A System Implementation For Rat Neural Signal Wireless Acquisition And Online Spike Sorting

The detection and classification of the spike signal is an important part of neural information encoding or decoding in brain science research.The traditional equipment for spike signal collecting and analyzing is bulky and expensive,and is usually a wired transmission method.The wire connection increases the external interference to small freely moving animals,and influences the small animal neural signal under the specified movement experiments.Therefore,basing on the characteristics of embedded system,this paper designs and implements an experimental system for neural signal wireless acquisition and online spike classification,and uses simulation signals and real spike signals of experimental animals to separately test and verify the online detection and classification functions.The system mentioned in this paper mainly includes a front-end spike detection device and a back-end PC-based online spike classifier.With the microcontroller as the core,the front-end acquisition detection device configures the front-end acquisition chip through SPI interface and then completes the update of adaptive detection threshold,spike detection and other pretreatments in the microcontroller;then,the wave data of spike will be sent through the wireless RF module.After receiving the data,the Lab VIEW upper computer combined with the MATLAB script will run spike online classification algorithm to classify the spike in real time.The verification and test results show that the proposed rat's neural signal wireless acquisition and online spike classification can achieve all the predetermined functions.In the simulation experiment,the classification accuracy of the signal of the analog signal generator can reach 100%at most,and for the free moving rat spike signal,the classification result of the proposed system is well consistent with the traditional PC offline classification system,and the similarity of spike classification between the system and traditional PC offline classification system can at least reach 75.0%while highest reach 93.9%,indicating that the performance between this wireless system and commercial offline system is very close.However,the size,portability(35mm x 30mm× 18mm,18g below)and power consumption(3.7V,180mAh lithium battery power supply,stable work of up to 40 minutes)of the proposed system are superior to the traditional PC system,which can better meet the need of free-moving rat's spike detection experiment.In summary,the design and implementation of this system lays a good foundation for the subsequent research on embedded,miniaturized and wireless neural signal analysis systems.