Conductivity Tensor Imaging of the Human Brain Using Water Mapping Techniques

• Published in: Frontiers in neuroscience Vol. 15; p. 694645
• Main Authors: Marino, Marco, Cordero-Grande, Lucilio, Mantini, Dante, Ferrazzi, Giulio
• Format: Journal Article
• Language: English
• Published: Lausanne Frontiers Research Foundation 30.07.2021; Frontiers Media S.A
• Subjects: Brain; Brain mapping; Brain tumors; Cerebrospinal fluid; conductivity tensor imaging; diffusion tensor imaging; EEG; Electric properties; Electrical conductivity; electrical properties tomography; Helmholtz equations; Magnetic resonance imaging; Magnetoencephalography; Mapping; Neuroimaging; Neuroscience; Noise; Open source software; Pipelines; Spatial discrimination; Substantia alba; Substantia grisea; Tissues; water content mapping
• ISSN: 1662-453X; 1662-4548; 1662-453X
• DOI: 10.3389/fnins.2021.694645

Conductivity tensor imaging (CTI) has been recently proposed to map the conductivity tensor in 3D using magnetic resonance imaging (MRI) at the frequency range of the brain at rest, i.e., low-frequencies. Conventional CTI mapping methods process the trans-receiver phase of the MRI signal using the MR electric properties tomography (MR-EPT) technique, which in turn involves the application of the Laplace operator. This results in CTI maps with a low signal-to-noise ratio (SNR), artifacts at tissue boundaries and a limited spatial resolution. In order to improve on these aspects, a methodology independent from the MR-EPT method is proposed. This relies on the strong assumption for which electrical conductivity is univocally pre-determined by water concentration. In particular, CTI maps are calculated by combining high-frequency conductivity derived from water maps and multi b-value diffusion tensor imaging (DTI) data. Following the implementation of a pipeline to optimize the pre-processing of diffusion data and the fitting routine of a multi-compartment diffusivity model, reconstructed conductivity images were evaluated in terms of the achieved spatial resolution in five healthy subjects scanned at rest. We found that the pre-processing of diffusion data and the optimization of the fitting procedure improve the quality of conductivity maps. We achieve reproducible measurements across healthy participants and, in particular, we report conductivity values across subjects of 0.55 ± 0.01 S m , 0.3 ± 0.01 S m and 2.15 ± 0.02 S m for gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF), respectively. By attaining an actual spatial resolution of the conductivity tensor close to 1 mm in-plane isotropic, partial volume effects are reduced leading to good discrimination of tissues with similar conductivity values, such as GM and WM. The application of the proposed framework may contribute to a better definition of the head tissue compartments in electroencephalograpy/magnetoencephalography (EEG/MEG) source imaging and be used as biomarker for assessing conductivity changes in pathological conditions, such as stroke and brain tumors.