Detection and Characterization of Neural Spikes

In this paper the characteristics of neural spikes in deep brain recording were investigated. The adaptive spike thresholding was used to detect the neural spikes, and the autoregressive model was proposed to differentiate neural spikes and background potentials. Our preliminary results indicated there appears a higher firing rate, more concentrated inter-spike interval histogram and more correlated spike train in the neural signal recorded in substantia Nigra compared to the signals in reticulate subthalamic nucleus. (4 pages)