Figure 2. Feature-based attention: behavioral tasks and effects of attention on accuracy and certainty.

A, Timing of events for an example feature-based attention trial. The test stimulus consisted of two interlacing RDKs differing in color and was presented randomly left or right of the fixation point (1–1.5 s), modulation of the independent RDKs otherwise matched the spatial attention tasks. Color changes of the fixation point to red or green before (0.5–1 s) and after the presentation (1.5–2 s) of this test stimulus instructed subjects which dot elements to direct attention to and which to actually indicate the direction of global motion for. The other cueing and response modalities were identical to the spatial attention tasks. B–I, Percentage of correct responses or certainty index, respectively, vs. motion coherence for all subjects. Conventions are identical to the spatial attention tasks in Fig. 1. * tags p<0.05; *** p<0.001 derived from model comparison statistics using Monte-Carlo simulations of the two respective fits, missing asterisk in I: no significant difference between fits. Feature-based attention with numerical certainty ratings (FN): B–E, feature-based attention with certainty wager (FW): F–I. B, F compare accuracy for valid and invalid cues. C, G compare certainty for valid and invalid cues. D, H compare accuracy and certainty for valid trials. E, I compare accuracy and certainty for invalid trials.