EEG data space and feature space distribution

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21460%2F18%3A00325791" target="_blank" >RIV/68407700:21460/18:00325791 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.prolekare.cz/casopisy/ceska-slovenska-neurologie/2018/dokumenty/32_slovensky_a_cesky_neurologicky_zjazd_a_65_spolocny_slovensky_a_cesky_zjazd_klinickej_neurofyziologie-1-64" target="_blank" >https://www.prolekare.cz/casopisy/ceska-slovenska-neurologie/2018/dokumenty/32_slovensky_a_cesky_neurologicky_zjazd_a_65_spolocny_slovensky_a_cesky_zjazd_klinickej_neurofyziologie-1-64</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

EEG data space and feature space distribution

Popis výsledku v původním jazyce

Physiological and pathological processes of the brain may have their characteristic manifestations in the electroencephalographic (EEG) curve. Computer processing offers different views of EEG data. It is possible to examine the continuous amplitude course over time or its individual segments, the scalp electric potential maps, or the spectral space. It is important to investigate the layout of the EEG space in order to design a criterion that would automatically classify the EEG data. The tested EEG data comes from patients with suspected epilepsy, sleep-walking patients and healthy individuals. The data were measured based on the approval of the Bulovka Hospital Ethics Commission and the Ethics Committee of the National Institute of Mental Health. This study uses both linear and nonlinear dimensional reduction techniques to detect the distribution of EEG space. We evaluated the feature space used to classify epileptic data and the space of topographic maps made from the independent components (IC). As the nonlinear dimensional reduction method was used tSNE and nonlinear PCA. The DBSCAN method was used as an auxiliary classifier. Based on our results, we assume nonlinear relationships in the EEG space. Epileptic activity is a separate set of data in the symptom space, on the contrary artifacts (EOG, EMG) are not completely separated by symptoms. When using ICs, artefacts (EOG, heart, EMG) are separated in the topomaps space.

Název v anglickém jazyce

EEG data space and feature space distribution

Popis výsledku anglicky

Physiological and pathological processes of the brain may have their characteristic manifestations in the electroencephalographic (EEG) curve. Computer processing offers different views of EEG data. It is possible to examine the continuous amplitude course over time or its individual segments, the scalp electric potential maps, or the spectral space. It is important to investigate the layout of the EEG space in order to design a criterion that would automatically classify the EEG data. The tested EEG data comes from patients with suspected epilepsy, sleep-walking patients and healthy individuals. The data were measured based on the approval of the Bulovka Hospital Ethics Commission and the Ethics Committee of the National Institute of Mental Health. This study uses both linear and nonlinear dimensional reduction techniques to detect the distribution of EEG space. We evaluated the feature space used to classify epileptic data and the space of topographic maps made from the independent components (IC). As the nonlinear dimensional reduction method was used tSNE and nonlinear PCA. The DBSCAN method was used as an auxiliary classifier. Based on our results, we assume nonlinear relationships in the EEG space. Epileptic activity is a separate set of data in the symptom space, on the contrary artifacts (EOG, EMG) are not completely separated by symptoms. When using ICs, artefacts (EOG, heart, EMG) are separated in the topomaps space.

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GA17-20480S" target="_blank" >GA17-20480S: Časový kontext v úloze analýzy dlouhodobého nestacionárního vícerozměrného signálu</a><br>

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2018

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů