Identification of stimulus-response relations for cultured neuronal networks using features of multiple temporal resolution levels

Abstract

In recent years cultured neuronal networks have been used with the aim of unraveling how biological information transmits between neurons. Investigating the evoked activities of cultured neuronal networks helps acquire a better understanding of neural decoding. However, it is still challenging to quantitatively describe and predict evoked patterns for them. This study focuses on evoked patterns of cultured neuronal networks and aims to identify stimulus-response relations with extracted feature sets including spike-based and rate-based features. The majority of neural information is encoded in evoked activities of the post-stimulus intervals. By partitioning post-stimulus intervals, features with multiple temporal resolution levels were constructed. This study investigates the impact of temporal resolution level on accuracy in recognizing stimulus-response relations.