Abstract
On-implant spike sorting methods employ static feature extraction/selection techniques to minimize the hardware cost. Here we propose a novel framework for real-time spike sorting based on dynamic selection of features. We select salient features that maximize the geometric-mean of between-class distances as well as the associated homogeneity index effectively to best discriminate spikes for classification. Wave-shape classification is performed based on a multi-label window discrimination approach. An external module calculates the salient features and discrimination windows through optimizing a replica of the on-implant operation, and then configures the on-implant spike sorter for real-time online operation. Hardware implementation of the on-implant online spike sorter for 512 channels of concurrent extra-cellular neural signals is reported, with an average classification accuracy of ~88%. Compared with other similar methods, our method shows reduction in classification error by a factor of ~2, and also reduction in the required memory space by a factor of ~5.
Highlights
On-implant spike sorting methods employ static feature extraction/selection techniques to minimize the hardware cost
We propose an automated method for online spike sorting dedicated to high-density, highspeed brain implants
We introduced a novel framework for on-implant spike sorting
Summary
On-implant spike sorting methods employ static feature extraction/selection techniques to minimize the hardware cost. 1234567890():,; In the realization of brain implants and neural prostheses, one of the main challenges is to increase the number of recording channels. Spike sorting can be performed through the following general steps: (i) filtering the raw neural signal (from 0.3 to 6 kHz) to preserve only the useful frequency content of neural spikes, (ii) detection of spike events upon the firing of neurons, (iii) extraction of spike wave-shapes from the filtered neural signal (for details of our spike detection and extraction method, refer to our previous work2), (iv) temporal alignment of the spike wave-shapes, to avoid additional hardware cost, in this work spike wave-shapes are aligned to the detection (first threshold-crossing) points, (v) mapping of the extracted spike wave-shapes into a feature space, known as feature extraction, this step is to enhance the discrimination between spikes and noise, and between different spike classes ( referred to as between-class variability), (vi) selection of a minimal subset of features, known as feature selection, in order to reduce the dimensions of the data being processed, and (vii) classification or clustering of the wave-shapes into different spike classes as isolated units. To make the spike sorting procedure complete, onimplant classification of spike wave-shapes has been realized using distance-based classification and oblique decision tree classification methods
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
