Context-dependent coding in single neurons.

Area of Science:

• Computational neuroscience
• Neural coding and information theory

Background:

• The linear-nonlinear cascade (LN) model effectively represents neural encoding but struggles with history-dependent firing patterns.
• Direct current (DC) stimulus and post-spike refractory periods, like those from afterhyperpolarization (AHP) currents in motoneurons, introduce regularity and reduce available spike-timing entropy for encoding fluctuations.

Purpose of the Study:

• To address limitations of the LN model in capturing history-dependent neural firing.
• To explore alternative models that accurately represent single-neuron behavior under varying DC and stimulus variance.
• To quantify the tradeoff between encoding instantaneous fluctuations and DC stimulus information.

Main Methods:

• Investigated alternative neural encoding models beyond the standard LN model.
• Utilized a motoneuron simulation to compare model performance with and without the afterhyperpolarization (AHP) current.
• Quantified the tradeoff between encoding stimulus fluctuations and DC input.

Main Results:

• Models incorporating factors like DC stimulus and AHP currents show improved accuracy in reproducing neuron firing patterns.
• The presence of AHP current significantly regularizes firing behavior, impacting information encoding capacity.
• A quantifiable tradeoff exists between the neuron's ability to encode rapid fluctuations versus sustained DC input.

Conclusions:

• Alternative models are necessary to accurately capture the complex firing dynamics of neurons, especially those influenced by intrinsic currents like AHP.
• Understanding the interplay between DC input, intrinsic currents, and stimulus fluctuations is key to comprehending neural coding efficiency.
• This work provides a framework for analyzing the dual role of neurons in encoding both dynamic stimuli and sustained inputs.