Dissecting estimation of conductances in subthreshold regimes

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biophysics

Background:

• Linear regression methods are commonly used to estimate synaptic conductances from neuronal recordings.
• Existing methods often remove spiking activity but overlook the impact of subthreshold ionic currents.
• Subthreshold activity, crucial for neuronal function, can introduce significant biases in these estimations.

Purpose of the Study:

• To investigate the influence of subthreshold ionic currents on the accuracy of synaptic conductance estimation.
• To identify and quantify errors introduced by nonlinear subthreshold dynamics in linear estimation methods.
• To develop and validate an improved method for estimating synaptic conductances that accounts for subthreshold activity.

Main Methods:

• Utilized conductance-based models to simulate neuronal activity.
• Introduced various mechanisms to induce representative ionic subthreshold activity.
• Applied linear regression for initial conductance estimation and compared it with a novel quadratization-based method.

Main Results:

• Subthreshold ionic currents were shown to cause significant errors in linearly estimated synaptic conductances.
• The proposed quadratization method reduced relative errors in conductance estimation by over an order of magnitude.
• The findings underscore the limitations of current linear methods in the presence of subthreshold activity.

Conclusions:

• Linear estimation of synaptic conductances is unreliable in the presence of subthreshold ionic currents.
• The developed quadratization method offers a more accurate approach for estimating conductances, even with noisy experimental data.
• These results necessitate a re-evaluation of conclusions drawn from previous intracellular recordings and highlight the importance of considering nonlinear dynamics.