Focused-Ultrasound Embeded Optical Computing Tomography System for Real-Time in vivo Diagnosis and Controlled Anticancer Drug Delivery

Magnetic resonance imaging (MRI) technique has become an indispensable tool for human brain mapping, both in healthy volunteers for research purposes and in patients for clinical diagnosis. The participant is usually instructed to maintain a steady head position during an MRI exam. However, it is difficult for most people to stay immobilized. Practically, head motion is inevitable and results in significant image artifact, especially for pediatrics, elderly people or patients suffering from certain medical condition (e.g. stroke and Parkinsons disease). This problem not only happens in clinical imaging, but also in research. In functional MRI, head motion correlated with the stimulus onset time can produce false activations. In diffusion MRI, it produces significant bias in diffusion indices such as fractional anisotropy and mean diffusivity. The aim of this study is to develop a real-time head motion correction system, consisting of accelerometer sensors and two cameras, for existing MRI system. An embedded GPU development toolkit will be employed as the computing platform with the advantage of low cost and high computing power. The accelerometer is capable of monitoring the inclination/tile angles of head, and two out-bore cameras with higher frame rate will be used for tracking the head motion and then quantifying rigid body motion. The collected head pose information about translation/rotation will be sent to MRI console directly for adapting scan geometry to real-time compensate for patient motion during data acquisition. We also plan to develop a software system for real-time MRI data analysis using high-speed data communicate network to boost computing performance by GPU workstation. The constructed real-time head-motion correction system will bring great benefits to the major three applications: clinical MRI, resting-state functional MRI and diffusion MRI. Furthermore, we will design and release an application toolbox for advanced researches, including neurofeedback, sleep neuroimaging and pre-surgical planning. In brief, combining with the real-time MRI data analysis system, it undoubtedly facilitates the clinical diagnosis and academic progress of MRI investigations. Exclusively, the realization of this project will assist the research team to carry out the first MR-compatible real-time motion correction system in Taiwan, where this novel system and technique has highly potential for international patents and for industrial commercialization in the near future.

Project ID:PB10408-5710
External Project ID:MOST104-2221-E182-034