A Preliminary Analysis on Impact of Additive Flicker Noise on Detection Sensitivity of Neural Spikes.

Abstract

The ultimate objective of this study is to enhance the number of neurons that can be monitored in an extracellular recording captured by a single electrode. This paper focuses on exploring if additive 1/f noise can improve detection sensitivity of spikes captured in an extracellular medium. Detection performance of a simple bandpass filtering stage followed by a threshold-based spike detector is investigated when 1/f noise is added to the filtered signals. The study is performed on both extracellular recorded datasets and synthetic datasets. The preliminary data demonstrated that, threshold-based spike detection of recordings with small amplitude spikes in high background noise can be optimized by a non-zero additive flicker noise. The addition of non-zero additive flicker noise increased the spike detection sensitivity in neurons in regions successively further from the electrode from 0 to as high as 0.8. The results suggest that it may be possible for a single electrode to use multiple spike detection threshold stages to increase sensitivity in a future neural monitoring system.