In surgical practice, if the tumor is very close to the cortex or the associative pathways responsible for language or motor function, the surgeon must balance the radicality of resection and the preservation of functionally significant brain regions. The advantages of broad resection should be compared with the degree of neurological deficiency that arise due to the destruction of functionally significant cortical areas. Since the degree of individual anatomical variability of these areas is high enough, preoperative localization and intra-operative mapping are often required.
Intraoperative cortical stimulation mapping (CSM) is an invasive procedure to locate the function of specific brain regions. It is considered as a golden standard for identifying certain brain regions, while preoperative mapping is usually performed non-invasively by task-based fMRI (t-fMRI). In order to map motor and language areas, t-fMRI with motor and language paradigms is used. This method can complement the surgical treatment strategy by meaningful information obtained non-invasively. The result of t-fMRI study depends largely on the ability of the patient to perform a particular task. Because performing tasks for glioma patients could be hard or even impossible, attempts are made to replace t-fMRI with resting-state fMRI (rs-fMRI) – when a subject is at rest and not performing any particular task in scanner.
In this project we develop a presurgical planning pipeline for individual brain eloquent areas localization based on the combination of multimodal neuroimaging data: open t-fMRI and rs-fMRI databases, such as Human Connectome Project, historical CSM and tb-fMRI data as well as individual rs-fMRI data. Tensor decomposition as well as machine learning methods including transfer learning, are extensively used here.
Pavlov, S., Artemov, A., Sharaev, M., Bernstein, A., Burnaev, E., 2019. Weakly Supervised Fine Tuning Approach for Brain Tumor Segmentation Problem, in: 2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA). IEEE, pp. 1600–1605. https://doi.org/10.1109/ICMLA.2019.00263
Sharaev, M., Smirnov, A., Melnikova-Pitskhelauri, T., Orlov, V., Burnaev, E., Pronin, I., Pitskhelauri, D., Bernstein, A., 2018. Functional Brain Areas Mapping in Patients with Glioma Based on Resting-State fMRI Data Decomposition, in: 2018 IEEE International Conference on Data Mining Workshops (ICDMW). IEEE, pp. 292–298. https://doi.org/10.1109/ICDMW.2018.00049
Smirnov, A.S., Sharaev, M.G., Melnikova-Pitskhelauri, T. V., Zhukov, V.Y., Bikanov, A.E., Sharova, E. V., Pogosbekyan, E.L., Turkin, A.M., Fadeeva, L.M., Pitskhelauri, D. V., Kornienko, V.N., Pronin, I.N., 2018. Resting state fMRI in pre-surgical brain mapping. Literature review. Med. Vis. 6–13. https://doi.org/10.24835/1607-0763-2018-5-6-13