Cortical Mapping with Diffuse Optical Compressed Sensing
Diffuse optical tomography is an emerging non-invasive medical imaging technology which can be used to map the optical properties of biological tissues using near-infrared light. Such measurements are of significant clinical and scientific interest, since the differing absorption spectra of the oxygenated and deoxygenated forms of haemoglobin allow one to spectroscopically image haemodynamic activity and tissue oxygenation within the brain. The Biomedical Optics Research Laboratory (BORL), and the Centre for Medical Image Computing (CMIC) have both played a major role in the early development of this technology, previously demonstrating the ability to monitor localised changes in oxygenation in both infants and adults. Significant research effort is currently being directed to improving the spatial resolution and sensitivity of the technique. This effort is leading to the development of new instrumentation which employs dense arrays of optical sources and detectors which are placed upon, and fully encompass the head. Such devices generate vast amounts of data, and present new challenges in maintaining the fast acquisition rates which are a key advantage of this technology. New holistic approaches to the development of instrumentation and the requisite image reconstruction problem are now required to allow the technique to advance. With the appropriate development, the next generation of DOT imaging technology has the potential to offer real-time continuous monitoring of the brain at the bed- or cot-side, which would have a disruptive impact in neuroimaging, promote the targeted use of scarce rehabilitatory resources, and simultaneously lead to considerable advances in our knowledge of the pathophysiology of brain injury.
Aims and Objectives:
Mapping cortical function with near-infrared light is a novel technique which can generate a rich set of data corresponding to functional responses to external stimuli. Similar to functional MRI, the data gives an indirect response via the convolution of an activation signal with the haemodynamic response function, but in contrast to fMRI it provides direct estimation of both blood volume and oxygenation, and is much faster. In diffuse optical cortical mapping (DOCM), data are acquired using dense arrays of optical sources and detectors, and reconstruction calls for the inversion of a mapping from a function on the cortical surface to one on the scalp surface where measurements are taken. In this project we aim to exploit concepts from compressed sensing to accelerate the data acquisition to enable real time imaging, comparable to Electro-Encephalography (EEG) source imaging but with much higher resolution. Successful development will have a disruptive impact in neuroimaging.