Complex parameter landscape for a complex neuron model
2006
Understanding Neuron Models Through Parameter Optimization
Sample size: 20
publication
Evidence: high
Author Information
Author(s): Pablo Achard, Erik De Schutter
Primary Institution: University of Antwerp, Belgium
Hypothesis
How many models of a neuron can have similar input/output behavior?
Conclusion
The study found that different sets of ionic channel densities can produce similar electrical activities in neuron models.
Supporting Evidence
- The models produced similar outputs to current injections, including complex firing patterns.
- The parameter landscape of good models can be described as loosely connected hyperplanes.
- The study demonstrates that model parameters are not uniquely identifiable.
Takeaway
Scientists created 20 different models of a neuron that all acted similarly, showing that neurons can be flexible in how they work.
Methodology
An evolution strategy algorithm was used to optimize 24 channel density parameters in neuron models.
Limitations
The study assumes that the morphology of the cell is known and that channel kinetics are not modified.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website