Classification of action potentials in multi-unit intrafascicular recordings using neural network pattern recognition techniques

Neural network pattern recognition techniques were applied to classify action potentials in multi-unit neural recordings made from chronically and acutely implanted intrafascicular electrodes in cats. The success of unit potential classification with the neural network was compared to that with a previously described template system. The neural network reliably separated 89 of the 194 recorded units, while the template system only separated 31 of the units. The network was able to reliably separate 6 or 7 units per recording, on average. The results demonstrate the potential superiority of neural networks over template matching approaches to classification of neural activity in multi-unit recordings.