Claims 42
From the paper Transformations in prefrontal ensemble activity underlying rapid threat avoidance learning · claims file · checked 2026-10-21: ready
- data-neurons
The paper analysed 1394 PL neurons from 9 shocked mice and 847 neurons from 7 non-shocked mice on day 1, and 1031 (shocked) and 967 (non-shocked) cells on day 2.
We identified 1394 neurons across 9 mice in Shock animals and 847 neurons across 7 mice in Non-shock animals.
STAR Methods, Miniscope data processing; Figure 1K and 1L legendsHow it will be judged
Counts of accepted cells per session and of mice with imaging, by group and day. Tolerance zero, as these are counts of the recorded cells. The day 2 counts are from the Figure 1K and 1L legends ("Cells analyzed... Day 2 1031" and "Day 2 967").
- data-registered-cells
Cell registration across days tracked about 40% of neurons, 463 cells from 6 shocked mice and 322 cells from 7 non-shocked mice.
we used cell registration to track ~40% of neurons across sessions (shocked mice ... 463 cells from 6 mice, NS controls ... 322 cells from 7 mice
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 2; Figure 1N legendHow it will be judged
Counts of cells matched between day 1 and day 2 (CellReg on the Minian footprints, Methods), by group. Tolerance zero for counts. If the dataset holds no registration, the claim cannot be tested without redoing it, which the author reports.
- fig1e-platform-time
During learning, shocked mice spent a progressively larger fraction of the tone on the safety platform, and this differed from non-shocked controls.
During learning, shocked mice exhibited progressive increases in time spent on the platform ... These behaviors were significantly different than NS controls
Results, Tracking changes in PL activity during PMA; Figure 1EHow it will be judged
Per mouse, the fraction of tone time on the platform in blocks of three tones on each day (opening comment). Two-way repeated measures ANOVA (group by block) with Geisser-Greenhouse correction and Sidak's post hoc tests, per day (Methods; Figure 1 legend, day 1 Shk N=11, NS N=9; day 2 Shk N=8, NS N=8). Reproduced when in shocked mice the day 1 block means rise from block 1 to block 4 with at least one Sidak-corrected block difference at p < 0.05, and at least one Sidak comparison of shocked against NS mice at a block is p < 0.05 on day 1 or day 2. The Sidak family is not stated; readings are all pairwise block comparisons within the shocked group, and each block against block 1; both are run.
- fig1f-successful-trials
Shocked mice showed a progressive increase in the proportion of successful trials, in which the footshock was completely avoided, which differed from non-shocked controls.
in the proportion of successful trials (where mice completely avoided a footshock)
Results, Tracking changes in PL activity during PMA; Figure 1FHow it will be judged
As fig1e-platform-time, with the per-mouse proportion of successful trials per block (success rule in the opening comment).
- fig1g-freezing
Shocked mice showed a progressive increase in freezing during the tone, which differed from non-shocked controls.
and in freezing levels
Results, Tracking changes in PL activity during PMA; Figure 1GHow it will be judged
As fig1e-platform-time, with the per-mouse fraction of tone time freezing per block (BehaviorDEPOT freezing labels).
- fig1h-latency
Shocked mice showed a progressive decrease in the latency to enter the platform after tone onset, which differed from non-shocked controls.
and a progressive decrease in latency to enter the platform during the tone
Results, Tracking changes in PL activity during PMA; Figure 1HHow it will be judged
As fig1e-platform-time, with the per-mouse latency per block (opening comment), and with the day 1 block means falling from block 1 to block 4 in shocked mice.
- fig1kl-overall-encoding-similar
In each period (BL, D1, D2), the overall proportion of PL neurons encoding PMA features was similar between shocked and non-shocked mice, and it increased modestly across learning.
In each period (BL, D1, D2), the overall proportion of PL neurons encoding PMA features was similar between groups and modestly increased across learning
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 1; Figure 1K and 1L (black and gray pie charts)How it will be judged
Per period and group, the fraction of all recorded cells (Figure 1K and 1L denominators) responsive to at least one of tone, platform or freezing (opening comment). The paper compares pie charts with chi-square (Methods; two-sided with Yates' correction, Figure 1 legend) and reports no test for this. Similar is judged by that test, a 2x2 chi-square with Yates' correction of responsive against non-responsive cells, shocked against NS, in each period; it holds when p >= 0.05 in all three periods. The increase holds when in each group the D2 fraction exceeds the BL fraction. Run under each period and frame reading (opening comment).
- fig1kl-day1-categories-similar
On day 1, shocked and non-shocked mice had similar proportions of PL neurons encoding the tone, the safety platform, freezing, or their combinations.
On day 1, shocked and NS control mice had similar proportions of PL neurons that encoded the tone, safety platform, freezing behavior, or their combinations
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 1; Figure 1K and 1L (colored wheel plots); Figure S1F, GHow it will be judged
For BL and D1, among the responsive cells, the counts of tone-only, platform-only, freezing-only and mixed cells, shocked against NS. The paper's test for these pie charts is the two-sided chi-square with Yates' correction (Methods; Figure 1 legend), applied as for the day 2 mixed cells, a 2x2 table of the category against the other responsive cells. Reproduced when no category gives p < 0.05 in either period; the differences in percentage points are given either way. Run under each period and frame reading. Figure S1F and S1G were not read.
- disc-tone-cells-baseline
In both shocked and non-shocked mice, the proportion of tone-encoding PL neurons was greatest in the baseline period, and the activity of tone-encoding neurons at tone onset was similar between groups.
We also revealed that in both shocked and NS control mice, the proportion of tone-encoding neurons was greatest during the baseline period, when tones are novel, and their activity during tone onset was similar across groups.
Discussion, paragraph 3; Figure 1K, 1L, 1P1 and 1Q1How it will be judged
Proportion, per group and period (BL, D1, D2), the fraction of all cells (Figure 1K and 1L denominators) that encode the tone. Which cells count as tone-encoding is open, reading G1, any cell responsive to the tone category (opening comment), and G2, tone-only cells (the colored wheel plots). The paper gives no test; greatest holds when in each group the BL fraction exceeds the D1 and the D2 fractions. Activity, as fig1p-platform-exit-excited-day2 for the cells excited and the cells suppressed at tone onset (the code's toneonVector and 5 s window; 61 and 52 shocked, 24 and 25 NS on day 1, 34 and 37 shocked, 34 and 22 NS on day 2, STAR Methods, Cells Analyzed for Figure 1), Welch's t-test of shocked against NS on each day; similar holds when none gives p < 0.05, with the differences reported. Reproduced when both parts hold. Run under each period and frame reading and under G1 and G2; where they give different outcomes the claim is methods_underspecified.
- fig1k-mixed-day2
By day 2, shocked mice had a significantly larger proportion of mixed-selective PL neurons (encoding more than one PMA feature) than non-shocked mice, 51% against 25%.
By day 2, shocked mice had a significantly greater proportion of PL neurons exhibiting mixed selectivity (i.e., encoding multiple features of PMA; shocked ... 51%, non-shocked ... 25%
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 1; Figure 1K-MHow it will be judged
Mixed cells on day 2 (period D2, opening comment), shocked against NS, 2x2 chi-square with Yates' correction (Figure 1K legend). The denominator is open. Reading R1, the responsive cells (the multi-colored pie charts show "profiles of" encoding cells, Figure 1K legend); reading R2, all day 2 cells (1031 and 967). A chi-square near 43 with 51% against 25% needs about 300 cells per group, which fits R1 and not R2, but both are shown. Reproduced when under a reading the percentages fall within tolerance and the test gives p < 0.05 with shocked higher. Run under each period and frame reading.
- fig1o-tone-to-mixed
Among registered cells, neurons that selectively encoded the tone at the end of day 1 were more likely to be mixed-selective on day 2 in shocked mice than in non-shocked mice.
In shocked mice, neurons that selectively encoded the tone at the end of day 1 were more likely to exhibit mixed selectivity on day 2 compared to NS controls
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 2; Figure 1OHow it will be judged
Registered cells only (Methods, Tracking cell encoding across days). Among the cells that were tone-only at D1, the fraction that are mixed at D2, shocked against NS, 2x2 chi-square with Yates' correction (Figure 1O legend). Reproduced when the shocked fraction is higher and p < 0.05. Run under each period and frame reading.
- fig1o-nonmod-to-mixed
Among registered cells, neurons not modulated at the end of day 1 were more likely to become mixed-selective on day 2 in shocked mice than in non-shocked mice.
Non-modulated neurons were also more likely to develop mixed selectivity on day 2 in shocked mice
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 2; Figure 1OHow it will be judged
As fig1o-tone-to-mixed, for the cells non-modulated (no tone, platform or freezing response) at D1.
- fig1p-platform-exit-excited-day2
On day 2, platform-excited neurons of shocked mice showed a large increase in activity after platform exits compared with non-shocked mice.
Shocked mice had a large increase in the activity of platform-excited neurons following platform exits on day 2
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 3; Figure 1P3How it will be judged
Method from compile_pooled_avg_heatmaps.m and trialAvg.m, where the Methods are silent. Cells are those called excited for platform exits on day 2, by an ROC over the whole session (the code uses the session-wide vectors; cell counts in the STAR Methods, Cells Analyzed for Figure 1). For each cell, DeltaF/F averaged over all exits of the session in a window from 4 s before to 4 s after the exit (exits whose window leaves the recording are dropped), z-scored by the mean and SD of its pre-exit half, and the area (trapezoid sum over frames at 15 frames/s) of the post-exit half. Welch's t-test of these per-cell areas, cells pooled across mice, shocked against NS (Figure 1P legend). Reproduced when the shocked mean is higher and p < 0.05.
- fig1q-platform-exit-inhibited-day2
On day 2, the activity of platform-inhibited neurons after platform exits did not differ between shocked and non-shocked mice.
but no differences in the activity of platform-inhibited neurons
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 3; Figure 1Q3How it will be judged
As fig1p-platform-exit-excited-day2, for cells called suppressed for platform exits on day 2 (74 shocked, 82 NS in the STAR Methods). The paper's test is Welch's t-test at p < 0.05; reproduced when it gives p >= 0.05, with the difference of means reported.
- fig1pq-freezing-day2
On day 2 there were subtle differences between shocked and non-shocked mice in the activity of freezing-modulated neurons after freezing onset and offset.
On day 2, there were subtle changes in activity following freezing onset and offset
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 3; Figure 1P2 and 1Q2How it will be judged
As fig1p-platform-exit-excited-day2, for freezing-excited and freezing-suppressed cells on day 2 (171 and 283 shocked, 150 and 198 NS), aligned to freezing onset and to freezing offset with a 3 s window on each side (code). The paper does not say which comparison differed or in which direction. Reproduced when at least one of the four Welch's t-tests (excited or suppressed, onset or offset) gives p < 0.05; the four results are reported.
- fig1r-mixed-exit-day1
Excited mixed-selective neurons showed a significant increase in activity during platform exits in shocked mice on day 1 only, not on day 2.
Excited mixed selective neurons exhibited a significant increase in activity during platform exits only on day 1
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 3; Figure 1R3How it will be judged
As fig1p-platform-exit-excited-day2, for mixed cells (opening comment) that are excited, aligned to platform exits, on day 1 and on day 2. Which mixed cells enter is open, readings M1, mixed cells with an excited platform-exit call, and M2, mixed cells whose calls include any excited call (the code's split into excited and suppressed mixed cells counts a cell as excited when any call is excited). Reproduced when Welch's t-test gives shocked higher with p < 0.05 on day 1 and p >= 0.05 on day 2.
- figs1h-exit-not-shock
The larger platform-exit response of platform-excited neurons in shocked mice on day 2 was not due to premature exits followed by footshocks.
This was not due to premature platform exits resulting in footshocks
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 3; Figure S1HHow it will be judged
The Methods (ROC analysis) split exit bouts by whether the 2 s after the exit hold a shock. As fig1p-platform-exit-excited-day2 with exits followed by a shock within 2 s removed. Reproduced when shocked mice remain higher with Welch's p < 0.05. Figure S1H was not read.
- figs2-innate-behaviors
Escape, jumping and rearing were rare; shocked mice jumped more on day 1; few PL neurons encoded these behaviors and their activity was similar between groups, except escape-excited neurons, which were more active in shocked mice on both days.
While these behaviors were relatively rare ( Figure S2C - E ), shocked mice jumped more on day 1 ... Shocked mice had significantly higher levels of activity in escape-excited neurons on both days
Results, Learning-induced increase in PL neurons with mixed-selective encoding, paragraph 4; Figure S2How it will be judged
Figure S2 was not read, and its tests are not in the main text; the Methods give Welch's t-test for comparisons not otherwise named. Jumping, per mouse, the number of jumping bouts on day 1, shocked against NS, Welch's t-test p < 0.05 with shocked higher. Activity, as fig1p-platform-exit-excited-day2, for escape-excited cells aligned to escape onset, on each day, Welch's t-test p < 0.05 with shocked higher on both days. Similar activity, as fig1p-platform-exit-excited-day2 for the other cells excited or suppressed for escape, jumping or rearing, on each day, Welch's t-test; it holds when none gives p < 0.05 (the paper's test, with the differences reported). That these behaviors were rare and that few neurons encoded them is reported as counts and not judged, as Figure S2 was not read. The averaging window for escape, jumping and rearing is not given; the code uses 5 s for the tone, 4 s for the platform and 3 s for freezing on each side of the event, so the readings are 3, 4 and 5 s; where they give different outcomes the claim is methods_underspecified. Reproduced when the jumping, escape-excited and similar-activity parts all hold.
- fig2-day1-tone-stronger
On day 1, neurons encoding platform entries or exits were more strongly modulated when the entry or exit occurred during the tone than during the inter-tone interval, in both shocked and non-shocked mice.
On day 1, neurons encoding platform entries or exits were more strongly modulated during the tone for both shocked and NS control mice
Results, Enhanced encoding of avoidance actions during threatening cues, paragraph 1; Figure 2C1 and 2D1How it will be judged
Per cell called excited or suppressed for platform entries (or exits) during the tone, and separately during the ITI (code vector names such as itiPlatentVector), the auROC, with suppressed values folded as 1 - auROC (Methods, Frequency distributions of auROC values). Ordinary two-way ANOVA (group by tone/ITI) with Fisher's LSD (Methods). What enters the ANOVA is open, reading H1, the folded auROC of each cell pooled across mice, and H2, the per-mouse modulation scores of fig2e-modulation-entries. Reproduced when, for entries and for exits, the tone values exceed the ITI values with Fisher's LSD p < 0.05 in each group.
- fig2-day2-tone-further
On day 2, shocked mice showed a further increase in the strength of platform entry and exit encoding during the tone.
On day 2, shocked mice had a further increase in encoding strength during the tone
Results, Enhanced encoding of avoidance actions during threatening cues, paragraph 1; Figure 2C2 and 2D2How it will be judged
As fig2-day1-tone-stronger, on day 2. Reproduced when, for entries and for exits, the shocked tone values exceed the NS tone values with Fisher's LSD p < 0.05, and exceed the shocked day 1 tone values.
- fig2e-modulation-entries
Per animal, shocked mice had significantly higher modulation scores than non-shocked mice for platform entries occurring during the tones on day 2.
Again, shocked mice had significantly higher modulation scores for both platform entries and exits occurring during the tones on day 2
Results, Enhanced encoding of avoidance actions during threatening cues, paragraph 1; Figure 2EHow it will be judged
Per mouse, the folded auROC values of its cells called for platform entries during the tone (or during the ITI); mice with fewer than 3 such cells are left out (Methods). The modulation score is 1 minus the area under the mouse's cumulative frequency histogram (Methods). For an exact cumulative distribution over the auROC axis from 0.5 to 1, this is the mean folded auROC; over an axis rescaled to 0 to 1 it is 2 times that mean minus 1; the two give the same ANOVA. The histogram bins are not given (Prism); the exact cumulative distribution is used, a standard choice that changes the score only by binning. Two-way ANOVA (group by tone/ITI) with Fisher's LSD (Methods; Figure 2E legend, day 2 Shk N=7, NS N=7). Reproduced when on day 2 the shocked tone score exceeds the NS tone score with p < 0.05.
- fig2f-modulation-exits
Per animal, shocked mice had significantly higher modulation scores than non-shocked mice for platform exits occurring during the tones on day 2.
shocked mice had significantly higher modulation scores for both platform entries and exits occurring during the tones on day 2 ( Figure 2E , F )
Results, Enhanced encoding of avoidance actions during threatening cues, paragraph 1; Figure 2FHow it will be judged
As fig2e-modulation-entries, for platform exits (Figure 2F legend, day 2 Shk N=6, NS N=7).
- fig3a-decoding-shock-gt-ns
Decoding of trial outcome from all recorded PL neurons on day 1 was significantly better in shocked than in non-shocked mice in the pre-shock and shock epochs.
Decoder performance using all recorded neurons was significantly greater for shocked mice than NS controls in the pre-shock and shock epochs
Results, Population dynamics prospectively represent trial outcomes early in learning; Figure 3AHow it will be judged
Decoding as in the opening comment, all cells, day 1, per mouse (Figure 3 legend, Shk N=7, NS N=7). Two-way repeated measures ANOVA (group by epoch) with Geisser-Greenhouse correction and Sidak's test of shocked against NS at each epoch (Methods). Reproduced when shocked mice are higher with Sidak p < 0.05 in the pre-shock and in the shock epoch. Run under C1 and C2, N1 and N2; the code's pre-shock epoch (22 to 27 s) is shown as a variant.
- fig3a-decoding-before-tone
Decoding performance in shocked mice was already better than in non-shocked mice before tone onset.
Notably, decoding performance was already better even before the onset of the tone.
Results, Population dynamics prospectively represent trial outcomes early in learning, paragraph 1; Figure 3AHow it will be judged
As fig3a-decoding-shock-gt-ns. The Methods define no epoch before the tone; the code's pre-cue epoch, -5 to 0 s, is the one used. The paper reports no test for it. Reproduced when the shocked mean over that epoch exceeds the NS mean; the Sidak comparison is reported beside it and decides nothing.
- fig3a-decoding-peak-preshock
In shocked mice on day 1, decoding of trial outcome peaked immediately before the shock.
we found that decoder performance peaked immediately before the shock, and further showed that decoding trial outcomes was possible before animals established a stable avoidance strategy
Discussion, paragraph 5; Figure 3AHow it will be judged
Decoding as in fig3a-decoding-shock-gt-ns, all cells, shocked mice, day 1. The paper gives no test. What "peaked" is measured on is open, reading P1, of the four Methods epochs the pre-shock epoch (23 to 28 s) has the highest mean normalized F1 across mice; P2, the highest bin of the across-mouse mean trace from 0 to 38 s after tone onset lies in the pre-shock epoch. Run under C1 and C2, N1 and N2; where the readings give different outcomes the claim is methods_underspecified.
- fig3b-feature-cells-enhance
Decoding from only ROC-defined feature-encoding neurons enhanced decoder performance in shocked mice.
Using only feature-encoding neurons enhanced decoder performance in shocked mice
Results, Population dynamics prospectively represent trial outcomes early in learning, paragraph 2; Figure 3BHow it will be judged
Decoding as in fig3a-decoding-shock-gt-ns with modulated cells only (opening comment). What "enhanced" is compared with is open. Reading E1, against the all-cell decoder of the same shocked mice, the mean normalized F1 in the pre-shock and shock epochs is higher with modulated cells. Reading E2, against NS mice decoded the same way, the two-way repeated measures ANOVA with Sidak's test gives shocked higher with p < 0.05 in at least as many epochs as with all cells. Both are run, each under the cell readings K1 and K2; when they disagree the claim is methods_underspecified.
- fig3c-nonmod-cells-worsen
Decoding from only non-modulated PL neurons worsened decoder performance.
whereas using only the non-modulated PL neurons worsened performance
Results, Population dynamics prospectively represent trial outcomes early in learning, paragraph 2; Figure 3CHow it will be judged
As fig3b-feature-cells-enhance with non-modulated cells only, and with lower in place of higher (E1) and fewer significant epochs in place of at least as many (E2).
- fig4b-hybrid-model-best
The fraction of tone time on the platform was best fitted by the hybrid Rescorla-Wagner model, which learns from both shock and avoid trials.
The fraction of time on the platform was best modeled by the hybrid approach
Results, PL population activity states relate to individual learning rates, paragraph 1; Figure 4A-CHow it will be judged
The three models of the opening comment fitted per shocked mouse (N=9, Figure 4B legend). Per mouse and model, the mean over blocks of three trials of the squared residual between the predicted and observed fraction (Figure 4A and 4B legends). Repeated measures one-way ANOVA with Geisser-Greenhouse correction (Methods). Reproduced when the hybrid model has the lowest mean squared residual and the ANOVA gives p < 0.05. Run under T1 and T2, S1 and S2.
- fig4e-day1-onset-distance
On day 1, the mean Euclidean distance between latent trajectories of successful and unsuccessful trials in the tone-onset epoch was positively correlated with the avoid learning rate.
On days 1 and 2, the mean Euclidean distance during the tone-onset epoch, but not during the pre-shock epoch, was significantly correlated with learning rate
Results, PL population activity states relate to individual learning rates, paragraph 2; Figure 4EHow it will be judged
Mean Euclidean distance (opening comment) in the tone-onset epoch on day 1, per shocked mouse, against the hybrid model's avoid learning rate; Pearson's correlation, N=8 (Figure 4E legend). Reproduced when r > 0 (positive per the Discussion, "positively correlated with learning rates ( Figure 4E , G , J )") and p < 0.05. The paper's r and p are in Table S1, not read. Run under W1, W2 and W3, L1 and L2, T1 and T2, S1 and S2.
- fig4f-day1-preshock-distance-none
On day 1, the mean Euclidean distance in the pre-shock epoch was not significantly correlated with the learning rate.
but not during the pre-shock epoch
Results, PL population activity states relate to individual learning rates, paragraph 2; Figure 4FHow it will be judged
As fig4e-day1-onset-distance in the pre-shock epoch (23 to 28 s). The paper's test is Pearson's correlation at p < 0.05; reproduced when p >= 0.05, with r reported.
- fig4g-day2-onset-distance
On day 2, the mean Euclidean distance between successful and unsuccessful trajectories in the tone-onset epoch was positively correlated with the avoid learning rate.
On days 1 and 2, the mean Euclidean distance during the tone-onset epoch ... was significantly correlated with learning rate
Results, PL population activity states relate to individual learning rates, paragraph 2; Figure 4GHow it will be judged
As fig4e-day1-onset-distance on day 2, under D-a and D-b as well.
- fig4h-day2-preshock-distance-none
On day 2, the mean Euclidean distance in the pre-shock epoch was not significantly correlated with the learning rate.
but not during the pre-shock epoch, was significantly correlated with learning rate
Results, PL population activity states relate to individual learning rates, paragraph 2; Figure 4HHow it will be judged
As fig4f-day1-preshock-distance-none on day 2, under D-a and D-b as well.
- fig4i-day1-onset-trajectory-none
On day 1, trajectory distance in the tone-onset epoch on the trials of interest was not significantly correlated with learning rate.
On day 1, there were no significant correlations between trajectory distance and learning rate during the tone onset epoch
Results, PL population activity states relate to individual learning rates, paragraph 3; Figure 4IHow it will be judged
Trajectory distance (opening comment, Methods measure) in the tone-onset epoch on the first shock trial and on the first avoidance trial of day 1, per shocked mouse, against the avoid learning rate; Pearson's correlation, N=8 (Figure 4I legend). The paper's test is Pearson's at p < 0.05; reproduced when every trial of interest gives p >= 0.05, with r reported. The code's distance and its extra trials (last shock, last avoid) are shown as variants. Run under W1, W2 and W3, L1 and L2.
- fig4j-day1-preshock-first-avoid
On day 1, trajectory distance in the pre-shock epoch on the first avoidance trial was positively correlated with learning rate.
but there was a positive correlation during the pre-shock epoch on the first avoidance trial
Results, PL population activity states relate to individual learning rates, paragraph 3; Figure 4JHow it will be judged
As fig4i-day1-onset-trajectory-none in the pre-shock epoch, first avoidance trial. Reproduced when r > 0 with p < 0.05. The Discussion calls this correlation trend-level ("some trend-level correlations (e.g. Figure 4I , J )"); the Results are followed, and the r and p are reported whatever the outcome.
- fig4k-day2-onset-first-shock
On day 2, trajectory distance in the tone-onset epoch on the first shock trial was negatively correlated with learning rate.
On day 2, during the tone onset period, we observed a significant negative correlation between the trajectory distance and learning rate on the first shock trial
Results, PL population activity states relate to individual learning rates, paragraph 3; Figure 4KHow it will be judged
As fig4i-day1-onset-trajectory-none on day 2, first shock trial. Reproduced when r < 0 with p < 0.05. Run under D-a and D-b as well.
- fig4l-day2-preshock-trajectory-none
On day 2, trajectory distance in the pre-shock epoch on the trials of interest was not significantly correlated with learning rate.
and no significant correlations during the pre-shock period
Results, PL population activity states relate to individual learning rates, paragraph 3; Figure 4LHow it will be judged
As fig4i-day1-onset-trajectory-none on day 2 in the pre-shock epoch. Reproduced when every trial of interest gives p >= 0.05.
- figs3ab-day1-post-onset-distance
On day 1, the mean Euclidean distance 5 to 10 s after tone onset was positively correlated with learning rate, while the distance just before tone onset was not.
On day 1, we also observed a significant positive correlation between Euclidean distance and learning rate during the 5–10s period after (but not in the moments before) the tone onset
Results, PL population activity states relate to individual learning rates, paragraph 2; Figure S3A, BHow it will be judged
As fig4e-day1-onset-distance in the post-onset window (5 to 10 s) and the pretone window (-6 to -1 s) (Methods). Reproduced when the post-onset correlation has r > 0 and p < 0.05 and the pretone correlation has p >= 0.05. Figure S3 was not read.
- figs3cd-day1-distance-groups-similar
On day 1, the Euclidean distances between successful and unsuccessful population trajectories were comparable in shocked and non-shocked mice.
On day 1, Euclidean distances between population vectors for successful and unsuccessful trials were comparable for shocked and NS mice
Results, PL population activity states relate to individual learning rates, paragraph 4; Figure S3C, DHow it will be judged
Mean Euclidean distance (opening comment) on day 1 in the tone-onset and pre-shock epochs, per mouse, shocked against NS. Figure S3 was not read; the test is Welch's t-test (Methods, "All other comparisons"). Reproduced when both epochs give p >= 0.05, with the differences reported. Run under W1, W2 and W3.
- figs3ef-first-success-trajectory
Following the first successful trial, trajectory distance was significantly higher in shocked than in non-shocked mice in the tone-onset epoch and significantly lower in the pre-shock epoch.
Following the first successful trial, trajectory distance was significantly higher in shocked mice during tone onset, but significantly lower than NS controls during the pre-shock epoch
Results, PL population activity states relate to individual learning rates, paragraph 4; Figure S3E, FHow it will be judged
Trajectory distance (opening comment) on day 1, per mouse, shocked against NS, Welch's t-test (Methods). Which trial "following the first successful trial" means is open, reading A1, the trial after the first avoidance trial, and A2, the first avoidance trial itself (the code's event-aligned distances run from it to the trial after). Reproduced when shocked mice are higher at tone onset and lower at pre-shock, each with p < 0.05. Figure S3 was not read.
- figs3gh-behavior-similar
Behaviour of shocked and non-shocked mice was similar during the tone-onset and pre-shock epochs.
Behaviors of shocked and NS mice were similar during both epochs
Results, PL population activity states relate to individual learning rates, paragraph 4; Figure S3G, HHow it will be judged
Figure S3G and S3H were not read. The same paragraph names avoidance and freezing as the behaviours compared, so the measures are the fraction of time on the platform and the fraction of time freezing in the tone-onset and pre-shock epochs, per mouse. Which trials enter is open, reading B1, day 1 trials 4 to 12, and B2, the trial of interest of figs3ef-first-success-trajectory (the sentence follows that result). Welch's t-test (Methods), shocked against NS. Reproduced when no comparison gives p < 0.05, each reading shown.
- figs4ab-iti-midtone-none
In the ITI and mid-tone epochs, trajectory distance was not significantly related to learning rate.
We also examined epochs during the ITI and in the middle of the tone, finding no significant relationships between trajectory distance and learning rate in either
Results, PL population activity states relate to individual learning rates, paragraph 4; Figure S4A, BHow it will be judged
As fig4i-day1-onset-trajectory-none in the mid-ITI (-25 to -20 s) and mid-tone (14 to 19 s) windows (Methods), on days 1 and 2. Reproduced when no correlation gives p < 0.05. Figure S4 was not read.
- figs4cd-iti-before-first-avoid
The only differences in trajectory distance between shocked and non-shocked mice in the control epochs occurred during the ITI before the first avoidance trial.
The only differences in trajectory distance between shocked and NS mice occurred during the ITI before the first avoidance trial
Results, PL population activity states relate to individual learning rates, paragraph 4; Figure S4C, DHow it will be judged
As figs3ef-first-success-trajectory in the mid-ITI and mid-tone windows, on each trial of interest. Reproduced when the mid-ITI window on the first avoidance trial gives p < 0.05 and no other comparison does. Figure S4 was not read.
Pipeline
- Paper
- done (1 of 1)
- Claims
- done
- Claims check
- done (ready)
- Attempt
- todo
- Questions
- none
- Verified
- none