|
PresentationsGraph theory, topology methods, and hidden Markov models in the analysis of structural, functional, and effective brain connectomes of patients diagnosed with schizophrenia and control groupsLomonosov Moscow State University Institute for Advanced Brain Research, 27 Lomonosovsky Ave., Building 1, 119192, Moscow, Russia, +7 (495) 938-25-48, tiuq@yandex.ru The study of the mechanisms of the brain, the construction of a theory of consciousness and the solution of the difficult problem of consciousness is in the field of creating a mathematical apparatus for modeling neurocognitive processes, developing mathematics for describing cognitive and neurophysiological spaces, finding a computational description of emergence and a set of metrics for calculating connectivity matrices and causal relationships obtained on atlas and functionally homogeneous regions (FOR) of the brain. In this work, brain connectomes were constructed and analyzed based on functional MRI data of the resting state of the brain (10 min., TR: 720 msec, 1000 msec, 1100 msec) and structural connectomes from the database of the Alekseev Design Bureau No. 1 [1]. The fMRI data was preprocessed in the SPM12 program, and the diffusion tractography data was preprocessed in the DSI Studio program. The ODDS were obtained and analyzed based on our own developed software. To construct functional connectivity matrices, a set of metrics was used: the Pearson correlation coefficient, the h2 method, the value of mutual information MI, the coherence coefficient based on Wavelet and Fourier transforms, the Granger causality coefficient, the transfer entropy coefficient and the A(H) value (MVAR-frequency domain based techniques). The dynamic causal modeling (DCM) method was additionally used to build effective connectomes. The connectivity of the brain regions was analyzed using standard metrics for evaluating graph features, as well as based on the calculation of rich-club structures, using methods for analyzing simplices and calculating graph curvature. The method of hidden Markov models was used to identify the operating conditions of neural networks of the brain. As a result, differences were found in the neural network architectures of the brain connectomes of the healthy control group and patients diagnosed with schizophrenia. The developed approaches can be used in creating an objective assessment in the diagnosis of the disease and in studying the mechanisms of the brain's neural networks in cognitive processes. The research was carried out within the framework of the state assignment of the Lomonosov Moscow State University.
Literature 1. Database of MRI brain data of patients diagnosed with schizophrenia and norms // Registration number (certificate): RU2024622180.
|
