Design of a recognition system to predict movement during anesthesia

A Predictive Model of Anesthesia Depth Based on SVM in the Primary Visual Cortex

In this paper, a novel model for predicting anesthesia depth is put forward based on local field potentials (LFPs) in the primary visual cortex (V1 area) of rats. The model is constructed using a Support Vector Machine (SVM) to realize anesthesia depth online prediction and classification. The raw LFP signal was first decomposed into some special scaling components. Among these components, those containing higher frequency information were well suited for more precise analysis of the performance of the anesthetic depth by wavelet transform. Secondly, the characteristics of anesthetized states were extracted by complexity analysis. In addition, two frequency domain parameters were selected. The above extracted features were used as the input vector of the predicting model. Finally, we collected the anesthesia samples from the LFP recordings under the visual stimulus experiments of Long Evans rats. Our results indicate that the predictive model is accurate and computationally fast, and that it is also well suited for online predicting.

Data mining in healthcare and biomedicine: a survey of the literature

As a new concept that emerged in the middle of 1990's, data mining can help researchers gain both novel and deep insights and can facilitate unprecedented understanding of large biomedical datasets. Data mining can uncover new biomedical and healthcare knowledge for clinical and administrative decision making as well as generate scientific hypotheses from large experimental data, clinical databases, and/or biomedical literature. This review first introduces data mining in general (e.g., the background, definition, and process of data mining), discusses the major differences between statistics and data mining and then speaks to the uniqueness of data mining in the biomedical and healthcare fields. A brief summarization of various data mining algorithms used for classification, clustering, and association as well as their respective advantages and drawbacks is also presented. Suggested guidelines on how to use data mining algorithms in each area of classification, clustering, and association are offered along with three examples of how data mining has been used in the healthcare industry. Given the successful application of data mining by health related organizations that has helped to predict health insurance fraud and under-diagnosed patients, and identify and classify at-risk people in terms of health with the goal of reducing healthcare cost, we introduce how data mining technologies (in each area of classification, clustering, and association) have been used for a multitude of purposes, including research in the biomedical and healthcare fields. A discussion of the technologies available to enable the prediction of healthcare costs (including length of hospital stay), disease diagnosis and prognosis, and the discovery of hidden biomedical and healthcare patterns from related databases is offered along with a discussion of the use of data mining to discover such relationships as those between health conditions and a disease, relationships among diseases, and relationships among drugs. The article concludes with a discussion of the problems that hamper the clinical use of data mining by health professionals.

Automated detection of anesthetic depth levels using chaotic features with artificial neural networks

Monitoring the depth of anesthesia (DOA) during surgery is very important in order to avoid patients' interoperative awareness. Since the traditional methods of assessing DOA which involve monitoring the heart rate, pupil size, sweating etc, may vary from patient to patient depending on the type of surgery and the type of drug administered, modern methods based on electroencephalogram (EEG) are preferred. EEG being a nonlinear signal, it is appropriate to use nonlinear chaotic parameters to identify the anesthetic depth levels. This paper discusses an automated detection method of anesthetic depth levels based on EEG recordings using non-linear chaotic features and neural network classifiers. Three nonlinear parameters, namely, correlation dimension (CD), Lyapunov exponent (LE) and Hurst exponent (HE) are used as features and two neural network models, namely, multi-layer perceptron network (feed forward model) and Elman network (feedback model) are used for classification. The neural network models are trained and tested with single and multiple features derived from chaotic parameters and the performances are evaluated in terms of sensitivity, specificity and overall accuracy. It is found from the experimental results that the Lyapunov exponent feature with Elman network yields an overall accuracy of 99% in detecting the anesthetic depth levels.

Classifying depth of anesthesia using EEG features, a comparison

Various EEG features have been used in depth of anesthesia (DOA) studies. The objective of this study was to find the excellent features or combination of them than can discriminate between different anesthesia states. Conducting a clinical study on 22 patients we could define 4 distinct anesthetic states: awake, moderate, general anesthesia, and isoelectric. We examined features that have been used in earlier studies using single-channel EEG signal processing method. The maximum accuracy (99.02%) achieved using approximate entropy as the feature. Some other features could well discriminate a particular state of anesthesia. We could completely classify the patterns by means of 3 features and Bayesian classifier.

EEG signal processing in anaesthesia. Use of a neural network technique for monitoring depth of anaesthesia

BACKGROUND

The Bispectral Index (BIS) is a proprietary index of anaesthesia depth, which is correlated with the level of consciousness and probability of intraoperative recall. The present study investigates the use of a neural network technique to obtain a non-proprietary index of the depth of anaesthesia from the processed EEG data.

METHODS

Two hundred patients, who underwent general abdominal surgery, were recruited for our trial. For anaesthesia we used a total i.v. technique, tracheal intubation, and artificial ventilation. Fourteen EEG variables, including the BIS, were extracted from the EEG, monitored with an EEG computerized monitor, and then stored on a computer. Data from 150 patients were used to train the neural network. All the variables, excluding the BIS, were used as input data in the neural network. The output targets of the network were provided by anaesthesia scores ranging from 10 to 100 assigned by the anaesthesiologist according to the observer's assessment of alertness and sedation (OAA/S) and other clinical means of assessing depth of anaesthesia. Data from the other 50 patients were used to test the model and for statistical analysis.

RESULTS

The artificial neural network was successfully trained to predict an anaesthesia depth index, the NED (neural network evaluated depth), ranging from 0 to 100. The correlation coefficient between the NED and the BIS over the test set was 0.94 (P<0.0001).

CONCLUSION

We have developed a neural network model, which evaluates 13 processed EEG parameters to produce an index of anaesthesia depth, which correlates very well with the BIS during total i.v. anaesthesia with propofol.

Electroencephalogram processing using neural networks

The electroencephalogram (EEG), a highly complex signal, is one of the most common sources of information used to study brain function and neurological disorders. More than 100 current neural network applications dedicated to EEG processing are presented. Works are categorized according to their objective (sleep analysis, monitoring anesthesia depth, brain-computer interface, EEG artifact detection, EEG source-based localization, etc.). Each application involves a specific approach (long-term analysis or short-term EEG segment analysis, real-time or time delayed processing, single or multiple EEG-channel analysis, etc.), for which neural networks were generally successful. The promising performances observed are demonstrative of the efficiency and efficacy of systems developed. This review can aid researchers, clinicians and implementors to understand up-to-date interest in neural network tools for EEG processing. The extended bibliography provides a database to assist in possible new concepts and idea development.

Monitoring anesthesia using neural networks: a survey

New methods of data processing combined with advances in computer technology have revolutionized monitoring of patients under anesthesia. The development of systems based on analysis of brain electrical activity (EEG or evoked potentials) by neural networks has provided impetus to many investigators. Though not claiming to be the end-all in patient monitoring, the potential and efficiency of the combination does indeed stand out. Various strategies are presented and discussed, as well as suggestions for further investigation.

EEG complexity as a measure of depth of anesthesia for patients

A new approach for quantifying the relationship between brain activity patterns and depth of anesthesia (DOA) is presented by analyzing the spatio-temporal patterns in the electroencephalogram (EEG) using Lempel-Ziv complexity analysis. Twenty-seven patients undergoing vascular surgery were studied under general anesthesia with sevoflurane, isoflurane, propofol, or desflurane. The EEG was recorded continuously during the procedure and patients' anesthesia states were assessed according to the responsiveness component of the observer's assessment of alertness/sedation (OAA/S) score. An OAA/S score of zero or one was considered asleep and two or greater was considered awake. Complexity of the EEG was quantitatively estimated by the measure C(n), whose performance in discriminating awake and asleep states was analyzed by statistics for different anesthetic techniques and different patient populations. Compared with other measures, such as approximate entropy, spectral entropy, and median frequency, C(n) not only demonstrates better performance (93% accuracy) across all of the patients, but also is an easier algorithm to implement for real-time use. The study shows that C(n) is a very useful and promising EEG-derived parameter for characterizing the (DOA) under clinical situations.

Derived fuzzy knowledge model for estimating the depth of anesthesia

Reliable and noninvasive monitoring of the depth of anesthesia (DOA) is highly desirable. Based on adaptive network-based fuzzy inference system (ANFIS) modeling, a derived fuzzy knowledge model is proposed for quantitatively estimating the DOA and validate it by 30 experiments using 15 dogs undergoing anesthesia with three different anesthetic regimens (propofol, isoflurane, and halothane). By eliciting fuzzy if-then rules, the model provides a way to address the DOA estimation problem by using electroencephalogram-derived parameters. The parameters include two new measures (complexity and regularity) extracted by nonlinear quantitative analyses, as well as spectral entropy. The model demonstrates good performance in discriminating awake and asleep states for three common anesthetic regimens (accuracy 90.3 % for propofol, 92.7 % for isoflurane, and 89.1% for halothane), real-time feasibility, and generalization ability (accuracy 85.9% across the three regimens). The proposed fuzzy knowledge model is a promising candidate as an effective tool for continuous assessment of the DOA.

Depth of anesthesia estimation and control

A fully automated system was developed for the depth of anesthesia estimation and control with the intravenous anesthetic, Propofol. The system determines the anesthesia depth by assessing the characteristics of the mid-latency auditory evoked potentials (MLAEP). The discrete time wavelet transformation was used for compacting the MLAEP which localizes the time and the frequency of the waveform. Feature reduction utilizing step discriminant analysis selected those wavelet coefficients which best distinguish the waveforms of those responders from the nonresponders. A total of four features chosen by such analysis coupled with the Propofol effect-site concentration were used to train a four-layer artificial neural network for classifying between the responders and the nonresponders. The Propofol is delivered by a mechanical syringe infusion pump controlled by Stanpump which also estimates the Propofol effect-site and plasma concentrations using a three-compartment pharmacokinetic model with the Tackley parameter set. In the animal experiments on dogs, the system achieved a 89.2% accuracy rate for classifying anesthesia depth. This result was further improved when running in real-time with a confidence level estimator which evaluates the reliability of each neural network output. The anesthesia level is adjusted by scheduled incrementation and a fuzzy-logic based controller which assesses the mean arterial pressure and/or the heart rate for decrementation as necessary. Various safety mechanisms are implemented to safeguard the patient from erratic controller actions caused by external disturbances. This system completed with a friendly interface has shown satisfactory performance in estimating and controlling the depth of anesthesia.

Monitoring of the central nervous system

Clinical studies have shown a close relationship between variables such as hypoxia, increased intracranial pressure, arterial hypotension, or seizures and neurological outcome. This indicates the need for monitoring techniques of the central nervous system including measurements of cerebral blood flow, cerebral oxygenation and neuronal function. Semiquantitative changes in cerebral blood flow can be measured continuously using transcranial Doppler sonography. Measurements of jugular venous oxygen saturation or tissue oxygenation reflect the balance between cerebral oxygen delivery and cerebral oxygen demand. Near-infrared spectroscopy appears to be a technology with potential for non-invasive measurements of cerebral oxygen saturation and mitochondrial oxygen availability. The current technology is, however, of limited clinical utility. Brain electrical monitoring techniques such as electroencephalogram and evoked potentials are sensitive and specific to detect changes in neuronal function caused by cerebral ischaemia. Electroencephalogram and evoked potential measurements of depth of anaesthesia and specific electroencephalogram patterns for pharmacodynamic quantification of drug effects may gear the dosage of anaesthetics according to the anaesthetic effect.

An expert system for EEG monitoring in the pediatric intensive care unit

OBJECTIVES

was to design a warning system for the pediatric intensive care unit (PICU). The system should be able to make statements at regular intervals about the level of abnormality of the EEG. The warnings are aimed at alerting an expert that the EEG may be abnormal and needs to be examined.

METHODS

A total of 188 EEG sections lasting 6 h each were obtained from 74 patients in the PICU. Features were extracted from these EEGs, and with the use of fuzzy logic and neural networks, we designed an expert system capable of imitating a trained EEGer in providing an overall judgment of abnormality about the EEG. The 188 sections were used in training and testing the system using the rotation method, thus separating training and testing data.

RESULTS

The EEGer and the expert system classified the EEGs in 7 levels of abnormality. There was concordance between the two in 45% of cases. The expert system was within one abnormality level of the EEGer in 91% of cases and within two levels in 97%.

CONCLUSIONS

We were therefore able to design a system capable of providing reliably an assessment of the level of abnormality of a 6 h section of EEG. This system was validated with a large data set, and could prove useful as a warning device during long-term ICU monitoring to alert a neurophysiologist that an EEG requires attention.