Sveikas žmogus užmiega vidutiniškai per 15 min, o miegas prasideda lėtojo miego fazėje. Patologinis užmigimas yra diagnozuojamas narkolepsija sergantiems pacientams, kurie užmiega paradoksinio miego fazėje. Šiam miego sutrikimui diagnozuoti yra atliekamas - Pakartotinas Miego Latencijos Testas (anlg. Multiply sleep latency test (MSLT)). Testas atliekamas tam pačiam pacientui keturis kartus per dieną. Vertinamos keturios užmigimo trukmės, analizuojant registruojamą paciento žievės bioelektrinį aktyvumą polisomnografiniu tyrimu miego laboratorijoje bei vertinant miego latenciją. Miegas ir epilepsija yra susiję. Miego metu provokuojami priepuoliai, tuo tarpu priepuoliai įtakoja miego vientisumą, jo fazes. Kai kurių epilepsijos formų (idiopatinės židininės, infantilinių spazmų ir kt.) epilepsiniai priepuoliai aktyvuojasi užmingant, kitų (juvenilinės miokloninės)- prabundant. Miego pradžios nustatymas yra reikalingas epilepsija sergančių ligonių miego struktūros analizei, ankstyvųjų epilepsinių iškrūvių atpažinimui bei tiksliai narkolepsijos diagnostikai. Todėl klinikinėje praktikoje yra reikalingas automatinis diagnostinis metodas, kuris padėtų objektyviai, greitai ir tiksliai įvertinti miego pradžią. Mūsų tyrimo tikslas - įvertinti integruotos amplitudės EEG (aEEG) metodo patikimumą nustatant miego pradžią narkolepsija ir epilepsija sergantiems pacientams Tyrime dalyvavo Kopenhagos universitetinėje ligoninėje gydyti 25 narkolepsija sergantys bei 23 Vilniaus universiteto vaikų... [toliau žr. visą tekstą] / The onset of sleep under normal circumstances in young adult humans is normally through Non-REM sleep after 15 minutes of being awake. The abnormal entry into sleep through REM sleep can be diagnostic in patients with narcolepsy. One important investigation in diagnosing narcolepsy is the Multiple Sleep Latency Test (MSLT), where onset of sleep and occurrence of REM – if any is measured 4 times per day. This test is traditionally scored by visual analysis of sleep onset and onset of REM sleep. There is tight relation between sleep and epilepsy. During sleep epileptiform activity is frequent and seizures make influence on the quality of sleep ant it’s phases. In some types of epilepsy (e.g. idiopathic focal) seizures occur during onset of sleep, in others (juvenile myoclonic) – short time after awaking. Detection of sleep onset is very important for the evaluation of the sleep quality of epilepsy patient’s and for accurate diagnosis of narcolepsy. EEG interpretation strongly depends on the skills of the EEG reader. Therefore automatic sleep onset detection could be useful diagnostic tool for EEG interpretation. We have found it of interest to investigate if an automatic analysis amplitude – integrated electroencephalography (aEEG) method could reliable identify sleep onset by comparing it with visual analysis in 25 narcolepsy patients, who were treated in Copenhagen University hospital and 23 children with epilepsy (idiopathic focal and juvenile myoclonic), who were treated in... [to full text]

Introduction: Assessing brain maturation in preterm neonates is essential for the health of the neonates. Machine learning methods have been introduced as a prospective assessment tool for neonatal electroencephalogram(EEG) signals. Explainable methods are essential in the medical field, and more research regarding explainability is needed in the field of using machine learning for neonatal EEG analysis. Methodology: This thesis develops an explainable machine learning model that estimates postmenstrual age in very preterm neonates from EEG signals and investigates the importance of the features used in the model. Dual-channel EEG signals had been collected from 14 healthy preterm neonates of postmenstrual age spanning 25 to 32 weeks. The signals were converted to amplitude-integrated EEG (aEEG) and a list of features was extracted from the signals. A regression tree model was developed and the feature importance of the model was assessed using permutation importance and Shapley additive explanations. Results: The model had an RMSE of 1.73 weeks (R2=0.45, PCC=0.676). The best feature was the mean amplitude of the lower envelope of the signal, followed by signal time spent over 100 µV. Conclusion: The model is performing comparably to human experts, and as it can be improved in multiple ways, this result indicates a promising outlook for explainable machine learning model applications in neonatal EEG analysis.

No description available.