Title :
A Method for Solving Overlap Problem in Spike Sorting Based on Genetic Algorithm
Author :
Hu, Su-Rui ; Dai, Min
Author_Institution :
Tianjin Key Lab. of Intell. Comput. & Novel Software Technol., Tianjin Univ. of Technol., Tianjin, China
Abstract :
Overlap decomposition is one of the difficulties in spike sorting. A method based on genetic algorithm is proposed to deal with the overlapped signal from two single waves in this paper. Specifically, with the reliable signal wave templates given, the methods for overlap signals and for single wave signals are used simultaneously to process the unknown data from the signal acquisition equipment, and then the best is selected as the final solution. The method for overlap signals is of genetic algorithm, while the one for single wave signals is of template matching. Genetic algorithm, or even intelligence algorithm is seldom adopted by the past work of dealing with overlapped signals in spike sorting and the simulation experiments indicate that the method put forward in this paper possesses satisfying accuracy and robustness.
Keywords :
biology computing; data mining; feature extraction; genetic algorithms; neural nets; neurophysiology; pattern clustering; signal classification; signal detection; sorting; data mining; feature extraction; genetic algorithm; human nervous system; intelligent algorithm; neuron spike sorting; overlap decomposition; overlapped signal problem; reliable single wave signal template matching; signal acquisition equipment; signal classification technique; signal clustering technique; simulation experiment; spike detection; Computer vision; Educational technology; Genetic algorithms; Genetic engineering; Neurons; Optical recording; Signal processing; Signal processing algorithms; Sorting; Systems engineering education;
Conference_Titel :
Image and Signal Processing, 2009. CISP '09. 2nd International Congress on
Conference_Location :
Tianjin
Print_ISBN :
978-1-4244-4129-7
Electronic_ISBN :
978-1-4244-4131-0
DOI :
10.1109/CISP.2009.5304015
