Final Report Summary - OVOC (Ultra Fast Magnetic Resonance Imaging using One-Voxel-One-Coil Acquisition)
The goal of the project was to develop new methods and applications for ultrafast magnetic resonance imaging based on the principles of One Voxel One Coil (OVOC)-imaging published in 2006.
OVOC uses the sensitive volumes of multicoil arrays consisting of multiple small coils as primary source of spatial information. Additional spatial information is gathered by gradient encoding while still maintaining very high acquisition speed. The potential range of applications outlined in the proposal included human applications in the brain, cardiovascular system and oncology at 3T as well as at 7T as well as small animal experiments down to microMR on a submillimeter scale.
Given the groundbreaking nature of this new approach the project plan was to first explore the possible range of applications, identify the most promising candidate subprojects for further exploration and then to perform in-depth development and applications for these key target areas.
The first and most promising area of application has been the use of OVOC for fast and dynamic measurement of brain function. In analogy to electrophysiological measurements (EEG and MEG) the term MREG (MR-encephalography) was coined for the specific implementation.
Initial experiments revealed ECG- and respiratory signal fluctuations as dominant contributing sources of MREG-measurements. In order to distinguish these physiological confounds from functional signals specifications for MREG were set to an acquisition speed of 10 fps for full brain acquisition (64x64x64 matrix, 3 mm isotropic resolution). Development of MREG-methods to meet these extremely challenging targets was greatly successful and rewarding. We went through different acquisition strategies based on single shot 3D-acquisition with different space encoding schemes (Projections, Rosettes, Concentric Shells, Stack of Spirals (SoS)). With the recently introduced SoS-method we are able to meet our specifications at an image quality comparable to standard EPI but at an increase in acquisition speed by a factor of 10-20. As a pre-requisite to translate the technology to applications studies we were able to implement a reconstruction pipeline based on Inverse Image Reconstruction which allows reconstruction of a full dynamic MREG acquisition in a few hours (compared to days to weeks in the original implementation). This was afforded in part through hardware acceleration (parallel computing, GPUs), but mainly through optimization of the algorithms used.
With this methodological progress we have focused the application of MREG to two specific areas of applications: Simultaneous measurements of MREG with EEG allowed us to investigate the spatiotemporal development of the cortical correlates underlying interictal electrical discharges (IED in epilepsy patients. The successful demonstration of the localization of every single EEG-detected IED led to the development of the integrated MREG-EEG device ‘EPIFocus’, Pierre Levan as leader of the subproject was awarded the German High Tech Championship Award 2014 for this development, which has great potential in the management of patients with intractable epilepsy.
A second area of neuroscientific applications is the investigation of resting state networks (RSNs) in task-free MREG. As a recent highlight of this particular area of applications we were able to demonstrate correlated brain activity in a frequency windows of 0.5-0.8 Hz and 2-5 Hz, much faster than conventional RSNs (0.01-0.1 Hz). This allows the investigation of the dynamic reorganization of RSNs with a timeframe of 10-20 s, which is key to understanding of how the brain interacts with external inputs.
The method has been disseminated to several other research labs.
Additional areas of research for OVOC were targeted wholebody imaging at 7T, for which a 32-channel coil has been built.
As a further area of application development fast imaging (> 20 frames/s) techniques for the dynamic examination of the laryngopharyngeal system for investigation ranging from professional singers to speech disorders and pathologies of the laryngopharyngeal system have been successfully implemented.
Application of OVOC to microMR led to the development of microsystem technologies for manufacturing of multicoil array of scalable size and number of coil elements. Application studies focused on microMR of skin preparations and – most recently – on the investigation of hippocampal slices of mice brain.
The tremendous success of the developments has been afforded through strong interdisciplinary exchange both within the project (most notably J.Korvink IMTEK, for development of coils) but also through synergies with other projects and partners.
In summary, the model of the ERC Advanced Grant has worked out exceptionally well for our innovative research and is perceived to be the best and most efficient European Research Program encountered so far in my career.
OVOC uses the sensitive volumes of multicoil arrays consisting of multiple small coils as primary source of spatial information. Additional spatial information is gathered by gradient encoding while still maintaining very high acquisition speed. The potential range of applications outlined in the proposal included human applications in the brain, cardiovascular system and oncology at 3T as well as at 7T as well as small animal experiments down to microMR on a submillimeter scale.
Given the groundbreaking nature of this new approach the project plan was to first explore the possible range of applications, identify the most promising candidate subprojects for further exploration and then to perform in-depth development and applications for these key target areas.
The first and most promising area of application has been the use of OVOC for fast and dynamic measurement of brain function. In analogy to electrophysiological measurements (EEG and MEG) the term MREG (MR-encephalography) was coined for the specific implementation.
Initial experiments revealed ECG- and respiratory signal fluctuations as dominant contributing sources of MREG-measurements. In order to distinguish these physiological confounds from functional signals specifications for MREG were set to an acquisition speed of 10 fps for full brain acquisition (64x64x64 matrix, 3 mm isotropic resolution). Development of MREG-methods to meet these extremely challenging targets was greatly successful and rewarding. We went through different acquisition strategies based on single shot 3D-acquisition with different space encoding schemes (Projections, Rosettes, Concentric Shells, Stack of Spirals (SoS)). With the recently introduced SoS-method we are able to meet our specifications at an image quality comparable to standard EPI but at an increase in acquisition speed by a factor of 10-20. As a pre-requisite to translate the technology to applications studies we were able to implement a reconstruction pipeline based on Inverse Image Reconstruction which allows reconstruction of a full dynamic MREG acquisition in a few hours (compared to days to weeks in the original implementation). This was afforded in part through hardware acceleration (parallel computing, GPUs), but mainly through optimization of the algorithms used.
With this methodological progress we have focused the application of MREG to two specific areas of applications: Simultaneous measurements of MREG with EEG allowed us to investigate the spatiotemporal development of the cortical correlates underlying interictal electrical discharges (IED in epilepsy patients. The successful demonstration of the localization of every single EEG-detected IED led to the development of the integrated MREG-EEG device ‘EPIFocus’, Pierre Levan as leader of the subproject was awarded the German High Tech Championship Award 2014 for this development, which has great potential in the management of patients with intractable epilepsy.
A second area of neuroscientific applications is the investigation of resting state networks (RSNs) in task-free MREG. As a recent highlight of this particular area of applications we were able to demonstrate correlated brain activity in a frequency windows of 0.5-0.8 Hz and 2-5 Hz, much faster than conventional RSNs (0.01-0.1 Hz). This allows the investigation of the dynamic reorganization of RSNs with a timeframe of 10-20 s, which is key to understanding of how the brain interacts with external inputs.
The method has been disseminated to several other research labs.
Additional areas of research for OVOC were targeted wholebody imaging at 7T, for which a 32-channel coil has been built.
As a further area of application development fast imaging (> 20 frames/s) techniques for the dynamic examination of the laryngopharyngeal system for investigation ranging from professional singers to speech disorders and pathologies of the laryngopharyngeal system have been successfully implemented.
Application of OVOC to microMR led to the development of microsystem technologies for manufacturing of multicoil array of scalable size and number of coil elements. Application studies focused on microMR of skin preparations and – most recently – on the investigation of hippocampal slices of mice brain.
The tremendous success of the developments has been afforded through strong interdisciplinary exchange both within the project (most notably J.Korvink IMTEK, for development of coils) but also through synergies with other projects and partners.
In summary, the model of the ERC Advanced Grant has worked out exceptionally well for our innovative research and is perceived to be the best and most efficient European Research Program encountered so far in my career.