Experimental setup for contrast dependent competition.
- Author
- Frederik Beuth
Data: Busse, Wade, Carandini (2009), Fig. 4A
Significant results (p. 932 left)
- Significance tests were not conducted, but the authors reported the following results:
- Single stimulus: contrast changes modulate the amplitude of a tuning curve, but not its width. This finding is valid for single neurons as well as for the population response.
- Overlaid stimuli (plaid):
- Similar contrast (6% and 12%, or 12% and 12%) result in two peaks in the population response
- Large different contrast result in one peak in the population response
- Response to pair is lower as responses to single stimuli (cross-orientation suppression)
Setup
- Vary the stimulus contrast over the experiments with values of 0%, 12%, or 50% contrast. Use only these 3 values to prevent cluttering of the plot
- Record the full neuronal population in V1 (used 13 neurons in the model like in the data)
- Stimuli are oriented edges of 0° and 90°, hence use neurons 4 and 10
- Experiment uses overlapping stimuli, so place both stimuli at the same place
- Experiment uses no attention, so record the neurons only in the attend away conditions
- Stimuli sizes and RF sizes were not reported. We could not investigate the amount of surround suppression as it would be visible only in an attention condition which is not part of the data. Thus we assume no suppression as it is the simplest case, hence we assume that a stimulus falls only in a single receptive field.
Calibration of the fit:
- 1) Calibrate contrast response function parameters (sigma, eV, eP) to fit the peaks in conditions with a single stimulus
- 2) Set width and baseline of input tuning function
- 3) Calibrate amount of inhibition to fit data of conditions with a stimulus pair