Consequences of converting graded to action potentials upon neural information coding and energy efficiency

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biophysics

Background:

• Neural information processing involves graded and action potentials.
• Conversion between these potentials leads to information loss and reduced energy efficiency.

Purpose of the Study:

• Investigate the biophysical causes of information and energy loss during neural signal conversion.
• Compare spiking neuron models with generator and graded potential models.

Main Methods:

• Utilized stochastic voltage-gated Na(+) and K(+) channel models for spiking neurons.
• Compared these with models lacking voltage-gated Na(+) channels (generator and graded potentials).

Main Results:

• Identified three causes of information loss in generator potentials: increased noise, non-linearities from voltage-gated Na(+) channels, and action potential footprints.
• Generator potentials showed ~50% reduced information rates compared to spike trains.
• Both generator and graded potentials consumed significantly less energy than spike trains.

Conclusions:

• Action potential generation incurs a two-fold cost: information loss and increased energy consumption.
• Graded potentials are more energy-efficient than generator potentials due to higher information rates.
• Converting analog to digital neural signals involves substantial trade-offs.