Models for the cross-correlation between retinal ganglion cells

Area of Science:

• Neuroscience
• Computational Biology
• Retinal Physiology

Background:

• Vertebrate retinal ganglion cells exhibit correlated activity on millisecond timescales.
• Previous models involving simple linear addition of common variability failed to explain observed cross-correlations.

Purpose of the Study:

• To investigate and model the mechanisms underlying correlated activity in neighboring retinal ganglion cells.
• To test linear and nonlinear models against experimental observations of neural cross-correlation.

Main Methods:

• Utilized Monte Carlo simulations to evaluate various linear and nonlinear neural models.
• Focused on models incorporating common variability affecting neuronal leakage properties.

Main Results:

• Simulations confirmed that simple linear addition of common variability is insufficient.
• Models where common variability modulates the leakages of leaky integrate-and-fire neurons accurately reproduced the observed data.
• These successful models predicted cross-correlogram lag without requiring explicit temporal delay lines.

Conclusions:

• Common variability influencing neuronal leakage is a key mechanism for correlated ganglion cell activity.
• This finding offers a parsimonious explanation for millisecond-timescale cross-correlations in the retina.
• The proposed model provides a framework for understanding neural synchrony without invoking complex temporal processing.