Innovative techniques for the analysis of clinical signals
External reference persons - Reparto di Cardiologia Universitaria dell'Ospedale "Molinette" di Torino
- Reserach Center Vall D'Hebron (Barcellona, Spagna)
Thesis type THEORETICAL/EXPERIMENTAL
Description In recent years, medicine is revolutionized by the possibility of measuring in-vivo fundamental variables for the prevention and early diagnosis of serious diseases. These are measures over time or space-time, with different temporal/spatial resolution. These data have an enormous clinical potential; however, it is necessary to develop novel techniques for their analysis and to extract all the information useful for the clinicians.
The theses proposed here fall within this field and concern the following medical quantities:
- hemodynamic measurements (flow field) in the aortic arch obtained from high intensity magnetic resonance; the measures are on healthy subjects and on patients suffering from pathologies (Marfan's syndrome, valvular dysfunctions, etc.)
- hemodynamic and continuous pressure measurements over long periods of time (24 hours). Also in this case, large data cohorts are available on healthy subjects and sick subjects
- measures of heartbeat and pressure during paroxysmal atrial fibrillation attacks
- near infrared spectroscopy (NIRS) measurements on the brain, both in superficial areas and in deep intracranial areas; the measures concern patients before and after the cardioversion procedure.
The basic idea of the theses proposed here is to develop innovative methods of analysis based on the recent theory of complex networks. We want to transform time series (or spatiotemporal series) into (mono- or multi-layer) complex networks and, then, using (appropriately modified and tailored) analysis methods developed in the field of complex network theory. From a formal point of view, the transformation of signals into complex networks corresponds to converting the signals into geometric objects, whose (local and global) geometrical characteristics embed particular aspects (temporal or spatial-temporal) of the original signals. In this way, it is possible to reveal the peculiarity and properties of the signals that would be very difficult to understand while remaining in the temporal or spatial-temporal dimension.
Required skills Please, contact supervisors
Deadline 01/06/2021 PROPONI LA TUA CANDIDATURA