As part of the Imperial festival, myself and colleagues presented some our of work relating to fetal MRI.
MRI imaging is a non invasive imaging technique that uses the water content of tissues as contrast. This makes it a very safe imaging modality with exceptional resolution, and wide flexibility to also image, blood flow, brain function and tissue microstructure.
Currently, fetuses are most commonly imaged using ultrasound. This has the advantage of being realtime, but is noisy and therefore difficult to interpret visually. Fetal MRI, on the other hand has the potential to generate very sharp images of all the major organs, allowing us to build very sensitive models of healthy development, which could potentially be fed back to improve and support ultrasound use in the clinic.
One limitation of MRI imaging is that it takes some time to collect and image, and therefore it is common practice for subjects to be asked to lie still for some considerable time during an MRI scan. As this is not possible for fetuses and unsedated neonates we have been required to come up with sophisticated scanning protocols which take many sharp 2D snapshots of the fetus, in different profiles and cross sections. These are then pieced together, correcting for motion, utilising graphics processing techniques that allow radiographers to recover the corrected images almost instantaneously (http://www.imperial.ac.uk/people/b.kainz, https://spiral.imperial.ac.uk:8443/handle/10044/1/21436).
This is the first step towards our goal of collecting data on brain function and microstructure in the developing fetus. During the Human Connectome Project (http://www.humanconnectome.org/) we were able to collect high quality data representing the functional and structural ‘connectivity’ of the adult brain. For example this shows roughly 15 minutes of brain activation (sped up considerably!) on the cortical surface. Here yellow represents active, or ‘firing’ tissues.
As MRI imaging uses water as a contrast these are not direct measurements of neuronal firing or fiber structure, rather these measure the flow of oxygenated blood to ‘active’ tissues to feed neuronal firing (in the case of functional activation) and the diffusion of water preferentially along neuronal fiber bundles (in the case of structural connectivity).
Each of these specialised MRI contrasts are even more sensitive to motion than structural imaging. Therefore, considerable challenges lie before us. Nevertheless, we have been able to apply these imaging to neonates and we are confident that we will soon do the same for fetal imaging.
Such data is extremely powerful as it will allow us to learn the pattern of cognitive development, allowing us to build models of healthy brain growth in the rapidly changing period of 20-40 weeks post gestation. During this time the brain surface goes from being completely smooth to developing a complex pattern of folds. These increase the surface area of the brain, and are linked to the development of neuronal fiber connections, and therefore increase in network complexity of the brain.
One important pause for thought came when I was questioned by someone with Autism who was concerned, that people like her would become the ‘casualties’ of this type of work. It is apparent that we will face complex ethical challenges, as our research progresses.