Answers are grounded in this paper and cite their sources.
In this paper present the main existing methods and means of determining the depth of anesthesia. Among the methods considered, the following groups were distinguished: analysis of electroencephalograms (EEG), analysis of electrocardiograms (ECG), and complex analysis of biosignals. The results obtained in previous studies are systematized and prospective directions for further research are determined, in particular the use of machine learning tools for the analysis of electrocardiograms and their characteristics and training of deep neural networks.In order to study the capabilities of neural networks as systems for classifying biosignals, a deep convolution neural network was constructed to classify ECG signals with the presence or absence of arrhythmias. Input data for those network were 30-second ECG signals without preliminary processing, and the classification accuracy was 68.3%. The results of such a classification system indicate the advisability of using deep training to determine arrhythmias, and the need for preprocessing ECG and pre-identified characteristics, among which can be the magnitude of the signal energy in different frequency ranges, the signal after the removal of the trend and the signal parameters in the time domain.Current studies in the field of determining the depth of anesthesia by analyzing EEG and ECG signals make not only the technical limitations of currently used technologies, but also the fact that the number of anesthetics having a unique effect on the brain and its signals is constantly increasing.Among the technical constraints, the sensitivity of sedimentation depth to noise is the most important and the unequal response to anesthetics of different types. In this direction, many studies are devoted to improving the accuracy of determining the depth of the nerve system, by analyzing the EEG, EEG and ECG signals, only the ECG, and through the comprehensive analysis of the signals that can be obtained from the patient. In many studies, the authors conclude that alternative methods for determining the depth of a patient's sedation have better results, but at the same time they are insufficiently tested or have other technical limitations.A promising technology for determining the depth of nerve sedation via ECG analysis may also be the use of neural networks, since such systems are capable of training on labeled data, they can classify data in conditions of unknown patterns and are resistant to noise in incoming signals.Simultaneously with the use of neural networks perspective, one can use the methods of their analysis, as classification systems, in order to identify the best configurations of their parameters. Systems for the analysis of classification systems can serve both for selecting the best of them and for their adaptive combination in order to improve the classification result.Ref. 54, tabl. 1.
No takes yet. Share an insight, caveat, or question.
Fedorchuk et al. (2018) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: