Fig. 2: Goal-associated inhibitory reduction is required for new goal learning.
a, Optogenetic PV stimulation protocol with stimulation locations and intensities for goal and sham stimulation (stim.). The mouse illustration was adapted from scidraw.io. b, Disrupting normal goal-associated inhibitory firing reduction (blue) impaired the learning of new goal locations. Performance based on speed differentiation over 3 days of learning for all trial types (goal stimulation, blue, n = 5 mice; sham stimulation, orange, n = 5 mice; day 1 versus day 2, P = 0.0045; day 1 versus day 3, P = 0.0001; LMM followed by Tukey correction). c, The RZ identification speed as in b, but for trials with low (left), high (middle) or no (right) stimulation. There was a significant performance increase from day 1 with sham, but not goal, stimulation at all intensities (low stimulation, day 1 versus day 2, P = 0.006; day 1 versus day 3, P < 0.0001; high stimulation, day 1 versus day 3, P = 0.03; no stimulation, day 1 versus day 2, P = 0.01, day 1 versus day 3, P = 0.002; LMM with Tukey correction). d, The normalized change in pyramidal cell firing from the baseline over time with goal (blue) or sham (orange) stimulation across stimulation intensities (top). The dashed line indicates the stimulation onset. Bottom row, there were no significant differences in the mean firing rates in the 2-s period after the start of goal (n = 307 cells) or sham (n = 270 cells) stimulation across intensities. e, The change in firing rate as in d, but for PV interneurons. n = 19 (goal stimulation) and 16 (sham stimulation) cells. For b–e, data are mean ± s.e.m. NS, not significant. **P < 0.01. The brain (F. Claudi) in the mouse head (E. Tyler and L. Kravitz) diagrams in a were adapted under a CC BY 4.0 licence.
