Abstract-Monitoring the levels of sedation-analgesia may be helpful for managing patient stress on minimally invasive medical procedures. Monitors based on EEG analysis and designed to assess general anesthesia cannot distinguish reliably between a light and deep sedation. In this work, the Poincaré plot is used as a nonlinear technique applied to EEG signals in order to characterize the levels of sedation-analgesia, according to observed categorical responses that were evaluated by means of Ramsay Sedation Scale (RSS). To study the effect of high frequencies due to EMG activity, three different frequency ranges (FR1=0.5-110 Hz, FR2=0.5-30 Hz and FR3=30-110 Hz) were considered. Indexes from power spectral analysis and plasma concentration of propofol and remifentanil were also compared with the bispectral index BIS. An adaptive Neurofuzzy Inference System was applied to model the interaction of the best indexes with respect to RSS score for each analysis, and leave-one-out cross validation method was used. The ability of the indexes to describe the level of sedation-analgesia, according with the RSS score, was evaluated using the prediction probability (Pk). The results showed that the ratio SD1/SD2FR3 contains useful information about the sedation level, and SD1FR2 and SD2FR2 had the best performance classifying response to noxious stimuli. Models including parameters from Poincaré plot emerge as a good estimator of sedation-analgesia levels.
The refined multiscale entropy (RMSE) approach is commonly applied to assess complexity as a function of the time scale. RMSE is normally based on the computation of sample entropy (SampEn) estimating complexity as conditional entropy. However, SampEn is dependent on the length and standard deviation of the data. Recently, fuzzy entropy (FuzEn) has been proposed, including several refinements, as an alternative to counteract these limitations. In this work, FuzEn, translated FuzEn (TFuzEn), translated-reflected FuzEn (TRFuzEn), inherent FuzEn (IFuzEn), and inherent translated FuzEn (ITFuzEn) were exploited as entropy-based measures in the computation of RMSE and their performance was compared to that of SampEn. FuzEn metrics were applied to synthetic time series of different lengths to evaluate the consistency of the different approaches. In addition, electroencephalograms of patients under sedation-analgesia procedure were analyzed based on the patient’s response after the application of painful stimulation, such as nail bed compression or endoscopy tube insertion. Significant differences in FuzEn metrics were observed over simulations and real data as a function of the data length and the pain responses. Findings indicated that FuzEn, when exploited in RMSE applications, showed similar behavior to SampEn in long series, but its consistency was better than that of SampEn in short series both over simulations and real data. Conversely, its variants should be utilized with more caution, especially whether processes exhibit an important deterministic component and/or in nociception prediction at long scales.
El gráfico de Poincaré es un método no lineal que permite analizar la variabilidad de una señal, dibujando dicha señal contra ella misma retardada en el tiempo. Aunque este método ha sido aplicado al análisis del electroencefalograma (EEG) para la medición de niveles de anestesia general, su aplicación en la medición de niveles de anestesia ligera aún no se ha abordado con profundidad. En este trabajo, se consideran varios tiempos de retardo en el gráfico de Poincaré para caracterizar el nivel de anestesia ligera en 110 pacientes sometidos a endoscopia digestiva. Los índices de Poincaré son combinados con parámetros espectrales de señales que contienen información del EEG y del EMG (electromiografía) para modelar los niveles de anestesia ligera, teniendo como referencia la escala de sedación Ramsay. La mejor predicción, utilizando índices del gráfico de Poincaré, del nivel de sedación y analgesia (Pk = 0,711) se obtiene usando un retardo de 8 ms en el rango de frecuencia de 30-110 Hz y de 6 ms en el rango de frecuencia de 0,5 -30 Hz, respectivamente. La combinación de diferentes parámetros permitió obtener valores de Pk = 0,799.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.