Back to search

Aristotle University Of Thessaloniki (AUTH)

Ανάλυση σύνθεση και εποπτική αναλυτική πολυδιάστατων δεδομένων πολλαπλών πηγών με σκοπό τη βελτιστοποίηση αναγνώρισης μοτίβων και αξιολόγησης βιοδεικτών που σχετίζονται με τον ύπνο

Abstract

dc:description

In this thesis, a series of methodologies and computational tools were developed for the analysis of polysomnographic data with the aim of identifying and evaluating patterns and biomarkers related to sleep quality and supporting sleep specialists. Initially, appropriate methods were developed for the pre-processing and extraction of features from multichannel signals recorded from different sources: electroencephalogram and electrocardiogram. Subsequently, artificial intelligence models were trained to achieve high classification accuracy, combining the characteristics of biological signals and approaches from the fields of biological signal analysis, image processing (transformation of the extracted features into a data structure similar to that of the image) and deep learning. For automatic sleep stagινγ from biological signal recordings, it is necessary to extract features that can encode the changes observed in the signal waveforms due to changes in body functions. These features can be calculated as functional connectivity metrics between EEG electrodes or brain regions after an appropriate transformation, which can be considered as a graph adjacency matrix through which graph theory can be used to calculate graph metrics. To support experts, the integrated graphical user interface SmartHypnos eXpert was developed for the MATLAB programming environment. Through this interface, the user can select the input file whose signals are displayed in the interface. For each 30-second epoch, the expert can select the sleep stage to which it belongs, while at the same time a prediction is available from an artificial intelligence method, thus reducing the staging time. Respiratory event recognition is also supported. After the staging is complete, it is possible to save it in different file formats which can be loaded for any corrections or further analysis in the same or a different software tool. A summary of the staging and recognition of respiratory events is presented in a different tab of the environment in the form of graphs. This report can also be saved in different file formats. Finally, in another tab, the data from the Android application can be graphically presented. The SmartHypnos eXpert interface was evaluated by 11 sleep specialists with positive results. The SmartHypnos application for Android devices is targeted at a general audience. This application presents data related to the person's physical activity during the day and the subjective sleep quality. In this application, there is the possibility of completing questionnaires to assess sleep quality and stress levels. The combination of recorded data allows the production of personalized suggestions aimed at improving sleep quality and efficiency, which are generated through a decision tree, developed with the help of experts. All data is recorded on the user's device without the need for an internet connection. This application was evaluated by 48 users during a one-week pilot test with positive results. The methods and tools developed are expected to contribute to scientific research and clinical practice, particularly facilitating the work of sleep specialists.

Degree

thesis:*
Grantor dc:publisher
Aristotle University Of Thessaloniki (AUTH)
Year dc:date
2025

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Chriskos, Panteleimon
  • Χρίσκος, Παντελεήμων

Subjects

dc:subject × 14

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/59034

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Chriskos, Panteleimon; Χρίσκος, Παντελεήμων. Ανάλυση σύνθεση και εποπτική αναλυτική πολυδιάστατων δεδομένων πολλαπλών πηγών με σκοπό τη βελτιστοποίηση αναγνώρισης μοτίβων και αξιολόγησης βιοδεικτών που σχετίζονται με τον ύπνο. Aristotle University Of Thessaloniki (AUTH), 2025. http://hdl.handle.net/10442/hedi/59034