Novel Analysis Methods For Multi-channel Spike Signals And Their Applications In The Investigation Of Neuronal Firing Features In Rat Hippocampus

Since the development of micro-electrode arrays for extracellular recording of neuronal spikes and local field potentials, there comes up a lot of signal processing problems, such as detecting and sorting of neuronal extracellular action potentials (i.e. spikes), which are the basis of study on neural system encoding and processing mechanisms.In this dissertation, micro-electrode array with high density recording sites was used to record extracellular spikes and local field potentials in rat hippocampal CA1 areas. The main work and analyzing methods are list as follows:(1) To improve spike detecting and sorting results, a multi-channel threshold spike detecting method was designed based on the multi-channel recording signals; Then a spike detecting method based on principal component analysis (PCA) and nonlinear energy operator (NEO) was designed for signals with low signal-to-noise ratio (SNR); After detection, a spike sorting method based on multi-channel spike waveforms were designed. (2) In order to investigate the intrinsic relation of neuronal spike firing and local field potentials in CA1 area of rat hippocampus, the new designed detecting and sorting methods were applied to the analysis of signals recorded before and under tail clamping. Then the changes in pyramidal cells'firing, interneurons'firing, and the power of local field potentials were studied.The main results of this dissertation are summarized as follows:(1) The detecting ratio of multi-channel threshold detecting method promoted 30% camparing with single-channel threshold method, and more neuronal firings were detected in one trial. The accurate detecting ratio of PCA-NEO detecting method is up to over 92%, and detecting ratio of overlapped spikes was 20% higher than other three methods, meanwhile, the false detecting ratio was well controlled.(2) The multi-channel sorting method first connected the multi-channel spike waveforms end to end to make composite waveforms. Then the features of spikes were extracted from composite waveforms by PCA, and spikes were sorted by max-min distance feature clustering method. Through this sorting method, spikes were nearly completely classified correctly, and the number of extracted overlapped spikes were increased.(3) Pyramidal and interneuronal spikes were extracted out of multi-channel signals recorded in hippocampal CA1 area before and under tail clamping. Then analized by raster plot and firing histogram, it was found that under tail clamping stimulation, some pyramidal cells fired much less than normal while some interneurons fired much more. And the changes in neuronal firings induced the decrease of the power of low frequency area(8 rhythm) and the increase of the power of high frequency area(βrhythm, y rhythm and ripples) in local field potentials.In a word, the spike detecting and sorting methods designed in this dissertation have great advantages in spike detecting ratio, sorting accuracy and calculating efficiency, and provide better signal analyzing methods in neural electrophysiology research. Using these new methods, pyramidal and interneuronal firings were extracted from hippocampal CA1 area, and a new variation law of neuronal firing modes under tail clamping were found. As hippocampus is an important learning and memory area in brain, the results found in this dissertation may relate to the neuronal encoding process for sensing stimulation.