Viral vector expression and fiberoptic placement

(A) Anatomical distribution of eYFP fluorescence expression in distinct AP coordinates of the retrosplenial cortex after injection in the hippocampus (HIPP). (B) Injection spots in the HIPP (C) Fiberoptic placement two fiberoptic stubs at AP reversed 30 ° angles reaching the two hemispheres at two distinct AP levels of RSC (arrows point to fiberoptic tracks, please refer to Figure S1 for low-mag images of all animals and examples of high-mag images).

Delayed Non-Match to Trajectory protocol and baseline performance

(A) DNMP protocol. Rats explore the unblocked arm in a sample run, then wait in the starting area for a 15 sec delay. Once back on the track, with both arms unblocked, rats receive a reward if they choose the arm opposite the one explored in the sample run. (B) Training schedule, after reaching 75% performance, 3 baseline (no light delivered) sessions are followed by 10 sessions with illuminated trials interleaved with non-illuminated ones. (C) No differences in performance were observed in the baseline sessions (GLMM, p=0.212, Bonferroni post-hoc test)

Optogenetic silencing of HIPP terminals in RSC causes a significant and prolonged decrease in performance

(A) Illumination of RSC during DNMP leads to a significantly lower performance in eArch+, compared to empty-vector CTRL animals (CTRL: 0.81±0.03, ARCH: 0.73±0.08, GLMM, p=0.006, Bonferroni post-hoc test). (B) In eArchT+ animals, we found no differences between performance in TI and NI trials taken as groups, both significantly lower than both baseline (p<0.001), and the corresponding trial groups in CTRL animals (TI, p=0.02, NI, p=0.008). In CTRL animals no differences were found among the three groups (p=0.252 and 0.184) (C) Illumination on any given trial resulted in lower performance in the subsequent trial in eArch+, but not CTRL, animals (p=0.015), regardless of whether the illuminated trial is a correct or an incorrect trial. All summary data depicted is mean+/−SD.

Optogenetic silencing of HIPP terminals in RSC causes persistent error increases.

(A) eArch+ animals exhibit persistent errors (p<0.001, Bonferroni post-hoc test). (B) Persistent errors are absent in baseline (no light delivered) sessions (p=0.172, Bonferroni post-hoc test). All summary data depicted is mean+/−SD.

Behavioral impairments resulting from optogenetic silencing of HIPP terminals in RSC outlast illumination

(A) eArch+ animals exhibit decreased performance once silencing occurred at T-1 (p=0.01, Bonferroni post-hoc test), T-2 (p=0.033), T-3 (p=0.036), not T-4 (p=0.078). (B) Two consecutively non-illuminated trials after illumination still result in decreased performance on the subsequent trial (p=0.023, Mann-Whitney U). All summary data depicted is mean+/−SD.

Animals spend more time at the choice point in Correct trials unless HIPP terminals in RSC are silenced

(A) CTRL animals spend more time at the choice point than eArch+ animals (p=0.02, Fixed Effect Omnibus Test). (B) This is true specifically of Correct trials both in NI and TI (respectively p=0.008 and p=0.045, Simple Effects Parameter Estimates). All summary data depicted is mean+/−SD.