www.imodel.org

Dog.1.1

2007 Jul 18
Summary   This model uses a difference-of-Gaussians linear filter like that of Dog.0.1, but it has a mechanism for non-classical surround suppression. It also has a stronger driven firing rate than Dog.0.1 because of changes to the scale and bias parameters that control the signal that is passed from the DOG filter to the integrate-and-fire mechanism. Other differences compared to Dog.0.1 are: a finer spatial resolution (sscale is 0.04) and a larger spatial grid (xn and yn are 64).

Dog.1.1.mod

Data

Results

The results here were obtained using this response file, which records spikes from an ON-center unit near the center of the stimulus grid.

Sine Tuning

• Spatial frequency
• Temporal frequency
• Size
• Contrast
• Static phase

Dynamic Random Stimuli

• Random counter-phase vs. contrast
• Random counter-phase vs. SF

Discussion

The size tuning curves taken at TF 6 Hz differ substantially from those at 16 Hz. The ones at 6 Hz are typical of LGN cells, but the substantial increase in suppression at 16 Hz is probably not realistic.