Fuzzy-Based Expert Supervision System for Feedback Controlled Oxygenation.

Supervision of mechanical ventilation is currently still performed by clinical staff. With the increasing level of automation in the intensive care unit, automatic supervision is becoming necessary. We present a fuzzy-based expert supervision system applicable to automatic feedback control of oxygenation. An adaptive fuzzy limit checking and trend detection algorithm was implemented. A knowledge-based fuzzy logic system combines these outputs into a final score, which subsequently triggers alarms if a critical event is registered. The system was evaluated against annotated experimental data. An accuracy of 83 percent and a precision of 95 percent were achieved. The automatic detection of critical events during feedback control of oxygenation provides an additional layer of safety and assists in alerting clinicians in the case of abnormal behavior of the system. Clinical relevance - Automatic supervision is a necessary feature of physiological feedback systems to make them safer and more reliable in the future.

Identification of dental pain sensation based on cardiorespiratory signals.

The aim of this study is to investigate the feasibility of the detection of brief periods of pain sensation based on cardiorespiratory signals during dental pain triggers. Twenty patients underwent dental treatment and reported their pain events by pressing a push button while ECG, PPG, and thoracic effort signals were simultaneously recorded. Potential pain-indicating features were calculated from the physiological data (sample length of 6 s) and were used for supervised learning of a Random forest pain detector. The best feature combination was determined by Feature forward selection. The best feature combination comprises nine feature groups consisting of four respiratory and five cardiac related groups. The final algorithm achieved a sensitivity of 87% and a specificity of 63% with an AUC of 0.828. Using supervised learning it is possible to train an algorithm to differentiate between short time intervals of pain and no pain solely based on cardiorespiratory signals. An on-site and real-time detection and rating of pain sensations would allow a precise, individuum- and treatment-tailored administration of local anesthesia. Severe phases of pain could be paused or avoided, this would allow more comfortable treatment and yield better patient compliance.

Detection of acute periodontal pain from physiological signals.

OBJECTIVE: To investigate the feasibility of the detection of brief orofacial pain sensations from easily recordable physiological signals by means of machine learning techniques. APPROACH: A total of 47 subjects underwent periodontal probing and indicated each instance of pain perception by means of a push button. Simultaneously, physiological signals were recorded and, subsequently, autonomic indices were computed. By using the autonomic indices as input features of a classifier, a pain indicator based on fusion of the various autonomic mechanisms was achieved. Seven patients were randomly chosen for the test set. The rest of the data were utilized for the validation of several classifiers and feature combinations by applying leave-one-out-cross-validation. MAIN RESULTS: During the validation process the random forest classifier, using frequency spectral bins of the ECG, wavelet level energies of the ECG and PPG, PPG amplitude, and SPI as features, turned out to be the best pain detection algorithm. The final test of this algorithm on the independent test dataset yielded a sensitivity and specificity of 71% and 70%, respectively. SIGNIFICANCE: Based on these results, fusion of autonomic indices by applying machine learning techniques is a promising option for the detection of very brief instances of pain perception, that are not covered by the established indicators.