Experimental setup for tuning curves in biased competition with feature-based attention.
More...
Experimental setup for tuning curves in biased competition with feature-based attention.
- Author
- Frederik Beuth
Data: Treue and Martinez, 1999, Fig. 3B
Significant results (p. 577, left column):
- Significance test were not reported, probably due to lack of space in a Nature paper.
- Attention has a strong effect, it increases in average the response to 156%.
- No sharpening was observed.
Setup:
- Stimuli size is unknown, except that two stimuli fit in the RF size of area MT. => We model them as small stimuli compared to the RF size.
Calibration of the fit:
- Suppression results primarily from the distractor, hence amount of suppression in all 3 conditions results from feature-based suppression.
- Thus the suppression cannot be calibrated independently in each condition, rather the amount of attention creates the difference of the curves.
- If the variable stimulus is an anti-preferred one, additional surround suppression occurs. However, this effect is weak compared to the feature-based suppression influence, so it is not visible in the data.
- Strangely, the data shows also an amplification in condition attend pref. at -120°. This orientation is different to the preferred orientation, so it still receives strong feature-based suppression in the model and in other data sets, thus no amplification should be observed. Thus with the standard feature-based suppression weights, this effect should not be possible.