Experimental setup for biased competition - selectivity in relation to sensory interaction.
More...
Experimental setup for biased competition - selectivity in relation to sensory interaction.
- Author
- Frederik Beuth
Data: Reynolds, Chelazzi, Desimone (1999), Fig. 11a,b,d
Significant results (p. 932 left)
- 68% of the neurons are modulated by attention. The following results concern this sub-population:
- Attend away:
- Slope of the fitted line is 0.49, which is not significantly different from 0.5
- The line is significantly shifted upward by 0.06, indicating an increased response due to adding the second stimuli
- Attend probe (mostly anti-preferred) stimulus:
- Slope is significantly increased from 0.49 to 0.83, which is significantly different from 0.5
- Attend reference (mostly preferred) stimulus:
- Slope is significantly decreased from 0.49 to 0.21, which is significantly different from 0.5
Setup
- Record an arbitrary neuron in V4
- Use a pool of 16 stimuli with a mix of different orientations and colors
- In each experiment, arbitrary select a reference and a probe stimulus from them. They are arbitrary selected to investigate all possible stimulus conditions
- Record neuron in all of the 9 conditions in each experiment
- Calculate for each experiment the selectivity and sensory interaction for all 3 attention conditions
- Stimuli have a width of 0.25° and a heigth of 1°, eccentricity is unknown. However, this size is very small compared to V4 rf sizes at all eccentricities, so we assume that a stimulus falls only in the receptive field of a single V4 location.
Calibration of the fit:
- Pretty standard biased competition
- The unattended condition will mostly result in the correct slope of ~0.5
- Calibrate first amplification to fit attend reference condition
- Afterwards, calibrate suppression to fit attend probe case