fMRI – Noninvasive Imaging
The research presented here uses two very different methods for measuring brain activity during language production, and they complement each other in various ways.
Functional magnetic resonance imaging (fMRI) is by far the most widespread modern method for investigating the brain’s functions. It has many advantages, for instance it can visualize the whole brain, it is non-invasive (does not require surgery or other intervention with the body), and it can see subtle changes related to how metabolically demanding a given brain region is during a task (that is, how ‘thirsty’ it is for the oxygen and nutrients in blood). This is a stand-in for how much a given brain region is performing its dedicated calculations, and thus a way of visualizing its involvement in a cognitive task such as speaking. Thus, fMRI allows me to map out the large neural regions that are responsible in some way for the aspects of language that I study.
However, fMRI has many limitations. The main limitation is that it has low resolution in time: it blurs together all the brain activations that happen over the course of a typical mental event (it is like seeing a movie of a complex set of actions, collapsed into a single blurred frame). Even with modern signal-processing and data acquisition paradigms designed to address this, the fundamental temporal limitations are hard to overcome: because the blood-related dynamics it measures happen slowly and gradually, over the course of seconds. Furthermore, that timecourse is different in different parts of the brain (and likely across people) so that it is impossible to reconstruct a generalizable fine-scale timecourse of activity across brain regions. The spatial resolution is also limited, and MRI scanners tend to warp and distort images in complex ways so establishing the precise location of each activity across subjects is difficult. A third type of resolution, physiological resolution – the ability to discern different types of cellular activity from each other – is limited by the fact that fMRI measures physiological processes of brain cells only through the proxy of blood flow changes (and indeed the actual nature of the fMRI signal and how it relates to brain-cell activity is not known). Therefore, fMRI is fundamentally limited in the kinds of brain computations it can reveal.
For these reasons, I combine fMRI with another method, called ICE (read more about it here), which offers fine-scale temporal and physiological resolution, but is rare and does not cover as much of the brain as fMRI. The two methods complement each other and I record both from the same patients in many cases. This is possible through the large collaborative consortium built by Eric Halgren (see collaborators) and is a truly exciting opportunity, and a privilege. Dr. Halgren’s lab also includes other methods such as MEG and EEG.
One way we combine fMRI and ICE is to first use fMRI to map the wide territory of brain tissue that includes significant activity that relates to a given mental task. In separate studies or in the same patients, ICE is used to chronicle the timecourse, physiology, and spectral dynamics at the finer temporal and spatial scales possible with ICE.