Real-time brain imaging using pulsed radar may give “live” images of brain activity currently unavailable. Non-invasive brain imaging is difficult because the skull is hard to penetrate and body substances are absorbing most signals. Magnetic Resonance Imaging is the most accurate brain-imaging mode, but is expensive, large and limited in temporal resolution since creation of each image is time consuming (seconds).
As a part of the MEDIMA research effort at IFI we are developing a functional brain imaging platform based on a modified pulsed radar produced by the Norwegian company, Novelda AS (http://www.xethru.com). An adapted module with reduced signal frequency is used for brain imaging . The radar chip is a complex (SoC) with SPI hardware interface.
Based on experimental verification we have confirmed sensing of heartrate inside the brain encouraging us for continued efforts on more detailed brain imaging. Next step is to build a 16-module circular brain imager to fit around the head.
The complete instrument should be as small and light as possible for portable use. The radar modules deliver a significant volume of imaging data for collection by a suitable computer. In this project we have adapted Beagle Bone Black 
with reasonably good real-time features. The module is equipped with 2 SPI interfaces allowing for higher performance.
Our aim is to be able to operate the radar modules with a 100MHz clock delivering images with 100fps. In addition to data collection, some simple signal filtering and compression is desired before (wireless) forwarding of data. We are still in the initial design process of the brain imaging modules and additional computational power may be required perhaps using FPGA.
We are seeking two MSc students working together primarily on low level, real-time software instrumentation using linux-based Beagle Bone Black. The hardware platform is almost operational and the MSc students will work mostly together with Mathias Tømmer and Kristian Kjelgård at the NANO group.