Experimental setup for biased competition and pooling with spatial attention.
- Author
- Frederik Beuth
Data: Lee and Maunsell (2010, J Neurosc), Fig. 5
Significant results (p. 932 left)
Significance tests were not conducted, but the authors reported the following results:
- Response to a pair of null and preferred stimulus is between the response of each individually.
- Weak attentional modulation for a single stimulus (9% increase on average), strong for a stimulus pair.
- Shifting attention from null to preferred stimulus of the pair increase the response on average by 59% (for preferred to null).
- From attend away to attend preferred stimulus of the pair increase the responses by 28% (away to preferred)
- Attention modulation was weak for 50-125ms: 18% for preferred to null, 10% for away to preferred
- Attention modulation was strong for 125-250ms: 111% for preferred to null, 47% for away to preferred
Setup
- Stimuli are oriented moving gratings at 1) 90 and 270°, 2) 90 and 180° and 3) 90 and 90°, so execute 3 subexperiments
- Record the neuron preferring stimulus 1 (assuming 90° orientation) in V4
- Record neuron in all of the 4 conditions in each experiment
- Contrary to Fig. 2, we use the data of a single stimulus in the attended own condition here(Fig. 2 uses attended away). This will result in slightly different parameters.
- Stimuli have the size of about 0.7°-1.2° (Gabors with SD of 0.35°-0.6°) at unknown eccentricity. However, this size is very small compared to MT rf sizes at all eccentricities (4°-19° according to Lee and Maunsell), so we assume that a stimulus falls only in a single receptive field.
Calibration of the fit:
- Pretty standard biased competition, except a baseline has to be used as the data contains a peak after stimulus onset which is not modulated by attention.