Modelling spatiotemporal olfactory data in two steps: from binary to Hodgkin-Huxley neurones

Area of Science:

• Computational Neuroscience
• Neural Network Modeling
• Olfactory System Research

Background:

• Biological neural networks exhibit complex spatiotemporal patterns.
• Synchronously updated McCulloch-Pitts neuron models show similarities to biological neuronal activity.
• Formal models offer analytical tractability for understanding neural dynamics.

Purpose of the Study:

• To construct a formal network of Hodgkin-Huxley (HH) neurons that replicates experimentally observed neuronal codes.
• To investigate the feasibility of a two-step modeling strategy for complex neural systems.
• To analyze the robustness of neuronal codes to synaptic noise.

Main Methods:

• A two-step strategy was employed: first, a binary unit network was designed to match experimental codes.
• Second, this binary model guided the design of a more realistic network of formal HH neurons.
• The resulting HH neuron model was used to reproduce Wehr-Laurent olfactory codes and assess noise sensitivity.

Main Results:

• The proposed two-step strategy successfully generated a model of neuronal activity.
• The model accurately reproduced the Wehr-Laurent olfactory codes from locusts.
• The study investigated the resilience of these olfactory codes against synaptic noise.

Conclusions:

• A two-step approach, starting with binary units and progressing to Hodgkin-Huxley models, is effective for reproducing biological neuronal codes.
• This methodology allows for the creation of detailed neural network models that mimic experimental observations.
• The findings provide insights into the robustness of olfactory coding mechanisms in the presence of neural noise.