Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference
-
Annu Int Conf IEEE Eng Med Biol Soc · Jul 2019
Graph Theoretical Analysis of Cortical Networks based on Conscious Experience.
The aim of the study was to investigate differences in cortical networks based on the state of consciousness. Five subjects performed a serial-awakening paradigm with electroencephalography (EEG) recordings. We considered four states of consciousness: (1) non-rapid eye movement (NREM) sleep with no conscious experience, (2) NREM sleep with conscious experience, (3) rapid eye movement (REM) sleep with conscious experience, and (4) wakefulness. ⋯ In the beta band, functional integration in wakefulness was higher than in NREM sleep. Regarding functional segregation, in the theta band, transitivity and clustering coefficient in NREM sleep with no conscious experience were stronger than in wakefulness or REM sleep, but clustering in the beta band showed an opposite effect. The observed differences may be related to cortical bistability and add to previously observed neural correlates of consciousness.
-
Annu Int Conf IEEE Eng Med Biol Soc · Jul 2019
MODELHealth: Facilitating Machine Learning on Big Health Data Networks.
MODELHealth is a platform that aims to facilitate the implementation of Machine Learning (ML) techniques on medical data in order to upgrade the delivery of healthcare services. MODELHealth platform is a "holistic" approach to the implementation of processes for the development and utilization of ML algorithms in many forms, including Neural Networks, and can be used to assist clinical work and administrative decision-making. It covers the entire lifecycle of these processes, from pumping, homogenization, anonymization, and enrichment of the initial data, to the final disposal of efficient algorithms through Application Program Interfaces for consumption by any authorized Information System.
-
Annu Int Conf IEEE Eng Med Biol Soc · Jul 2019
High oxygen fraction during airway opening is key to effective airway rescue in obese subjects.
Apnea is common after induction of anesthesia and may produce dangerous hypoxemia, particularly in obese subjects. Optimal management of airway emergencies in obese, apneic subjects is complex and controversial, and clinical studies of rescue strategies are inherently difficult and ethically-challenging to perform. We investigated rescue strategies in various degrees of obesity, using a highly-integrated, computational model of the pulmonary and cardiovascular systems, configured against data from 8 virtual subjects (body mass index [BMI] 24-57 kg m-2). ⋯ Rescue using tidal ventilation with 100% oxygen provided rapid re-oxygenation in all subjects, even with small tidal volumes in subjects with large BMI. Overall, subjects with larger BMI pre-oxygenated faster and, after airway obstruction, developed hypoxemia more quickly. Our results indicate that attempts to achieve substantial tidal volumes during airway rescues are probably not worthwhile (and may be counter-productive); rather, it is the assurance of a high-inspired oxygen fraction that will prevent critical hypoxemia.
-
Annu Int Conf IEEE Eng Med Biol Soc · Jul 2019
BioTranslator: Inferring R-Peaks from Ambulatory Wrist-Worn PPG Signal.
Advancements in wireless sensor networks (WSN) technology and miniaturization of wearable sensors have enabled long-term continuous pervasive biomedical signal monitoring. Wrist-worn photoplethysmography (PPG) sensors have gained popularity given their form factor. However the signal quality suffers due to motion artifacts when used in ambulatory settings, making vital parameter estimation a challenging task. ⋯ Using one-dimensional Convolution-Deconvolution Network, we translate a single channel PPG signal to an electrocardiogram(ECG)-like time series signal, from which relevant R-peak information can be inferred enabling IHR measures. The proposed network configuration was evaluated on 12 subjects of the TROIKA dataset, involved in physical activity. The proposed network identifies 92.8% of R-peaks, besides achieving a mean absolute error of 51±6.3ms with respect to reference ECG-derived IHR.
-
Annu Int Conf IEEE Eng Med Biol Soc · Jul 2019
A Fast Brain Switch Based on Multi-Class Code-Modulated VEPs.
To realize asynchronous control of a brain-computer interface (BCI) system, a fast brain switch with low false positive rate (FPR) is required. This paper proposed a brain switch based on code-modulated visual-evoked potential (c-VEP), in which seven 8-bit pseudorandom codes were used to modulate the electroencephalogram (EEG) signal. This study optimized and demonstrated the control strategy through an offline and an online experiments. By decoding the brain state continuously with the task-related component analysis (TRCA) algorithm, the brain switch achieved an average reaction time (RT) of 1.72 seconds and an average idle time of 183.53 seconds without false positive events in the online experiment.