FIGURE 2.

Multiple suppression windows. (A) The waveforms for both stimuli (“dad” in the two upper panels, “sad” in the two lower panels) are shown with the “Early” and “Late” laser suppression windows overlaid in cyan. Approximate locations of the phonemes within each stimulus are indicated above each waveform. Early suppression coincides with the initial consonant (/d/ or /s/), and Late suppression coincides with the vowel (/a/). (B) Performance in overall percent correct for each mouse for 10,000 total trials is represented by connected dots representing each of the three types of trials. Bars represent mean performance for each trial type across all nine mice. Mice showed similar effects from either Early or Late suppression. (C) To partition suppression effects into accuracy and bias, performance for each mouse (same data as in B) was separated into “dad hit rate” (percent correct on “dad” trials) and “dad false alarm rate” (100 - percent correct on “sad” trials). Each mouse is represented with a large open circle corresponding to control trials, connected to small circles corresponding to Early (black) and Late (red) suppression conditions. The dashed line represents chance performance (50% correct). The lower left-hand corner (0,0) represents 100% “sad” responses (i.e., total bias toward “sad”). The upper right-hand corner (100, 100) represents 100% “dad” responses (i.e., total bias toward “dad”). Optogenetic suppression reduced accuracy for all mice, shifting performance toward the dashed line. The amount of response bias on suppression trials ranged from almost none (a shift perpendicular to the dashed line) to a strong bias toward “dad” (a shift toward the upper right-hand corner). If the effect of the laser was exactly the same for both suppression windows, the two small red and black circles would coincide. The effects of Early and Late suppression windows were very similar.