Quest for the Human Connectome: Imaging in the Information Age
All, here is another entry from my colleague in Imaging, Roshni Bhaglia:
Last month, a group of researchers from GE’s Visualization and Computer Vision Lab presented their work at the SPIE:Medical Imaging conference in San Diego. Conference keynote speeches typically have me ephemerally excited about the next scientific challenge until I get sucked back into the here-and-now deadline vortex at work. However, Jeffrey Lichtman’s excellent introduction to the Connectome, a wiring map of an organism’s neural network, had me so – well, wired – that on returning I delved deeper into the subject and stumbled upon the NIH funded 30 million dollar “Human Connectome Project” (HCP).
The idea of charting the human connectome (pronounced akin to genome), that is, building the connectivity map of the human brain, can be traced back to two Nobel winning luminaries who were at loggerheads: Golgi and Cajal. This notion has gathered a lot of traction in recent years with the inception of the Human Connectome Project. The HCP aims to develop and share knowledge about the structural and functional connectivity of the human brain.
There has been much debate around the relevance and feasibility of a project of this magnitude. Proponents applaud the projects’ foresight, claiming that the human connectome will be indispensible to studying how variations in brains are related to numerous neural disorders such as Alzheimer’s, Parkinson’s disease, Huntington’s disease and schizophrenia. In contrast skeptics doubt that a “blueprint brain connectome” even exists, noting that there is vast individual variability in the connections of human brains. Furthermore, the connections or information pathways of a brain adapt with experience and age and to make matters worse, there is no clear granularity level at which the connectome can be assumed to be complete. Is it enough to define connections between smaller sub-divisions of the brain or should the connectome include connections between individual neurons or even finer resolutions?
While both opinions are not without merit, it is evident that advances in imaging technology, machine learning and data mining will play a seminal role in realizing the goals of the HCP. In fact, the HCP would have been unthinkable without the advent of MR Spectroscopy, Diffusion Weighted MRI and Diffusion tensor Imaging. These techniques use diffusion of water molecules to study the cellular and molecular structure of fiber tracts in the white matter of the living human brain. While MRI can depict brain connections at a larger fiber bundle scale, development of numerous imaging solutions to systematically capture information about the connectome at s cellular level is currently underway. Some of the more recent forays in this area include the high throughput electron microscope, ATLUM and ofcourse, the brainbow.
Once this gargantuan amount of image data encoding the functional and structural interdependence of the human brain has been gathered, the second phase of HCP will likely emphasize data processing and analyses. Important advances in machine learning, data-mining and dimensionality reduction will be needed to infer the underlying neural connections that will make up the “Cerebral Blueprint”, if one exists!
Whether the Human Connectome Project will do for neuroscience what the Human Genome Project did for genetics remains to be seen. Regardless of the outcome however, this much is clear: the technological growth spurred in the imaging and data analyses fields to meet the goals of the project will be well worth the investment.