C. Canedo Ortega, A. Fernandez Santamónica, I. Fernández Martínez, Y. Larriba, C. Rueda Sabater
The problem of predicting the neurological recovery of patients after cardiac arrest from their EEG signals will be addressed. The analyzed EEGs were recorded with 18 channels during the next 72 hours after the cardiac arrest. These signals belong to a sample of nearly 1000 comatose patients that the ICARE consortium has made public this year 2023.
Our proposal considers the decomposition of the EEG segments into FMM waves. The parameters of each wave describe its location and sharpness simultaneously in all channels, as well as its amplitude and symmetry in each channel. From these parameters, features are created using statistics that collect the fundamental aspects of an EEG, resulting in an interpretable and useful alternative to identify EEGs from patients with a poor prognosis. In the final stage, other clinical and demographic variables are also considered.
This approach has been developed by our group as part of the George B. Moody PhysioNet Challenge 2023.
Keywords: EEG, Cardiac arrest, Frequency Modulated Mobius.
Scheduled
Data Analytics
November 10, 2023 4:00 PM
HC2: Canónigos Room 2