbors using linear interpolation. The interpolation coefficients are assumed to be invariant within an angular sector.

Interpolation coefficients are calibrated in the high-density, low k-space data and applied to interpolate missing data in the high k-space. The method is demonstrated in a human volunteer using BOLD fMRI and a simple block design visual-motor task. Results In vivo imaging results see Fig. 2. Fig. 2 Bold fMRI results see Fig. 3. Fig. 3 Discussion Dual-density spiral is an effective tool for self-calibrated partially parallel imaging. The dual-density spiral maintains a radial sampling regularity which allows segmented GRAPPA reconstruction and avoids the need for iterative reconstruction. Self-calibrated, partially parallel acquisitions are well suited for dynamic studies such as BOLD fMRI and cardiac imaging where the sensitivity information may be altered by motion or difficult to obtain. Acknowledgment This work was supported in part by the National Institutes of Health, Georgia Research Alliance and the Whitaker Foundation. In addition the authors would like to acknowledge Dr. Yasser Kadah and Dr. Shantanu Sarkar for their contributions to this work.

References

1. Pruessmann KP, Weiger M, B?rnert P, Boesiger P. Advances in sensitivity encoding with arbitrary k-space trajectories. Magn Reson Med 2001;46(4):638~51.
2. Griswold MA, Heidemann RM, Jakob PM. Direct parallel imaging reconstruction of radially sampled data using GRAPPA with relative shifts. In: Proceedings of the 11th Annual Meeting of ISMRM, Toronto, 2003. p 2349
3. Heberlein K, Hu X. Segmented spiral parallel imaging using GRAPPA. In: Proceedings of the 12th Annual Meeting of ISMRM, Kyoto, 2004. p 328
4. Heidemann RM, Griswold MA, Kr¨ąger G, Kannengie?er S, Kiefer B, Jakob PM. Fast parallel image reconstructions for spiral trajectories. In: Second International Workshop on Parallel MRI, Zurich, 2004. p 27
5. Heberlein K, Hu X. Auto-calibrated parallel imaging using dual-density spirals. In: Second International Workshop on Parallel MRI, Zurich, 2004. p 59