Blind source separation for spike sorting of high density microelectrode array recordings

High-density microelectrode arrays (HD-MEAs) with large numbers of densely packed electrodes potentially allow for recording from every cell on the array and generate large, redundant datasets. Blind-source-separation algorithms (BSS), used to separate mixtures of independent sources into the original signals, are an ideal means to be applied to the spike sorting of HD-MEA recordings. We show that recorded neuronal signals represent convoluted mixtures, and we present a BSS algorithm. The algorithm uses the nonlinear energy operator as preprocessor and an extended method of independent-component analysis to separate convoluted mixtures. The algorithm is applied to recordings from retinal ganglion cells, and its performance is evaluated.

[1]  Andreas Hierlemann,et al.  Switch-Matrix-Based High-Density Microelectrode Array in CMOS Technology , 2010, IEEE Journal of Solid-State Circuits.

[2]  Kari Torkkola,et al.  Blind separation of convolved sources based on information maximization , 1996, Neural Networks for Signal Processing VI. Proceedings of the 1996 IEEE Signal Processing Society Workshop.

[3]  J. F. Kaiser,et al.  On a simple algorithm to calculate the 'energy' of a signal , 1990, International Conference on Acoustics, Speech, and Signal Processing.

[4]  U. Frey,et al.  Microelectronic system for high-resolution mapping of extracellular electric fields applied to brain slices. , 2009, Biosensors & bioelectronics.

[5]  B. McNaughton,et al.  Tetrodes markedly improve the reliability and yield of multiple single-unit isolation from multi-unit recordings in cat striate cortex , 1995, Journal of Neuroscience Methods.

[6]  J. Csicsvari,et al.  Accuracy of tetrode spike separation as determined by simultaneous intracellular and extracellular measurements. , 2000, Journal of neurophysiology.

[7]  Sung June Kim,et al.  Neural spike sorting under nearly 0-dB signal-to-noise ratio using nonlinear energy operator and artificial neural-network classifier , 2000, IEEE Transactions on Biomedical Engineering.

[8]  Erkki Oja,et al.  Independent component analysis: algorithms and applications , 2000, Neural Networks.

[9]  Zbynek Koldovsky,et al.  Time-Domain Blind Separation of Audio Sources on the Basis of a Complete ICA Decomposition of an Observation Space , 2011, IEEE Transactions on Audio, Speech, and Language Processing.

[10]  S. Mukhopadhyay,et al.  A new interpretation of nonlinear energy operator and its efficacy in spike detection , 1998, IEEE Transactions on Biomedical Engineering.