IEEE transactions on bio-medical engineering
-
IEEE Trans Biomed Eng · Jan 1994
Classification of action potentials in multi-unit intrafascicular recordings using neural network pattern-recognition techniques.
Neural network pattern-recognition techniques were applied to the problem of identifying the sources of action potentials in multi-unit neural recordings made from intrafascicular electrodes implanted in cats. The network was a three-layer connectionist machine that used digitized action potentials as input. ⋯ As the number of units present in the recording increased beyond this limit, the number separable by the network remained roughly constant. The results demonstrate the utility of neural networks for classifying neural activity in multi-unit recordings.
-
Closed loop control of ventilation is traditionally based on end-tidal or mean expired CO2. The controlled variables are the respiratory rate RR and the tidal volume VT. Neither patient size or lung mechanics were considered in previous approaches. ⋯ Furthermore, feasibility tests were done in 6 patients undergoing surgical procedures in total intravenous anesthesia. In the model studies, the controller responded to step changes between 48 seconds and 81 seconds. It did exhibit an overshoot between 5.5% and 7.9% of the setpoint after the step change.(ABSTRACT TRUNCATED AT 250 WORDS)