vmPFC PV signals active avoidance

(A) Fiber photometry schematic. (B) GCaMP6f expression in vmPFC PV neurons. Scale bar, 100 μm. (C) Active avoidance task schematic. The green box indicates presence of the tone, which lasts till animal crosses. (D) Example vmPFC PV ΔF/F (red) and speed (black) during two successful avoidance trials. Vertical line indicates chamber crossing. Tone, light green. (E) vmPFC PV ΔF/F during successful avoidance trials, data aligned to chamber crossing. White ticks: chamber crossing. Black ticks: tone onset. Same example mouse as D. (F) Example average vmPFC PV ΔF/F (red) and speed (grey), aligned to chamber crossing (left) and movement initiation (right). Same example mouse as D. (G) Average vmPFC PV ΔF/F before chamber crossing (pre, 4 to 2 seconds before cross) and during chamber crossing (peri, 1 second before to 1 second after crossing) (N = 8 mice). **p < 0.01, paired t-test. Shaded regions indicate SEM.

vmPFC PV neural activity reflects the avoidance movement, not the predictive tone

(A) Schematic of a modified version of avoidance task including 10% short tone trials (S), where tone only lasts for 1.5 seconds regardless of animal’s action,10% long tone trials (L), where the tone lasts until 1.5 seconds after a successful avoidance, and 80% regular trials (R). (B-C) Average vmPFC PV ΔF/F aligned to (B) chamber crossing and (C) tone offset (N = 7 mice). S: short tone trials; R: regular trials; L: long tone trials. (D) Comparison between average vmPFC PV calcium activity between 0 to 0.1 seconds after avoidance chamber crossings in regular trials (R) and long tone trials (L). Each circle mark indicates average calcium activity across trials of one animal. (E) Comparison among average vmPFC PV calcium activity between 0 to 0.1 seconds after tone offsets in regular trials (R), short tone trials (S), and long tone trials (L). Each circle mark indicates average calcium activity across trials of one animal. (F-G) Average avoidance kernel (F) and average tone offset kernel (G) across all animals (N=7 mice) calculated by a linear model. (H) Loss of predictive power (ΔR2) in a reduced model with shuffled avoidance or tone offset time points. n.s. p>0.05, *p < 0.05, **p < 0.01, (D): one-way repeated measure ANOVA, (E): paired t-test, (H): one-sample t-test. Error bars indicate SEM. Shaded regions indicate SEM.

vmPFC PV activity does not signal reward-approach movements during reward seeking behavior.

(A) Reward approach task schematic. (B) Example vmPFC PV ΔF/F (red) and speed (black) during two successful approach trials. Vertical line indicates chamber crossing. Tone, light green. (C) vmPFC PV ΔF/F during successful approach trials, data aligned to chamber crossing. Black ticks: tone onset, white ticks: chamber crossing, magenta ticks: licks. Same example mouse as B. See Figure 1E for comparison, wherein vmPFC PV are active at crossing (time = 0) during active avoidance task. (D) Example average vmPFC PV ΔF/F (red) and speed (grey), aligned to chamber crossing (left). Same example mouse as B. See Figure 1F for comparison, wherein vmPFC PV are active at crossing (time = 0) during active avoidance task. (E) Average vmPFC PV ΔF/F before chamber crossing (pre, 4 to 2 seconds before cross) and during chamber crossing (peri, 1 second before to 1 second after crossing) (N = 7 mice). n.s. p > 0.05, paired t-test. Shaded regions indicate SEM.

vmPFC PV activity becomes correlated to movement after shock.

(A-B) Example vmPFC PV ΔF/F (red) and speed (black) during (A) habituation prior to the first-ever shock exposure, and (B) the inter-trial interval after avoidance training. (C-D) Evolution of cross-covariance between vmPFC PV activity and speed over trials (upper panel, black) and corresponding average avoidance success rate across animals (bottom panel, cyan) on (C) the first day of avoidance training (N = 7 mice) and (D) in a well-learned phase in approach (N = 6 mice). Error bars indicate SEM.

Suppressing vmPFC PV neurons delays avoidance.

(A) Optogenetic schematic. (B) NpHR-eYFP expression in vmPFC PV neurons in a PV-Cre mouse. Scale bar: 200 μm. (C) Optogenetic inactivation schematic. Interleaved simulation blocks were introduced, where vmPFC PV neurons were optogenetically inhibited during the interval from 0.5 to 2.5 seconds after tone onset. (D) Speed of NpHR-expressing mice during avoidance with (red) and without (dark brown) illumination. (E) Speed of NpHR- and eYFP-expressing mice, averaged over the laser stimulation period, with and without illumination. (F) Distribution of crossing latencies during avoidance in NpHR-expressing animals with (red) and without illumination (brown). (G) Ratio of illuminated/non-illuminated chamber crossing probabilities. (H-K) Same as (D-G) for approach task. (L-M) Same as (D-E) but for OFT with dim light. (N-O) Same as (L-M) but for OFT with bright light. *p < 0.05, **p < 0.01. For (E), (I), (M), and (O), 2- way ANOVA interaction term. For (G) and (K), unpaired t-test. Shaded regions indicate SEM. NpHR: N=8 mice, eYFP: N=5 mice.

Suppressing vmPFC PV neurons delays avoidance and prolongs freezing

The video displayed the mouse’s behavior in response to shock-predicting cues without suppression of vmPFC PV neurons (as shown in the clip “avoidance without stimulation”) and with suppression of vmPFC PV neurons (as shown in the clip “avoidance with vmPFC PV suppression”). The shock- predicting cues were on when the caption “Tone on” was displayed. The laser stimulation to suppress vmPFC PV neurons was given in the time frames where “Laser on” was displayed.

(A) Average vmPFC PV ΔF/F (red) and speed (grey), aligned to chamber crossing in avoidance from PV::GFP mice. (N=2 mice). (B) vmPFC PV ΔF/F during successful avoidance trials, data aligned to tone onset. White ticks: cue onset. Black ticks: chamber crossing. (C) Example average vmPFC PV ΔF/F (red) and speed (grey), aligned to tone onsets. Same example mouse as B. (D) Example average vmPFC PV ΔF/F (red) and speed (grey) of long latency trials (avoidance latency longer than 3 seconds), aligned to tone onset. Same example mouse as B. (E) Average vmPFC PV ΔF/F before tone onset (pre, 4 to 2 seconds before tone onset) and during tone onset (peri, 0 to 2 seconds after tone onset) of long latency trials (avoidance latency longer than 3 seconds) (N = 8 mice). (F) Correlation between maximum speed and peak vmPFC PV ΔF/F at avoidance. Each dot represents an avoidance trial, and trials from one animal are marked in the same color (N = 8 mice). (G) Distribution of slope of linear correlation between maximum speed and peak vmPFC PV ΔF/F at avoidance. One dot represents one animal (N = 8 mice). (H) Clustering of distribution of z score of maximum speed and z score of vmPFC PV ΔF/F at avoidance. (I-K) It is the same as (F-H), but the correlation was calculated between avoidance latency and maximum vmPFC PV ΔF/F at avoidance (N = 8 mice). For (E), paired t-test. For (G) and (J), one-sample t-test. Shaded regions indicate SEM.

(A) Average vmPFC PV ΔF/F (red) and speed (grey), aligned to chamber crossing in approach from PV::GFP mice. (N=2 mice). (B) Schematic of a modified reward approach task with reward omissions. (C) Average vmPFC PV ΔF/F (top) and speed (bottom) of rewarded and omitted trials, aligned to chamber crossing (N=5 mice). (D) Average vmPFC PV ΔF/F (top) and speed (bottom) of rewarded trials, aligned to the first lick after chamber crossing (N=5 mice). (E) OFT schematic. (F) Example vmPFC PV ΔF/F (red) and speed (black) during OFT under interleaving bright lighting (white shading) and dim lighting epochs (grey shading). (G) vmPFC PV ΔF/F during OFT, aligned to maximum speed of each movement epoch in OFT. White ticks: maximum speed of movement epoch. Same example mouse as F. (H) Example average vmPFC PV ΔF/F (red; purple) and speed (grey; dark blue), aligned to maximum speed of each movement epoch in OFT under bright lighting (red; grey) and under dim lighting (purple; dark blue). Same example mouse as F. (I) Average vmPFC PV ΔF/F before movement (pre, 4 to 2 seconds before peak movement) and during movement (peri, 1 second before to 1 second after peak movement) under bright ceiling lighting (red) and dim ceiling lighting (blue) (N = 7 mice). (J) Average vmPFC PV ΔF/F (red; purple) and speed (grey; dark blue), aligned to the maximum speed of each movement epoch in OFT from PV::GFP mice (N=2 mice) under bright lighting (red; grey) and under dim lighting (purple; dark blue). For (I), n.s., p > 0.05, paired t-test. Shaded regions indicate SEM.

Cross covariance between movement speed and PV activity in avoidance and approach

(A) Evolution of cross-covariance between movement speed and PV activity in habituation and inter-trial intervals on the first day of avoidance training (N=7 mice). (B) Evolution of cross-covariance between movement speed and PV activity in habituation and inter-trial intervals of well-trained animals in an approach task (N=6 mice).

Suppression of vmPFC PV neurons in control animals.

(A) Speed of eYFP-expressing mice during avoidance with (red) and without (dark brown) illumination. (B) Distribution of crossing latencies during avoidance in eYFP-expressing animals with (red) and without illumination (brown). (C-D) Same as (A-B) for approach. (E) Freezing probabilities of NpHR-expressing animals during avoidance with (red) and without (dark brown) illumination. (F) Difference in freezing probabilities between with (on) and without (off) illumination in NpHR-expressing mice (red) and in eYFP- expression mice (grey). (G) Difference in freezing probabilities, averaged over the laser stimulation period, between with (on) and without (off) illumination in NpHR-expressing mice (red) and in eYFP- expression mice (grey). (H) Change in ratio of freezing duration/avoidance latency of NpHR-expressing and eYFP-expression mice during illumination compared to non-illumination. (I) Crossing latencies during avoidance in NpHR-expressing mice, grouped by relative position prior or within a stimulation block. (J) Average crossing latencies of all non-illuminated trials (off), 1st trials (1st stim) and 2nd-6th trials in illuminated blocks (2-6th stim). (K) Crossing latencies during avoidance in NpHR-expressing mice, grouped by relative position within or after a stimulation block. (L) Average crossing latencies of all illuminated trials (stim), 1st trials (1st off) and 2nd-6th trials (2-6th off) in non-illuminated blocks. (M) Post- crossing optogenetic inactivation schematic. (N) Speed of NpHR-expressing mice during avoidance with (red) and without (dark brown) illumination. (O) Same as (N), but the y-axis was plotted on a log scale. (P) Change in speed of NpHR-expressing and eYFP-expression mice during illumination (0 to 2 s from avoiding chamber crossing) compared to non-illumination trials. (Q) The latency of the first trials in stimulation blocks is compared to the latency of the second trials in stimulation blocks, and the latency of the first trials after stimulation blocks is compared to the latency of the second trials after stimulation blocks. n.s. p > 0.05, *p < 0.05, **p < 0.01. For (G) and (H), un-paired t-test. For (J), (L) and (Q), paired t- test. For (P), 2-way ANOVA interaction term. Shaded regions indicate SEM. NpHR: N=8 mice, eYFP: N=5 mice.