English
 
User Manual Privacy Policy Disclaimer Contact us
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Poster

Spike sorting and detection with optimal multichannel filters

MPS-Authors
/persons/resource/persons84098

Munk,  MHJ
Department Physiology of Cognitive Processes, Max Planck Institute for Biological Cybernetics, Max Planck Society;
Max Planck Institute for Biological Cybernetics, Max Planck Society;

Fulltext (public)
There are no public fulltexts stored in PuRe
Supplementary Material (public)
There is no public supplementary material available
Citation

Franke, F., Natora, M., Boucsein, C., Munk, M., & Obermayer, K. (2008). Spike sorting and detection with optimal multichannel filters. Poster presented at 38th Annual Meeting of the Society for Neuroscience (Neuroscience 2008), Washington, DC, USA.


Cite as: http://hdl.handle.net/21.11116/0000-0003-8B6F-5
Abstract
Multi-electrode extracellular unit recordings have been extensively utilized for the investigation of mechanisms of neural information processing. The obtained signals usually contain action potentials from multiple cells which need to be separated by spike sorting. This is especially difficult in the case of overlapping spike waveforms. The ability to discriminate spikes from different neurons is greatly influenced by the choice of recording position. Thus, it is desirable to know the quality of spike sorting already during the experiment to allow for detection of favorable recording positions. Linear filtering with optimal multichannel filters [1,2] is a promising approach to spike sorting. For every recorded neuron - and its specific waveform - an optimal multichannel filter (OMF) is constructed by solving a constraint optimization problem. The output energy of the OMF to the correct waveform is constrained to one while minimizing the output energy for noise and waveforms of other neurons. The task of spike sorting is then reduced to a simple detection of high peaks in the filter output. Here we apply OMFs to spike sorting and present a fully autonomous online algorithm. From the generative model approach of the filter construction, a theoretical measure for electrode recording position quality is derived. The algorithm consists of three main building blocks: (1) A general spike detection, which aims at detecting every spike. (2) Clustering of spikes which were not detected by the OMFs to initialize new OMFs. (3) Spike detection and sorting with learned OMFs and adaptation of the OMFs. We evaluated the detection and sorting performance of spike sorting with OMFs in respect to false positives (FP) and false negatives (FN) on simultaneous intra and extracellular recordings and on simulated data. On simulated data we systematically varied the signal to noise ratio (SNR) and the percentage of overlapping spikes. We show that the number of FNs and FPs can be significantly decreased (up to 40% under low SNR) using OMFs compared to existing spike detection algorithms. Furthermore, up to 50% of all spikes being overlaps they can be successfully disentangled since both the filter operation and the superposition of spikes are linear. We conclude that OMFs have great advantages for spike sorting and detection and could be used for real time spike sorting and automated electrode positioning systems.