摘要： Electrical brain stimulation provides therapeutic benefits for patients with drug-resistant neurological disorders. But it has restricted access to cell-type selectivity. Optogenetics, in contrast, enables precise targeting of a specific cell type which can address the issue with electrical brain stimulation. Optogenetics is a light-based stimulation method, in which the target cells are transfected with light-sensitive ion channels called opsin. This technique can be used for both excitation and inhibition of cells based on light wavelength and opsin properties. To online modulation of neurons need a closed loop controlling. Closed-loop optogenetics system compares the neural signals with a predefined value at every moment and then decides what to do to achieve the desired value by optical stimulation [1]. One of the important brain signal recordings is single-unit recording (SUR) [2] whiles the most of signal processing method is offline. In this paper, an algorithm was proposed to online spike detection and sorting them based on wavelet transform which was programmed by Labview software. The algorithm consists of three main steps such as detecting spikes, extracting the features of every detected spike and assigning similar spikes to one group. The background noise in multi-unit recording has Gaussian distribution and it contains spikes which the mean value of the signal was considered in signal processing. Time of calculating the main value is about 6.04 μs. If the mean value of a dataset changes noticeably from the estimated value, a candidate spike has occurred. Detected spikes are a discrete sample of the real spikes which the waveforms should be interpolated to improve the detection spike [3]. With detecting a spike, its features must be extracted as the input of spike sorting algorithm. Spikes were sorted in different clusters based on their shapes. Wavelet transform compares the input signal with a predefined function known as a wavelet in different scales. In this algorithm was used 20 scales and Mexican hat wavelet to transform the SUR signal [4]. Total time for interpolation and applying wavelet transform takes about 5.3 ms. The final step of the algorithm is sorting the detected spikes using the extracted features based on applying the wavelet transform and compared all wavelet coefficients of a detected spike with the corresponding coefficients of all the clusters. If the wavelet coefficient in the clusters was within 30% of the corresponding coefficient of the detected spike in all scales, spike put in one group. This part of the algorithm takes about 3.1 ms. Schematic diagram of spike sorting and a photo of the front panel of designed software was shown in Fig.1. In conclusion, designed software has a high processing speed and can be used for online and fully automatic spike sorting for studying and controlling the neural network system with closed loop optogenetics technique.