Combined MEG and Cytoarchitectonic Data Imaging

In most MEG and other functional imaging studies the meaning of an activated cortical area is usually interpreted in relation to the surrounding gyral and sulcal landmarks. In contrast, previous studies showed that such landmarks do not have a fixed relationship to cytoarchitectonic boundaries. Moreover, it is not possible to identify any cortical area from functional imaging or anatomical data alone. In our MEG research group we developed a dedicated software (ROIfinder) to automatically extract strong neural activations from the huge amount of time series of reconstructed current densities for all voxels. First the neuromagnetic activity is estimated by means of the magnetic field tomography (MFT) providing time courses of fully three-dimensional cerebral current source density volumes with millisecond time resolution. We then apply the ROIfinder to search for strong neural activity within each whole reconstruction space that fulfils certain spatial and temporal connectivity requirements. However, it remains difficult to unambiguously attribute the underlying neuronal activity to its corresponding neuroanatomical structure. For this reason, the activation patterns are mapped onto MRI scans together with so-called probability maps providing information about the frequency of a given anatomical area being located at a given position in the stereotaxic space of the reference brain. Such maps were derived from cytoarchitectonic analyses of 10 human post-mortem brains that were transferred into the same space. Using a pattern reversal task experiment (black and white checkerboard) we demonstrate the feasibility of automatically extracted neuromagnetic activity that is mapped together with cytoarchitectonic probability maps. With this experiment we combine the necessary information from MEG data analysis, gyral and sulcal landmarks and cytoarchitectonic boundaries to better understand the meaning of neuromagnetic activations.