Variability of responses to sinusoidal modulation

Area of Science:

• Neuroscience
• Computational Neuroscience
• Vision Science

Background:

• Visual neurons are often studied using sinusoidally modulated stimuli.
• Neural response variability is a key factor in understanding neural coding.
• The fundamental Fourier component is commonly used to quantify neural responses.

Purpose of the Study:

• To investigate the variability of the fundamental sinusoidal components of visual neuron firing.
• To theoretically analyze the complex and scalar variance of neural responses.
• To compare theoretical predictions with experimental data from cat retinal ganglion cells.

Main Methods:

• Theoretical analysis of Fourier component variance.
• Definition and analysis of complex variance and scalar variance.
• Stimulation of cat retinal ganglion cells with sinusoidally modulated lights of varying contrasts.

Main Results:

• Theoretical analysis predicts complex variance is nearly independent of response amplitude.
• Scalar variance is predicted to decrease as response magnitude approaches zero.
• Experimental data from cat ganglion cells largely agree with theoretical predictions.
• Some individual cell data suggest potential nonlinearities in the rate-variance relationship.

Conclusions:

• The complex variance of neural responses to sinusoidal stimuli is largely independent of amplitude.
• This finding holds even when the variance of the firing rate itself increases with the firing rate.
• The results provide insights into the coding properties of visual neurons and the nature of neural variability.