Ph. D. Project
Title:
Clinical data fusion with deep neural networks for improved diagnosis of sleep apnea syndrome
Dates:
2023/04/07 - 2026/04/06
Supervisor(s): 
Other supervisor(s):
BOUGRAIN Laurent (laurent.bougrain@loria.fr) , GUYOT Pauline (pguyot@noviga.eu)
Description:
Sleep apnoea syndrome is a disorder of nocturnal ventilation characterised by the repetitive and transient occurrence of decreased (hypopnoea) or complete cessation (apnoea) of breathing during sleep. There are two types of syndrome: the obstructive syndrome (which accounts for 90% of cases) and the central syndrome. In both cases, the cessation of breathing causes a decrease in the level of oxygen in the blood and forces the brain to react: the person wakes up to resume breathing without being aware of it.

The goal of this thesis is to integrate deep learning methods in a sleep apnea detection algorithm based on an single ECG signal. Indeed, many works have been published during the last five years [5,6,7,8] proposing a detection of apneic events.The fusion of clinical data in a network, whatever its nature (CNN, LSTM, etc.) is a recent subject and a few articles deal with this issue but none on the subject of sleep apnea. Most of these publications are based on 2D networks with MRI images [11,12,13] or skin images [14].
Department(s): 
Biology, Signals and Systems in Cancer and Neuroscience