Background

Functional Magnetic Resonance Imaging (fMRI) uses Echo Planar Imaging (EPI) to measure Blood Oxygen Level Dependent (BOLD) response in the brain to neural activity. The research applications are wide and varied, particularly for cognitive neuroscientists, however fMRI is also increasingly used in clinical practice for improved neurosurgical planning.
The acquisition of this data has been somewhat and modestly optimised over the years as field strengths increased, and RF and gradient technology improved. More recently Multiband (MB) MRI, also known as Simultaneous Multi Slice (SMS), has allowed for multiple slice locations to be excited simultaneously. The use of this technique in conjunction with EPI and multi-channel receive coil reconstruction techniques has led to a dramatic improvement in fMRI, including increased spatial (< 1mm isotropic) and temporal resolution (TR<1s), while maintaining whole brain coverage at 7T [1, 2]. It has also allowed for the introduction of Multi-Echo fMRI which can be a successful way to increase the BOLD signal and contrast in small sub-cortical brain structures.

Teaching Point

1. The addition of SMS to EPI means whole brain coverage can be achieved with sub millimetre slices and sub second temporal resolution for fMRI at 7T.

Whilst the brain is exposed to a task there are changes in the blood flow in capillaries of the cortex . BOLD imaging uses the susceptibility difference oxyhaemoglobin and deoxyhaemoglobin to quantify brain activation. Timing of the acquisition is vital in imaging these changes in blood flow and with this requirement comes the consideration of TR (which dictates temporal resolution), voxel size (spatial resolution) and TE (contrast) [3, 4]. Correct TE selection is vital as the BOLD effect relies on T2* weighting differences between deoxyhaemoglobin and oxyhaemoglobin. This T2* effect increases with the increase in magnetic field leading to significant increases in BOLD contrast at ultra-high fields (7T or higher) [5] (Fig 1). This together with the increase in MRI signal to noise ratio (SNR) at 7T means there is the possibility to explore improved spatial and temporal resolution to sub millimetre, sub second (Fig 2). However, with thinner slices, associated with higher resolution images, whole brain coverage becomes a major issue. The introduction of multiband has made more coverage possible.
The Good- Multiband imaging uses composite RF pulses to simultaneously excite multiple slice locations through the use of slice selective gradients [6]. The MB factor determines how many slices are acquired within one TR. A high MB factor can increase the coverage or inversely allow a reduced TR to improve temporal resolution [7]. One advantage is that because thinner slices can be achieved there is less through slice dephasing which appears as signal dropout in the frontal and hippocampus areas using traditional fMRI [8] (Fig 2). These advantages of whole brain coverage with higher spatial and temporal resolution through the acquisition with multiband can be seen in a comparison of whole brain imaging at 3T and 7T (Fig 3).
The Bad – Using multiband doesn’t come without challenges. Both artefacts and significantly reduced SNR can be produced if the MB factor is too high [9]. This occurs when the simultaneously excited slices become too close together and the receive coil array fails to adequately separate the signals. When stacked slices are viewed from the perpendicular plane one sees significant slice banding (Fig 5a).

2. Adding multiple echoes to the MB-EPI sequence further increases fMRI signal and contrast.

The Future - As we move forward with fMRI and the improvements of BOLD imaging with MB we have seen the introduction of Multi Echo fMRI. Recently ME fMRI has moved from a technique being itself researched to both a research and potentially clinical technique. It involves the acquisition of multiple echoes within one TR on simultaneously excited slices [7]. This technique has potential to dramatically improve fMRI for two reasons: 1. The sum of the echoes increases SNR. 2. Optimal BOLD contrast is when the TE = T2*, therefore multiple TE’s will increase BOLD contrast [10]. Since different gray matter structures have a different T2* (particularly sub cortical) there are more chances of capturing the highest BOLD contrast from as many tissues as possible (Fig 4). The combination of the signals across the multiple TEs can then produce a higher sensitivity than a single TE fMRI [10]. This is supported in cases where there is signal dropout from susceptibility artefact in one TE but not in another. (Fig 5b.) It can also mean activation in subcortical layers can be separated by the different activation times for increased specificity [11].

Summary

The addition of MB to EPI means that higher temporal and spatially resolved fMRI can be realised without compromise of spatial coverage. In addition, combining this with a multi-echo readout can further increase the BOLD contrast, particularly for sub-cortical grey matter structures, increasing the sensitivity and specificity. This can be of great assistance to fMRI and the advancement to applied neuroscience and clinical neuroscience research. In principle, the MB technique can work for any slice selective sequences and as such the technology could be used to decrease slice thicknesses without compromise of coverage for a whole range of applications.

Acknowledgements

We acknowledge the facilities, and the scientific and technical assistance of the Australian National Imaging Facility, a National Collaborative Research Infrastructure Strategy (NCRIS) capability, at the Melbourne Brain Centre Imaging Unit of the University of Melbourne. The work was also supported by a research collaboration agreement with Siemens Healthineers. Acknowledgment also to Centre for Magnetic Resonance Research, Minnesota for the development and supply of the Multiband EPI sequence required for this project.