What causes a neuron to spike?

Area of Science:

• Computational Neuroscience
• Neural Coding
• Systems Neuroscience

Background:

• Neuronal computation involves stimulus dimensional reduction and response nonlinearity.
• Reverse correlation (spike-triggered average/covariance) is standard for characterizing neuronal sensitivity and nonlinearity.
• Existing methods struggle with simple models like the leaky integrate-and-fire neuron.

Purpose of the Study:

• To apply and adapt reverse correlation techniques to the leaky integrate-and-fire model.
• To develop novel methods for accurately characterizing neural computation in this model.
• To investigate implications for understanding neural adaptation.

Main Methods:

• Application of reverse correlation to the leaky integrate-and-fire model.
• Development of novel techniques analyzing isolated spikes and preceding silent periods.
• Comparison of new methods against standard reverse correlation approaches.

Main Results:

• Standard reverse correlation techniques fail to accurately recover known computations in the leaky integrate-and-fire model.
• Novel reverse correlation methods, accounting for isolated spikes and silences, successfully characterize the model's computation.
• The developed methods offer improved analysis of neural adaptation.

Conclusions:

• Novel reverse correlation techniques are necessary for accurate analysis of even simple model neurons.
• These methods provide a more robust framework for understanding neural computation and adaptation.
• Findings are relevant for analyzing spike trains from both model and real neurons.