Neural dynamics of PSTNTac1 neurons during active avoidance learning.

(A) Schematic of Cre-dependent GCaMP7f injection into the PSTN of Tac1-IRES-Cre mice.

(B) Example image showing GCaMP7f expression and fiber location in the PSTN. Yellow dashed line shows fiber track; white dashed line delineates the PSTN. Scale bar, 200 µm.

(C) Schematic of two-chamber active avoidance (2AA) task (upper) and the training pipeline (bottom). Red arrow lines refer to shuttle direction; Gray boxes and green boxes refer to 3-day habituation and 7-day training and recording, respectively.

(D) Behavioral protocol and definition of different behavioral readouts. In 2AA, a test mouse experienced a tone (Conditioned stimulus, CS) followed by an aversive foot shock (Unconditioned stimulus, US). Each trial has a constant 10-second-long time window and the trial-trial interval is ∼ 15–30 s. According to the shuttle onset in a given trial, it can be divided into three types: an avoidance trial (shuttle within 0–5 s and no shock), an escape trial (shuttle within 5–10 s and a short foot shock) and a failed trial (no shuttle and a maximum of shock). Green-filled boxes, CS duration; red-filled boxes, shock duration; purple lines, shuttle onset.

(E) Time course and line graph illustrating average avoidance rate and mean latency along 7-day training procedure. One-way repeated-measures ANOVA with Bonferroni post-hoc correction (For avoidance rate: ***P < 0.001; For mean latency: #P < 0.05, ##P < 0.01).

(F, G) Scatterplots illustrating average avoidance rate across five sessions (20 trials/session) on day 1 (F) and on day 7 (G). One-way repeated-measures ANOVA with Bonferroni post-hoc correction (*P < 0.05 and **P < 0.01).

(H, I) Heatmaps for pooling normalized fluorescence signals from the first 20 trials (early trials, H) and the last 20 trials (late trials, I) of each mouse. Time window for each trial is from 2 s before CS to 10 s after CS. Scale bar refers to % ΔF/F.

(J) The peri-stimulus time histogram (PSTH) plots for all early trials and late trials during the time window.

(K) Violin plots for the area under curve (AUC) of the average responses of early trials and late trials during CS. Two-tailed paired t test (***P < 0.001).

(L, M) Heatmaps for pooling normalized fluorescence signals from avoidance trials (L) and failure trials (M) of each mouse. Time window for each trial is from 2 s before CS to 10 s after CS. Scale bar refers to % ΔF/F.

(N) The PSTH plots for all avoidance trials and failure trials during the time window. O, Violin plots for the AUC of the average responses of avoidance trials and failure trials during CS. Two-tailed unpaired t test (***P < 0.001).

In (E-O), n = 6 mice; in (E-K), n = 20 early trials and 20 late trials each animal; in (G), n = 5 sessions from 100 trials each animal; and in (L-O), n = 536 avoidance trials and n = 64 failure trials. In (E, F, G, J, K, N and O), data was shown as mean ± SEM; in (K, O), shown as median. For detailed statistics information, see Supplementary Table 1.

Necessity of PSTNTac1 neurons mediating active avoidance learning.

(A) Schematic of Cre-dependent caspase3 injection into the PSTN of Tac1-IRES-Cre mice.

(B) Example images showing caspase injection (bottom) effectively ablated PSTN Tac1+ neurons whereas control GFP injection (upper) did not. Scale bar, 200 µm.

(C) Bar graph illustrating caspase ablation significantly reduced the number of GFP+ cells in the PSTN compared to the controls. Two-tailed Mann Whitney test (***P < 0.001).

(D) Schematic of two-chamber active avoidance task and training protocol. Gray boxes and orange boxes refer to 3-day habituation and 7-day training, respectively.

(E) Behavioral protocol and definition of different behavioral readouts.

(F, G) Average cumulative distribution of avoidance rates (F) and violin plots of latencies (G) for Caspase-ablated mice and EGFP-expressing controls on days 1-3, 5 and 7 of 2AA task. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction in (F) (ns = not significant; ***P < 0.001) and two-tailed unpaired t test in (G) (***P < 0.001).

(H, I) Line graphs of average avoidance rate (H) and mean latency (I) for caspase3-experessing mice and GFP-expressing controls along 7-day training. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction (**P < 0.01 and ***P < 0.001).

(J) Schematic showing viral injection and fiber implantation for optogenetic inhibition of PSTN Tac1+ neurons.

(K) Example image of GtACR expression and fiber location in the PSTN. Yellow dashed line shows fiber track; white dashed line delineates the PSTN. Scale bar, 200 μm.

(L) Schematic of two-chamber active avoidance task and training protocol. Gray boxes and purple boxes refer to 3-day habituation and 7-day training and inhibition, respectively.

(M) Behavioral protocol and definition of different behavioral readouts.

(N, O) Average cumulative distribution of avoidance rates (N) and violin plots of latencies (O) for GtACR-expressing mice and mCherry-expressing controls during CS inhibition on days 1-3, 5 and 7. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction in (N) (*P < 0.05, **P < 0.01 and ***P < 0.001) and two-tailed unpaired t test in (O) (***P < 0.001).

(P, Q) Line graphs of average avoidance rate (P) and mean latency (Q) for GtACR-expressing mice and mCherry-expressing controls along 7-day training. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction (***P < 0.001).

(R, S) Average cumulative distribution of avoidance rates (R) and violin plots of latencies (S) for GtACR-expressing mice and mCherry-expressing controls during US inhibition on days 1-3, 5 and 7. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction in (R) (*P < 0.05 and ***P < 0.001) and two-tailed unpaired t test in (s) (***P < 0.001).

(T, U) Line graphs of average avoidance rate (T) and latency (U) for GtACR-expressing mice and mCherry-expressing controls along 7-day training. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction (***P < 0.001).

In (F-I), n = 8 control mice and 8 Caspase-3 ablated mice; in (N-Q), n = 9 control mice and 7 GtACR-expressing mice; in (R-U), n = 8 control mice and 8 GtACR-expressing mice. In (F-I, N-Q, and R-U), data was shown as mean ± SEM; in (G, O and S), shown as median. For detailed statistics information, see Supplementary Table 1.

Sufficiency of PSTNTac1 neurons mediating active avoidance learning.

(A) Schematic showing viral injection and fiber implantation for optogenetic activation of PSTN Tac1+ neurons.

(B) Example image of injection site and ChR2 expression in the PSTN of Tac1-IRES-Cre animals. Yellow dashed line shows fiber track; white dashed line delineates the PSTN. Scale bar, 200 μm.

(C) Schematic of two-chamber active avoidance task and training pipeline. Gray boxes and blue boxes refer to 3-day habituation and 5-day training, respectively.

(D) Behavioral protocol and definition of different behavioral readouts.

(E, F) Average cumulative distribution of avoidance rates (E) and violin plots of latencies (F) for ChR2-experessing mice and mCherry-expressing controls on each day. Two-way repeated-measures with Bonferroni post-hoc correction in (E) (ns = not significant; *P < 0.05 and ***P < 0.001) and two-tailed unpaired t test in (F) (***P < 0.001).

(G, H) Line graphs of average avoidance rate (G) and mean latency (H) for ChR2-experessing mice and mCherry-expressing controls along 5-day training. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction (**P < 0.01 and ***P < 0.001).

(I) Schematic of experiment model.

(J) Schematic showing real-time place preference assay (RTPP). Light blue area indicates the chamber paired with light stimulation when the animal enters.

(K) Representative heatmaps showing locomotion trajectories of control (top) and ChR2-experessing (bottom) mice in the RTPP test.

(L) ChR2-expressing animals displayed a significant decrease in time spent in stimulation-coupled chamber compared to mCherry-expressing controls. Two-tailed Mann-Whitney test (***P < 0.001).

(M) Schematic of conditioned place aversion (CPA) test and experimental procedures.

(N) During post preference test on day 5, ChR2-expressing animals exhibited a remarkable reduction in time spent in preferred side on day 1 compared to mCherry-expressing controls, suggesting that activation of MeA GABAergic neurons could induce a long-term aversive memory for stimulation-coupled chamber. Two-tailed Mann-Whitney test (*P < 0.05).

In (E-H), n = 8 control mice and 14 ChR2-expressing mice; in (L), n = 9 control mice and 9 ChR2-expressing mice; in (N), n = 11 control mice and 8 ChR2-expressing mice. In (E-H, L, and N), data was shown as mean ± SEM; in (F), shown as median. For detailed statistics information, see Supplementary Table 1.

The PSTN projects to the PVT/IMD to mediate active avoidance learning.

(A) Schematic of Cre-dependent axon GCaMP6s injection into the PSTN and fiber placement in the PVT.

(B) Example image showing axon-GCaMP6s expression and fiber location in the PVT. Orange dashed line indicates fiber location and red dashed line delineates the PVT. Axon-GCaMP signals were strengthened by immunostaining with anti-GFP. Scale bar, 100 µm.

(C) Schematic of two-chamber active avoidance (2AA) task (upper) and the training pipeline (bottom).

(D) Behavioral protocol and definition of different behavioral readouts.

(E) PSTH plots for all early trials and late trials during the time window. Time window for each trial is from 2 s before CS to 10 s after CS.

(F) Violin plots for AUC of average responses of early trials and late trials during CS. Two-tailed paired t test (*P < 0.001).

(G) PSTH plots for all avoidance trials and failure trials during the time window. Time window for each trial is from 2 s before CS to 10 s after CS.

(H) Violin plots for AUC of average responses of avoidance trials and failure trials during CS. Two-tailed unpaired t test (*P < 0.001).

(I) Schematic showing eNpHR injection and fiber implantation for optogenetic inhibition of PSTN-PVT circuit.

(J) Example image of eNpHR expression in axon terminals in the PVT. Orange dashed line indicates fiber location and white dashed line delineates the PVT. Scale bar, 200 μm.

(K, L) Line graphs of average avoidance rate (K) and latency (L) for eNpHR-expressing mice and mCherry-expressing controls along 7-day training. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction (**P<0.01 and ***P < 0.001).

(M) Schematic showing viral injection and fiber implantation for optogenetic activation of PSTN-PVT circuit.

(N) Example image of axon terminal expression of ChR2 in the PVT. Orange dashed line indicates fiber location and red dashed line delineates the PVT. axon terminal mCherry signals were strengthened by immunostaining with anti-mCherry. Scale bar, 200 μm.

(O, P) Line graphs of average avoidance rate (O) and mean latency (P) for ChR2-experessing mice and mCherry-expressing controls along 5-day training. Two-way repeated-measures ANOVA with Bonferroni post-hoc correction (*P < 0.05).

In (E-H), n = 3 axon-GCaMP-expressing mice; in (G, H), n = 213 avoidance trials and n = 87 failure trials; in (K, L), n = 4 control mice and 8 eNpHR-expressing mice; in (O, P), n = 8 control mice and 9 ChR2-expressing mice. In (G-J, K, L, O, P), shown as data was shown as mean ± SEM; in (F, H), shown as median. For detailed statistics information, see Supplementary Table 1.