Automatic parameter estimation of multicompartmental neuron models via minimization of trace error with control...

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Biophysics

Background:

• Accurately modeling biological neurons is crucial for understanding neural function.
• Traditional methods struggle with precise action potential timing when fitting models to voltage traces.
• Injected noise can perturb neuron activity for more robust data acquisition.

Purpose of the Study:

• To introduce a new technique for fitting conductance-based neuron models to intracellular voltage recordings.
• To improve the phenomenological match between model and biological neuron voltage traces, particularly spike timing.
• To address the challenge of inevitable action potential timing mismatches in existing fitting methods.

Main Methods:

• Utilized intracellular voltage traces from isolated biological neurons recorded in current-clamp with injected pink (1/f) noise.
• Developed a novel algorithm to fit multicompartmental model neurons to recorded voltage traces.
• Incorporated a weak control adjustment, inspired by Luenberger observers, to promote alignment during fitting.
• Tested the method on both synthetic data and biological recordings from crab stomatogastric ganglion neurons.

Main Results:

• The new algorithm successfully fitted conductance-based neuron models to voltage traces.
• The control adjustment significantly improved spike-timing accuracy between model and biological neurons.
• The technique demonstrated flexibility, working even when synthetic data used different conductance models than the fitting model.
• Accurate spike-timing was achieved for both synthetic and biological datasets.

Conclusions:

• The developed technique offers a robust solution for fitting neuron models to experimental voltage data.
• Improved spike-timing accuracy enhances the reliability of computational neuron models for neuroscience research.
• This approach provides a powerful tool for advancing our understanding of neural dynamics and circuit function.