Ph.D. Thesis Proposal - Jacopo Pinccini
Candidate: Jacopo Pinccini
Date and Time: January 13, 14:00 - 15:00 room M110
Ph.D. Thesis Proposal - Jacopo Piccini
Supervisor: Eliahu August, Reykjavík University
Committee:
María Óskarddóttir, Reykjavík University
Erna Sif Arnardóttir, Reykjavík University
Rajdeep Nath, University of Eastern Finland
Abstract:
Abstract: Electrodermal activity (EDA) is one of the longest-known and most accessible physiological signals. EDA reflects changes in skin potential due to sweating, which, during sleep, has a thermoregulatory function. Eccrine sweat glands, the sweat glands activated during sleep, are innervated by the sympathetic nervous system (SNS) only, with no parasympathetic input. Despite this direct connection between EDA and the SNS during the night, the signal has been so far used in studies of diurnal phenomena. One of the main reasons for neglecting EDA in sleep studies is the complexity of the recorded signals. Long-term EDA recordings are susceptible to noise from various sources causing signal artefacts. Despite this issue, EDA remains one of the most information-rich and accessible signals. The broader goal of this thesis is to design methods to both allow the use of EDA for clinical purposes and to gain insight into the mechanisms governing EDA dynamics during sleep.
In the first part of this thesis, we developed an algorithm to identify the relevant phenomena in the EDA signals. We designed it using traditional signal processing techniques rather than machine learning (ML) methods. Secondly, we used supervised ML algorithms to perform sleep staging and detect sleep-breathing disorders such as obstructive sleep apnea (OSA) solely based on the EDA signal.
Microsoft Teams meeting
Join on your computer, mobile app or room device
Click here to join the meeting
Meeting ID: 331 290 590 88
Passcode: JJeXit
