Recently, Hidden Markov Models (HMM) have been used to reduce the dimensionality of complex and high-dimensional data. In particular, it is known that neural activity recorded through electroencephalogram (EEG) displays low-dimensional global patterns of coordinated activity, termed "microstates". In this thesis we will first review the framework of HMM to then apply it on EEG data recorded from healthy subjects on 64 channels at rest (and during task). Finally, we will characterize the statistics on the low dimensional space, such as the average duration, the frequency of occurrence and the transition probabilities.
Recentemente, gli Hidden Markov Models (HMM) sono stati usati per ridurre le dimensioni di dati complessi e ad alta dimensionalità. In particolare, è noto che l'attività neurale registrata tramite elettroencefalogramma (EEG) presenta pattern globali di attività coordinate a bassa dimensionalità, chiamati "microstati". In questa tesi esamineremo prima il framework degli HMM per poi applicarlo ai dati EEG registrati in soggetti sani su 64 canali a riposo (e durante lo svolgimento di un compito). Infine, caratterizzeremo la statistica dello spazio a bassa dimensionalità, come la durata media, la frequenza di occorrenza e le probabilità di transizione.
Hidden Markov Models for dimensionality reduction of neural activity
BEDIN, VERONICA
2021/2022
Abstract
Recently, Hidden Markov Models (HMM) have been used to reduce the dimensionality of complex and high-dimensional data. In particular, it is known that neural activity recorded through electroencephalogram (EEG) displays low-dimensional global patterns of coordinated activity, termed "microstates". In this thesis we will first review the framework of HMM to then apply it on EEG data recorded from healthy subjects on 64 channels at rest (and during task). Finally, we will characterize the statistics on the low dimensional space, such as the average duration, the frequency of occurrence and the transition probabilities.File | Dimensione | Formato | |
---|---|---|---|
Bedin_Veronica.pdf
accesso aperto
Dimensione
2.45 MB
Formato
Adobe PDF
|
2.45 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/41577