Fig. 8: Response time (RT) comparison between XOR and conjunction tasks.
From: Visual perceptual learning of feature conjunctions leverages non-linear mixed selectivity
a In the XOR transfer task, subjects had to respond with an eye movement to the left for red-anticlockwise and purple-clockwise stimuli, and to the right for red-clockwise and purple-anticlockwise stimuli. Colors and orientations in the XOR task were identical to the ones used in the conjunction task (and are only displayed for illustration purposes here). b Response times (RT) in the XOR task were generally slower than in the conjunction task (RT bias toward XOR), but this effect was larger in Experiment 2 than in Experiment 1 (mixed repeated measure ANOVA; Experiment: F(1,10) = 0.0072, p = 0.934, η2 = 0.0007, Task: F(1,850) = 763.4, p < 10−15, η2 = 0.4732, Experiment×Task: F(1,850) = 26.72, p < 10−6, η2 = 0.0305). Dashed blue line signifies RT equality (XOR = conjunction). Histograms of RT biases (distance from RT equality XOR=conjunction) are shown in the top right corner (black: Experiment 1, red: Experiment 2). c Biases from the equality line where RT in XOR = RT in conjunction (bias=0, signified by dashed blue line). The bias towards XOR is significantly larger in Experiment 2 than in Experiment 1 (p < 10−5). Error bars represent standard deviation. *** denotes p < 0.001.
