|
Graph Neural Networks in Epilepsy Surgery
Hrtonová, Valentina ; Filipenská, Marina ; Klimeš, Petr
Epilepsy surgery presents a viable treatment option for patients with drug-resistant epilepsy, necessitating precise localization of the epileptogenic zone (EZ) for optimal outcomes. As the limitations of currently used localization methods lead to a seizure-free postsurgical outcome only in about 60% of cases, this study introduces a novel approach to EZ localization by leveraging Graph Neural Networks (GNNs) for the analysis of interictal stereoelectroencephalography (SEEG) data. A GraphSAGE-based model for identifying resected seizure-onset zone (SOZ) electrode contacts was applied to a clinical dataset comprising 17 patients from two institutions. This study uniquely focuses on the use of interictal SEEG recordings, aiming to streamline the presurgical monitoring process and minimize risks and costs associated with prolonged SEEG monitoring. Through this innovative approach, the GNN model demonstrated promising results, achieving an Area Under the Receiver Operating Characteristic (AUROC) score of 0.830 and an Area Under the Precision-Recall Curve (AUPRC) of 0.432. These outcomes along with the potential of GNNs in leveraging the patient-specific electrode placement highlight their potential in enhancing the accuracy of EZ localization in drug-resistant epilepsy patients.
|
|
Investigating very high-frequency oscillations phenomena in human brain
Kolajová, Martina ; Cimbálník, Jan (referee) ; Klimeš, Petr (advisor)
Identifikace zóny vzniku záchvatu (SOZ) – oblasti mozku, kde začínají záchvaty – je klíčová pro úspěch chirurgické resekce epileptického ložiska. Význam vysokých vysokofrekvenčních oscilací (VHFO) spočívá v jejich roli jako více specifického biomarkeru této nemoci, přestože mechanismus jejich vzniku zůstává nejasný. Jednotlivé neurony nejsou fyziologicky schopné generovat oscilace v rozsahu VHFO. Je proto předpokládáno, že VHFO zaznamenané makroelektrodami mohou vznikat sumací aktivity různých neuronálních skupin, které pálí asynchronně. Tato diplomová práce se věnuje analýze jevů VHFO s důrazem na jejich vztah k epilepsii. Pomocí metod s využitím měření na mikroelektrodovém poli (MEA) jsou zkoumány mechanismy vzniku VHFO prostřednictvím analýzy dat v programovacím jazyce Python.
|
|
Reconstruction and strengthening of reinforced concrete ceiling construction
Klimeš, Petr ; Zich, Miloš (referee) ; Perla, Jan (advisor)
The final thesis deals with strengthening of cast-in-place concrete ceiling in existing construction for couple of escalators. New prestressed beam, columns in new positions and strengthened girder with unbonded strands are building in this adjustments. Every member, including the existing construction, is verified for ultimate and serviceability limit state. Prestressed members are alternatively designed with bonded strands and these variants are comparison in terms of utilization and calculation.
|
|
Monolithic concrete building at sewerage plant
Klimeš, Petr ; Bažant, Zdeněk (referee) ; Strnad, Jiří (advisor)
Bachelor's thesis deals with design of monolithic concrete construction with a timber roof truss for a device of sewerage plant. This construction consists of these elements- strip foundation, walls with lintels, slab and timber trusses. This thesis contains static analysis of the component parts and their drawing documentation.
|
|
Analysis of epileptogenic tissue response to intracranial electrical stimulation
Formánková, Zuzana ; Klimeš, Petr (referee) ; Cimbálník, Jan (advisor)
This work deals with the methods of intracranial electrical stimulation and their usage in the localization of epileptogenic tissue. The aim of the thesis is to assess, with help of the proposed markers, the reaction of pathological tissue on the electrical stimulation. Among the suitable markers high-frequency oscillations were classified, interictal spikes, changes in the connectivity, and the signal power within the frequency zones. The markers were detected on the iEEG records taken at the Fakultní nemocnice u sv. Anny in Brno. A software in the Python language has been designed for the purpose of analysis and detection; the software uses the detection algorithms of the EPYCOM library. In the final part of the thesis, the occurrence of the markers was analyzed in terms of dependency on the electrical stimulation. The influence of the electrical stimulation on the iEEG records of patients with epilepsy has been proved.
|
|
Model of traffic lights crossroads
Klimeš, Petr ; Zuth, Daniel (referee) ; Marada, Tomáš (advisor)
The purpose of this bachelor work is concept and realisation of traffic lights crossroad model which is controlled by programmable logic controller (PCL). Model is supposed to be used as a laboratory device during practical lessons of programmable logic controllers to demonstrate basic principle of semaphore lights controlled traffic crossroad.
|
| |
| |
|
Establishing Mutual Links among Brain Structures
Klimeš, Petr ; Hlinka,, Jaroslav (referee) ; Krajča,, Vladimír (referee) ; Halámek, Josef (advisor)
The Human brain consists of mutually connected neuronal populations that build anatomically and functionally separated structures. To understand human brain activity and connectivity, it is crucial to describe how these structures are connected and how information is spread. Commonly used methods often work with data from scalp EEG, with a limited number of contacts, and are incapable of observing dynamic changes during cognitive processes or different behavioural states. In addition, connectivity studies almost never analyse pathological parts of the brain, which can have a crucial impact on pathology research and treatment. The aim of this work is connectivity analysis and its evolution in time during cognitive tasks using data from intracranial EEG. Physiological processes in cognitive stimulation and the local connectivity of pathology in the epileptic brain during wake and sleep were analysed. The results provide new insight into human brain physiology research. This was achieved by an innovative approach which combines connectivity methods with EEG spectral power calculation. The second part of this work focuses on seizure onset zone (SOZ) connectivity in the epileptic brain. The results describe the functional isolation of the SOZ from the surrounding tissue, which may contribute to clinical research and epilepsy treatment.
|
|
Evaluation of blinking and the observed place on the monitor
Klimeš, Petr ; Jirgl, Miroslav (referee) ; Janáková, Ilona (advisor)
The purpose of this thesis is to create a system for user’s gaze and eye analysis in a driving simulator. There is a research of existing methods for facial landmark detection, head pose and gaze estimation and eye blink detection. Software and hardware of the simulator is described and a camera is added for user recording. In the main part, the Eye Aspect Ratio method is tested for blink detection. Next, head pose detection from facial landmarks is tested. Furthermore, Convolutional Neural Networks are created and trained for gaze estimation. From existing algorithms, OpenFace library is tested. OpenFace library is added to existing software in Unreal Engine and used for gaze and blink detection. A gaze calibration procedure is created and tested for better accuracy.
|