Community Research and Development Information Service - CORDIS


CoBCoM Report Summary

Project ID: 694665
Funded under: H2020-EU.1.1.

Periodic Reporting for period 1 - CoBCoM (Computational Brain Connectivity Mapping)

Reporting period: 2016-09-01 to 2018-02-28

Summary of the context and overall objectives of the project

One third of the burden of all the diseases in Europe is due to problems caused by diseases affecting the brain. Although exceptional progress has been obtained for exploring it during the past decades, the brain is still terra-incognita and calls for specific research efforts to better understand its architecture and functioning.

CoBCoM is primarily devoted to develop new generation of computational models and methodological breakthroughs for brain connectivity mapping. To solve the limited view of the brain provided just by one imaging modality, our models are solidly grounded on advanced and complementary integrated non invasive imaging modalities: diffusion Magnetic Resonance Imaging (dMRI) and Electro & Magneto-Encephalography (EEG & MEG).

To take up this immense challenge, CoBCoM has the overall objectives to develop advanced dMRI and MEEG source reconstruction methods for structural and functional brain connectivity mapping and build integrated dynamical brain networks from dMRI & M/EEG. This will help to push far forward the state-of-the-art in these modalities, to better understand and reconstruct the structural and functional brain connectivity and to provide a clinical added value to better identify and characterize abnormalities in brain connectivity. Clearly, this represents a fantastic scientific challenge as well as a pressing clinical need that, when solved, will greatly open new perspectives in neuroimaging and positively impact the unacceptable burden of brain diseases.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

We started to develop advanced dMRI and M/EEG methods for structural and funtional connectivity mapping and contributed tp develop

i) Generative and ground-breaking models for advanced acquisition and processing of dMRI data,
ii) New concepts and approaches grounded on microstructures from dMRI and apply them to Microstructure based tractography,
iii) Advanced M/EEG source reconstruction methods with dMRI based spatial regularization,
iv) Parcellation techniques to help reconstructing the information flow in the brain from EEG/MEG data and dMRI information.

i) Advanced acquisition and processing of dMRI data :

We developed technologies to compute new sampling schemes for the diffusion MRI acquisitions sampling the space of direction, magnetic gradient strength and diffusion time, namely qt-dMRI. To design an acquisition scheme that can cater to specific scanner settings and study protocols we have formulated the problem in terms of finding the
optimal subsample of very dense sample acquisition. This will enable experimenters to acquire once a very dense sample and then explore different acquisition possibilities in accordance to their experimental requirements and constraints. We have shown that this problem is currently unsolvable in the general case as it is NP-Hard. In fact, it is equivalent to the well-known knapsack problem in theoretical computer science. To provide approximate solutions, we have developed a probabilistic relaxation of the problem and implemented it in a 11.7T small animal scanner and different 3T clinical and pre-clinical scanners. This contribution has been presented in CD-MRI Miccai'2017 [Ref. 7].

We developed an effective representation of the four-dimensional diffusion MRI signal – varying over three dimensional q-space and diffusion time t – a sought-after and still unsolved challenge in diffusion MRI (dMRI). In particular, we proposed a functional basis approach that is specifically designed to represent the dMRI signal in this qt-space. This qt-dMRI can be seen as a time-dependent realization of q-space imaging by Paul Callaghan and colleagues. We used GraphNet regularization – imposing both signal smoothness and sparsity – to drastically reduce the number of diffusion-weighted images that is needed to represent the dMRI signal in the qt -space. As the main contribution, qt-dMRI provides the framework to – without making biophysical assumptions – represent the qt -space signal and estimate time-dependent q-space indices (qt-indices), providing a new means for studying diffusion in nervous tissue. In the hopes of opening up new t-dependent venues of studying nervous tissues, qt-dMRI is the first of its kind in being specifically designed to provide open interpretation of the qt-diffusion signal. The contributions related to this work have been published in Medical Image Analysis 2017 [Ref. 5] and Fick's PhD thesis [Ref. 9]

ii) Microstructure imaging and tractography:

Non-invasive estimation of brain white matter microstructure features using dMRI – otherwise known as Microstructure Imaging – has become an increasingly complex and difficult challenge over the last decade. Within the framework of Fick's PhD thesis [Ref. 9], we provided an extensive review, analysis, validation and discussion on state-of-the-art modeling approaches in Microstructure Imaging. We deconstructed and classified microstructure models by their components and the microstructural interpretation they lend to their model parameters, with a particular effort to expose their assumptions and limitations. We followed this with a validation of intra-axonal volume fraction estimation between different microstructure modeling approaches, using both synthetic Monte-Carlo generated data and spinal cord data with registered diffusion MRI and ground truth histology. We also addressed current concerns about the degeneracy of the solutions of multi-compartment models when the diffusivities are not fixed, current

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

We have timely started to work towards our objectives and harvested very important results in developing advanced dMRI signal modeling for dMRI representation and acquisition in space and time, in developing new concepts and approaches grounded for tissue microstructure characterization from dMRI and apply them to develop Microstructure based tractography. We also developed advanced M/EEG source reconstruction methods with dMRI based spatial regularization, and developed parcellation techniques to help reconstructing the information flow in the brain from EEG/MEG data and dMRI information. To this end, we addressed the problem of groupwise structural parcellation of the whole cortex, started to work on the developemnt of a large brain effective network from EEG/MEG data and dMRI and started to contribute to the inference and visualization of information flow using dMRI and EEG.

Upon completion of CoBCoM, we are confident that we will succeed to introduce pioneering and advanced dMRI and M/EEG ground-breaking methods and build a joint dynamical structural-functional brain connectivity network which will be exploited to identify and characterize white matter connectivity abnormalities in brain diseases. This will be a major scientific accomplishment and will certainly impact clinical practice in dealing with neurological disorders. Even if the overall objectives of CoBCoM can be considered as extremely challenging and very ambitious, we are confident that many important progress with significant scientific results and clinical impact will occur in support of the main objectives. We are convinced that the scientific drive and the product of our efforts will push forward and far beyond the state of the art measurements and modeling approaches that span across the disciplines of neuroimaging and health, computer sciences and applied mathematics.

Thanks to our rolling start, we succeeded to publish 23 papers among which 7 papers in the most selective journal (Neuroimage, Medical Image Analysis, Human Brain Mappig..) and the most important conferences of the domain (ISMRM, ISBI, IPMI, MICCAI...). We will continue to promote our results and disseminate the product of our efforts via publications in the most selective journals and conferences of our domains, via the organization of the second COBCOM workshop and via the software development of libraries such our dMipy library and our OpenMEEG software platform, a C++ opensource software available on GitHub for quasistatic electromagnetics, solving forward problems of EEG, MEG, ECoG, intracerebral EEG, and integrated into several software suites for MEG/EEG analysis and processing (Brainstorm, Fieldtrip, SPM).

Related information

Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top