Abstract:Nonlinear mechanisms are at the heart of neuronal information processing, for example to fire an action potential, the membrane voltage must exceed a threshold nonlinearity. Even though, linear encoding schemes are commonly used and often successfully describe large parts of sensory encoding nonlinear mechanisms such as thresholds and saturations are well known to be crucial to encode behaviorally relevant features in the stimulus space not captured by linear methods. Here we analyze the role of bursts in p-type electroreceptor afferents (P-units) in the weakly electric fish Apteronotus leptorhynchus. It is long known that subpopulations of these cells fire bursts of action potentials while others do not. Previous research suggests, that the non-bursting cells are better at encoding the stimulus time-course while bursting neurons are better suited to encode special features in the stimulus. We here show, based on the analysis of experimental data and modelling, that burst affect the linear as well as the nonlinear encoding. Theoretical work predicts that in simple leaky-integrate-and-fire model neurons, two periodic stimuli interact nonlinearly when the sum of the two frequencies matches the neuron's baseline firing rate as quantified by the second-order susceptibility. Indeed, such nonlinear responses have been found in non-bursting P-units when stimulated by two beats simultaneously but only in those cells, that exhibit very low levels of intrinsic noise. In this study, we found that bursts strongly enhance these nonlinear responses which may play a critical role for the detection of weak intruder signals in the presence of a strong female signal, i.e. an electrosensory cocktail party.

What problem does this paper attempt to address?

This paper mainly explores the nonlinear mechanisms in neural encoding, particularly the impact of bursting in primary afferents (P-units) on nonlinear encoding. The research focuses on P-units in weakly electric fish Apteronotus leptorhynchus, some of which exhibit burst firing while others do not. The study reveals that non-bursting cells perform better in encoding temporal processes of stimuli, while bursting neurons are more suitable for encoding specific features of stimuli. The paper suggests that in a simple leaky integrate-and-fire model neuron, it is theoretically predicted that two periodic stimuli with a sum of frequencies equal to the neuron's baseline firing rate would interact nonlinearly. The experimental results confirm the existence of this nonlinear response in non-bursting P-units, but only in cells with very low intrinsic noise. Furthermore, the study finds that burst firing significantly enhances this nonlinear response, which may play a critical role in detecting weak invading signals against a strong signal background (e.g., the "cocktail party" effect in electroreception). The paper also compares the linear and nonlinear stimulus encoding of bursting and non-bursting P-units, and discovers that bursting increases nonlinear susceptibility and expands the range of stimuli that lead to nonlinear interactions. Through simulation studies, the authors propose that bursting increases the second-order susceptibility not through the timing of the first action potential, but by increasing the number of action potentials within a burst. Additionally, similar nonlinear encoding patterns are observed even in non-bursting P-units of Eigenmannia virescens, indicating that nonlinear encoding may not solely depend on burst behavior. In conclusion, this paper reveals the significant role of bursting in neural encoding, particularly in processing complex stimuli and enhancing weak signal detection capabilities through nonlinear effects.

Bursts boost nonlinear encoding in electroreceptor afferents