When recording action potentials (spikes) from many neurons simultaneously via multichannel micro-electrodes the overlapping of spikes from different neurons is a demanding problem for detection and classifi-cation of spikes (spike sorting). Since multichannel electrodes provide better possibilities to separate the superimposed waveforms, we refined an algorithm for separation of overlapping spikes for the use on multichannel recordings and tested it on simulated data with different numbers of signal channels and with several signal parameters. We show that the larger the number of signal channels the better the separation that may be achieved, especially under demanding recording conditions.

Multi-neuron extracellular recordings with multichannel microelectrodes, like tetrodes or heptodes, are used to get insights in the neuronal information processing in the brain. Spike sorting is the detection and classification of extracellular action potentials (spikes) recorded with these electrodes. Here we report on the improvement of spike sorting performance when using heptodes rather than tetrodes by simulating tetrode/heptode recordings and analysing them automatically with our spike sorting algorithm. Simulation of signals is performed via a simple model of neurons and electrodes that has the ability to emulate preferably realistic recording conditions without too much complexity. Results indicate that spike sorting performance is better when heptodes instead of tetrodes are used to record from highly active local clusters consisting of many neurons and when the signals have a low signal-to-noise ratio.

Extracellular recordings of several neurons simultaneously with multichannel electrodes, like tetrodes or heptodes, are frequently used to explore the information processing in the brain. The main goal is then to determine the times of the extracellular recorded action potentials (spikes) and to classify them into groups, assuming that the spikes of one group originate from the same neuron. This process of detection and classification of spikes is called spike sorting. The main part of this work is the examination of the spike sorting improvement by using multichannel electrodes with more channels. To achieve this, we simulate heptode recordings, analyse them with our spike sorting algorithm and compare the results to recordings with fewer channels. Our results indicate that the spike sorting performance increases, especially under suboptimal signal conditions, when more channels are used.

A method for calculating the instantaneous energy of extracellularly recorded action potentials using the analytic signal has been developed and tested. The energy is computed from the recorded signal with Hilbert filter based multi-resolution energy filters with optimized signal to noise ratio. The computed instantaneous energy is used for threshold based detection of action potentials in the signals.