Claims 56
From the paper Hippocampal place code plasticity in CA1 requires postsynaptic membrane fusion · claims file · checked 2026-10-21: ready
- fig1c-novelty-preference
In the freely moving novel-arm T-maze test trial, control mice preferred the previously unseen arm while Cre mice explored the arms at chance, and the novel minus familiar difference score differed between groups.
During the test trial, control mice preferred to explore the previously unseen arm, while Cre-injected mice explored the different arms at chance levels
Figure 1C; Results, Stx3-dependent postsynaptic membrane fusion in CA1 is critical for novel environment preference and reward anticipation, paragraph 1How it will be judged
Difference score per mouse is time in the novel arm minus time in the familiar arm in the 2 min test trial, which begins when all four paws leave the start arm (Methods, Freely-Moving Novelty T-Maze). Unpaired t-test of control against Cre (Fig 1C legend), run as Student's (the Prism default for an unpaired t-test) and as Welch's. Reproduced when p < 0.05 with the control mean higher under both. The paper gives no test of the Cre score against zero; its mean and a one-sample t-test are reported in the notes only.
- figs1c-training-occupancy-no-difference
During novelty T-maze training, control and Cre mice occupied the three maze compartments at similar rates (no virus effect).
Control and Cre mice occupied the 3 zones of the maze during training at similar rates
Figure S1CHow it will be judged
A claim of absence. Per mouse, the fraction of time in each of the three training compartments (start, intermediate zone, familiar arm), averaged over all training trials (Fig S1C legend); repeated measures ANOVA with compartment within and virus between, as the legend's rmANOVA. Reproduced when the virus main effect has p >= 0.05, not reproduced when p < 0.05; the group difference is reported either way.
- figs1e-open-field-distance-no-difference
CA1 Stx3 deletion does not change the distance travelled in the open field.
CA1 deletion of Stx3 does not affect basal locomotion as demonstrated by the summary plot of distance traveled in the open field
Figure S1EHow it will be judged
A claim of absence. Total distance in the 30 min open field session (Methods, Open field), unpaired t-test as in Fig 1C. Reproduced when p >= 0.05; the difference in means is reported either way.
- figs1f-epm-closed-arm-no-difference
CA1 Stx3 deletion does not change the percentage of time spent in the closed arms of the elevated plus maze.
CA1 deletion of Stx3 does not affect anxiety-like behavior as demonstrated by the summary plot of %time spent in the closed arm of the elevated plus maze
Figure S1F, leftHow it will be judged
A claim of absence. Percentage of the 5 min session spent in the closed arms (Methods, Elevated plus maze), unpaired t-test as in Fig 1C. Reproduced when p >= 0.05.
- figs1f-epm-distance-no-difference
CA1 Stx3 deletion does not change the distance travelled in the elevated plus maze.
CA1 deletion of Stx3 does not affect locomotion in the elevated plus maze as demonstrated by the summary plot of distance traveled
Figure S1F, rightHow it will be judged
A claim of absence. Total distance in the 5 min session, unpaired t-test as in Fig 1C. Reproduced when p >= 0.05.
- fig2ab-example-cell-counts
The example sessions of Fig 2A-B hold 1208 cells and 120 familiar trials (Ctrl-3, day 5), 1568 cells and 20 novel trials (Ctrl-3, day 1), 1087 cells and 17 familiar trials (Cre-3, day 5) and 936 cells and 17 novel trials (Cre-3, day 1).
(Ctrl-3, N=1208 cells, 120 trials). Right - Same as (Left) for novel trials on day 1 (N=1568 cells, 20 trials)
Figure 2A and 2B legendsHow it will be judged
Counts of mCherry+ cells (all analysed cells of the session) and of completed trials of the given type in the named sessions. Each count is matched within its tolerance. The Cre-3 numbers are in the Fig 2B legend ("Day 5 familiar trial decoding - N=1087 cells, 17 trials, Day 1 novel trial decoding - N=936 cells, 17 trials").
- fig1f-anticipatory-licking-familiar
On familiar arm trials, control mice licked at a higher rate than Cre mice as they approached the reward zone, across the six days.
Control mice licked at a higher rate as they approached the reward zone for both familiar and novel arms (Fig. 1F-G)
Figure 1F; Results, Stx3-dependent postsynaptic membrane fusion in CA1 is critical for novel environment preference and reward anticipation, paragraph 4How it will be judged
Anticipatory lick rate is the lick rate (licks/s) in the 30 cm preceding the reward zone of the arm (Methods, VR behavioral analyses), on familiar arm trials, averaged over the trials of each day (Fig 1F legend, "across trial average for each mouse"), days 1 to 6. Mixed effects ANOVA (virus between, day within). Open and run as readings, the Cre mice (a) Cre-1 to Cre-7 or (b) Cre-1 to Cre-8 (legend N=8, df 15); and the rate (i) licks over time spent in the 30 cm per trial or (ii) the mean of the three 10 cm bins' rates. All familiar trials of the day are used, rewarded or not, as no exclusion is stated. Reproduced when the virus main effect has p < 0.05 with control higher. The printed F and p disagree (F(1,15)=10.66 gives p near 0.005); either p is below 0.05.
- fig1g-anticipatory-licking-novel
On novel arm trials, control mice licked at a higher rate than Cre mice as they approached the reward zone.
Control mice licked at a higher rate as they approached the reward zone for both familiar and novel arms (Fig. 1F-G)
Figure 1G; Results, Stx3-dependent postsynaptic membrane fusion in CA1 is critical for novel environment preference and reward anticipation, paragraph 4How it will be judged
As fig1f-anticipatory-licking-familiar, on novel arm trials (block 6 of days 1 to 5, all novel trials of day 6), with the same readings of the Cre mice and of the rate. Reproduced when the virus main effect has p < 0.05 with control higher.
- vr-reward-location-learning
Both control and Cre mice learned the context-dependent reward locations, licking more and running slower only in the bins just before the reward zones, and anticipatory licking on familiar trials changed over days.
Both Cre and control mice rapidly learned the context-dependent reward locations, evidenced by increased licking and decreased running speed only in the spatial bins immediately preceding the reward zones
Results, Stx3-dependent postsynaptic membrane fusion in CA1 is critical for novel environment preference and reward anticipation, paragraph 4; Figures 1F and S2B-EHow it will be judged
The paper gives no test of the spatial pattern; the comparison its sentence makes is between the bins preceding the reward zone and the rest of the track. On day 6, for each group and each arm (left, right), from the across-animal mean of position-binned normalized lick rate and normalized running speed (each mouse's binned value over its overall mean, Methods, VR behavioral analyses; Fig S2C and S2E), the pattern holds when the mean normalized lick rate in the 30 cm before the reward zone exceeds that of every bin outside the reward zone and those 30 cm, and the normalized speed in those 30 cm is below 1. The day effect holds when the day main effect of the fig1f ANOVA has p < 0.05 with the day 6 mean above the day 1 mean. Reproduced when both hold, partially when one does. Readings of the Cre mice as in fig1f.
- figs2f-novelty-slowing-familiar
When the novel arm block began, control mice slowed down more than Cre mice on familiar arm trials, most clearly on day 1.
Control animals slow down more than Cre animals in familiar arm trials after the novel arm block is removed
Figure S2F; Results, Stx3-dependent postsynaptic membrane fusion in CA1 is critical for novel environment preference and reward anticipation, paragraph 4How it will be judged
Normalized speed is log10 of the mean running speed on the block 6 trials divided by the mean running speed in the 10 trials preceding block 6 (Methods, VR behavioral analyses). The Methods use the whole block; the legend plots "the first two familiar arm trials of block 6". Readings, run both, (a) the first two familiar trials (legend), (b) all familiar trials of block 6 (Methods). The Methods do not say how "average running speed" is taken; readings, run both, (i) the mean of the position binned speeds, (ii) distance over time (time-averaged speed). Days 1 to 5; mixed effects ANOVA; posthoc t-test per day with Holm correction over the five days. Reproduced when the virus main effect has p < 0.05 with control lower (more slowing) and the day 1 posthoc has p < 0.05.
- figs2g-novel-arm-slowing-no-difference
Control and Cre mice slowed similarly on novel arm trajectories.
Control and Cre animals slow similarly to novel arm trajectories
Figure S2GHow it will be judged
A claim of absence. As figs2f-novelty-slowing-familiar ("Same as (F)") on novel arm trials, with the first two novel trials of block 6 as the measure and all novel trials of block 6 as a variant. Reproduced when the virus main effect has p >= 0.05.
- figs2i-reward-approach-speed-cv
Cre mice have more variable running speed during reward approach (higher coefficient of variation) than control mice, on familiar and novel trials.
Cre animals have more variable running speed during reward approach indicated by a higher coefficient of variation
Figures S2H-I; Results, Stx3-dependent postsynaptic membrane fusion in CA1 is critical for novel environment preference and reward anticipation, paragraph 4How it will be judged
CV per 10 cm bin is the standard deviation over trials of running speed divided by its mean over trials, for the trials of a type on a day; the value per mouse and day is the mean CV over the bins in the 30 cm preceding the reward zone (Methods, VR behavioral analyses). Days 1 to 6, mixed effects ANOVA per trial type. Readings of the Cre mice as in fig1f. Reproduced when the virus main effect has p < 0.05 with Cre higher for both trial types; partially when for one.
- fig2cd-decoding-no-group-difference
Controlling for the number of cells, position decoding error did not differ between control and Cre mice on familiar or novel trials.
When controlling for the number of cells used in the decoders, we did not find significant differences in decoder error between models trained on data from control or Cre animals (Fig. 2C-D)
Figures 2C-D; Results, CA1 Stx3 deletion does not affect gross measures of place code quality and stability, paragraph 1How it will be judged
A claim of absence. Naive Bayes decoder with uniform prior (maximum likelihood), rate maps from the trial-averaged activity rate of each cell in each bin on the training trials, Poisson log likelihood with continuous rates (the quasi-Poisson form of the Methods, whose equation the text omits), leave-one-trial-out cross-validation; error is the mean absolute distance between the most likely bin and the true bin, in cm; per mouse, 50 models each with 128 cells drawn with replacement, averaged (Methods, Naive Bayes decoding; Fig 2C legend). Per trial type, days 1 to 6, mixed effects ANOVA. Reproduced when the virus main effect has p >= 0.05 for both trial types.
- fig2cd-decoding-improves-over-days
Decoding error fell as the mice gained experience over days, on familiar and novel trials, with day 1 worse than every later day.
the decoders’ performance improved as the animals gained experience in the environments over days
Figures 2C-D; Results, CA1 Stx3 deletion does not affect gross measures of place code quality and stability, paragraph 1How it will be judged
Decoding error as in fig2cd-decoding-no-group-difference. Day main effect of the same ANOVA; posthoc paired t-tests over all 15 pairs of days, both groups pooled, Holm corrected (Fig 2C-D legends). Reproduced when for both trial types the day effect has p < 0.05 and day 1 error is significantly higher than each of days 2 to 6 (as the paper reports for both); partially when the day effect holds but not all five day 1 contrasts.
- figs3a-spatial-information-fraction-no-difference
The fraction of cells with significant spatial information did not differ between control and Cre mice, on familiar or novel trials.
There was also no difference between groups in the fraction of cells with significant spatial information or the stability of spatial representations within a day
Figure S3A; Results, CA1 Stx3 deletion does not affect gross measures of place code quality and stability, paragraph 1How it will be judged
A claim of absence. Fraction of the session's mCherry+ cells that are place cells (header, Place cells), per trial type and day, days 1 to 6; mixed effects ANOVA. Reproduced when the virus main effect has p >= 0.05 for both trial types.
- figs3b-within-day-stability-no-difference
Within-day stability of spatial representations, the mean trial by trial population vector correlation, was similar between control and Cre mice.
Average trial x trial population vector correlation for familiar (Left) and novel (Right) trials on each day shows similar within-day stability between conditions
Figure S3BHow it will be judged
A claim of absence. For each pair of trials of a type on a day, the correlation of their population vectors; the mean over pairs per mouse and day; days 1 to 6; mixed effects ANOVA. The paper does not define the population vector correlation. Readings, (a) Pearson correlation of the flattened cells by bins matrices, (b) the mean over bins of the Pearson correlation of the cell vectors at each bin; cells, (i) all cells, (ii) the place cells of that trial type and day. Reproduced when the virus main effect has p >= 0.05 for both trial types under every reading; when readings disagree the claim is methods_underspecified.
- figs3c-familiar-novel-remapping
Place codes were largely uncorrelated between the familiar and novel arms (global remapping) in both groups, and the familiar to novel population vector correlation did not differ between control and Cre mice.
Place cell codes were largely uncorrelated between the novel and familiar arms of the Y-maze in both groups (i.e. “global remapping”), indicating that contextual discrimination is intact after Stx3 deletion
Figure S3C; Results, CA1 Stx3 deletion does not affect gross measures of place code quality and stability, paragraph 1How it will be judged
A claim of absence for the group comparison. Population vector correlation between the average familiar trial activity and the average novel trial activity on each day (Fig S3C legend), days 1 to 6, mixed effects ANOVA, with the readings of the correlation and cells of figs3b. Reproduced when the virus main effect has p >= 0.05. The "largely uncorrelated" part has no test or number in the paper; the notes give each group's mean familiar to novel correlation beside its within-type correlation from figs3b. The text cites Fig 3E-F for remapping, which is the place field analysis; Fig 2E-F and S3C show it.
- fig2gh-across-day-pv-correlation-no-difference
The across-day population vector correlation of place cells tracked over all six days did not differ between control and Cre mice, for familiar or novel trials.
the population vector correlation of average place cell responses across days did not significantly differ between groups (Fig. 2G-H)
Figures 2G-H; Results, paragraph after Figure 3 legend ("In order to compare the long-term stability")How it will be judged
A claim of absence. Cells tracked across all six days; each cell's trial-averaged rate map per day; the day by day PV correlation matrix per mouse and its mean over the off-diagonal entries (the "average across day PV correlation"); unpaired t-test control against Cre. Readings, cells (a) all tracked cells (Fig 2G legend, "each cell"), (b) tracked place cells (Fig 2E-F, Results "place cell responses"), with the PV correlation readings of figs3b. Reproduced when p >= 0.05 for both trial types under every reading.
- figs3eg-across-day-decoding-no-difference
Position decoders trained on one day and tested on other days did not differ in accuracy between control and Cre mice.
Spatial decoders trained across days also did not significantly differ in their accuracy between control and Cre animals (Fig. S3D-G)
Figures S3D-G; Results, paragraph after Figure 3 legend ("In order to compare the long-term stability")How it will be judged
A claim of absence. Decoder as in fig2cd, using all cells tracked over the six days (Methods, Naive Bayes decoding), fit on the trials of one day and tested on the trials of each other day; the per mouse value is the average of the day by day error matrix (Fig S3E legend). Readings, (a) off-diagonal entries only, (b) with the diagonal filled by leave-one-trial-out within-day error. Unpaired t-test; reproduced when p >= 0.05 for both trial types under both readings.
- figs4-morph-task-context-discrimination
In a separate morph VR task, Cre mice showed normal experience-dependent contextual remapping, with trial by trial similarity matrices like those of control mice.
As a stronger test of intact contextual coding, Cre mice showed normal experience dependent contextual remapping patterns in a separate VR task71 (Fig. S4)
Figure S4D-E; Results, CA1 Stx3 deletion does not affect gross measures of place code quality and stability, paragraph 1How it will be judged
The paper shows example similarity matrices only, for session 8 of example mice, and gives no measure, test or count of mice; the task's methods are cited to Plitt & Giocomo 2021 (ref 71) and not given here. Not specified by the paper; recorded so the finding is covered.
- figs5-ca3-dg-place-fields
Place fields were present in CA2/3 and dentate gyrus imaged in Cre mice, on familiar and novel trials.
place fields were intact in the CA3 region and the dentate gyrus in a small cohort of Cre mice in which we additionally injected the same GCaMP virus in these regions
Figure S5; Results, Spatial coding in Cre animals may be passively inherited from upstream activity, paragraph 1How it will be judged
The paper shows example cells only (Cre-8 CA2/3 on day 1; Cre-7 DG in a session repeating day 6) and gives no count or test. The test the Methods allow is the place cell criterion of the header on those sessions; its 95% shuffle threshold admits about 5% of cells by chance. Reproduced when, on familiar and on novel trials of each session, the fraction of significant cells exceeds 5% with a binomial test p < 0.05. The 5% chance level follows from the paper's threshold; the paper states no count to match.
- fig3bc-model-stable-place-cells
In the CA3-CA1 K-Winners-Take-All model, both the Plasticity and the No-Plasticity versions produce place-selective CA1 units with a spatially stable code from the first trial.
In both models, individual units were clearly place-selective, and the population maintained spatially stable place coding from the first trial (Fig. 3B-C)
Figures 3B-C; Results, Spatial coding in Cre animals may be passively inherited from upstream activity, paragraph 1How it will be judged
A simulation, not a test on recorded data. The Methods give M=N=1000, input variance 0.5 on x in [0,10], K=50, output noise 0.5N(0,1), stimulus noise 0.05N(0,1), eta 1e-4 and alpha 9e-4 (Plasticity) or 0 and 0 (No-Plasticity), weight noise 0.001N(0,1), wmax 10, lognormal initial weights (mean 0, sd 0.5). The text read omits the weight update equation and the input tuning function, and gives no number of trials, no position sampling and no measure of stability or place selectivity. Not specified by the paper; recorded so the finding is covered.
- fig3d-model-wider-fields-with-plasticity
Place fields are wider in the Plasticity model than in the No-Plasticity model.
place fields were wider in the Plasticity model than the No-Plasticity model (Fig. 3D)
Figure 3D; Results, Spatial coding in Cre animals may be passively inherited from upstream activity, paragraph 2How it will be judged
A simulation, as fig3bc-model-stable-place-cells. The paper does not say how a model place field or its width is defined, nor the trial on which it is measured. Not specified by the paper; recorded so the finding is covered.
- fig3e-model-more-fields-without-plasticity
Units in the No-Plasticity model have more place fields per cell than units in the Plasticity model.
place cells from the No-Plasticity model were more likely to have multiple place fields per cell (Fig. 3E)
Figure 3E; Results, Spatial coding in Cre animals may be passively inherited from upstream activity, paragraph 2How it will be judged
A simulation, as fig3bc-model-stable-place-cells; the definition of a model place field is not given. Not specified by the paper; recorded so the finding is covered.
- fig3f-narrower-fields-cre-familiar
On familiar trials, place fields of Cre mice are narrower than those of control mice.
Place fields from Cre mice were narrower than those from control mice (Fig. 3F-G)
Figure 3F; Results, Spatial coding in Cre animals may be passively inherited from upstream activity, paragraph 2How it will be judged
Fields from the single place field method of the header on familiar trials of the day; width is the number of contiguous significant bins times 10 cm; per mouse and day, the mean over fields ("across cell average", Fig 3F legend; readings, mean over fields, or mean over cells of each cell's mean width). Which fields enter is open; readings, (a) all candidate fields of at least 20 cm, (b) only fields with a formation lap, as for the other single-field analyses. Days 1 to 5 mixed effects ANOVA; day 6 unpaired t-test. Reproduced when the virus main effect has p < 0.05 with Cre narrower and the day 6 test has p < 0.05; partially when only one holds.
- fig3g-narrower-fields-cre-novel
On novel trials, place fields of Cre mice are narrower than those of control mice.
Place fields from Cre mice were narrower than those from control mice (Fig. 3F-G)
Figure 3G; Results, Spatial coding in Cre animals may be passively inherited from upstream activity, paragraph 2How it will be judged
As fig3f-narrower-fields-cre-familiar, on novel trials.
- fig3h-more-fields-per-cell-cre-familiar
On familiar trials, place cells of Cre mice have more place fields per cell than those of control mice.
place cells from Cre mice had more place fields on average than place cells from control mice (Fig. 3H-I)
Figure 3H; Results, Spatial coding in Cre animals may be passively inherited from upstream activity, paragraph 2How it will be judged
Number of fields per cell from the single place field method, with the field readings of fig3f; cells counted are, readings, (a) cells with at least one field, (b) place cells (SI significant) for that trial type and day. Per mouse and day the mean over cells; days 1 to 5 mixed effects ANOVA; day 6 unpaired t-test. Reproduced when the virus main effect has p < 0.05 with Cre higher and the day 6 test has p < 0.05; partially when only one holds.
- fig3i-more-fields-per-cell-cre-novel
On novel trials, place cells of Cre mice have more place fields per cell than those of control mice.
place cells from Cre mice had more place fields on average than place cells from control mice (Fig. 3H-I)
Figure 3I; Results, Spatial coding in Cre animals may be passively inherited from upstream activity, paragraph 2How it will be judged
As fig3h-more-fields-per-cell-cre-familiar, on novel trials.
- fig4e-block6-familiar-activity-lower-cre
During block 6, population activity on familiar trials, relative to the preceding baseline, is lower in Cre mice than in control mice.
control animals showed an overall larger increase in population activity rate than Cre animals in both trial types
Figures 4C and 4E; Results, CA1 Stx3 is necessary for increased neural activity in novel environments, paragraph 1How it will be judged
Each cell's position-binned activity rate on each block 6 trial is divided by that cell's mean activity rate over the 10 trials preceding block 6; the population value is log10 of the mean over all cells (Methods, Novel arm population activity rate); the per mouse and day value is its mean over bins and over the block 6 familiar trials. The Fig 4A legend instead takes the log per cell before averaging; that is undefined for bins with zero activity, which the paper does not treat, so it is shown as a variant where defined and does not decide. Days 1 to 5, mixed effects ANOVA; posthoc unpaired t-tests per day, Holm corrected. Reproduced when the virus main effect has p < 0.05 with Cre lower.
- fig4e-block6-novel-activity-lower-cre
During block 6, population activity on novel trials, relative to the preceding baseline, is lower in Cre mice than in control mice.
control animals showed an overall larger increase in population activity rate than Cre animals in both trial types
Figures 4C and 4E; Results, CA1 Stx3 is necessary for increased neural activity in novel environments, paragraph 1How it will be judged
As fig4e-block6-familiar-activity-lower-cre, on the block 6 novel trials. Reproduced when the virus main effect has p < 0.05 with Cre lower.
- fig4e-novel-minus-familiar-larger-control
The difference between novel and familiar trial population activity in block 6 is larger in control mice than in Cre mice.
The difference in the population activity rate between novel arm and familiar arm trials was also greater in control animals compared to Cre animals
Figure 4E; Results, CA1 Stx3 is necessary for increased neural activity in novel environments, paragraph 1How it will be judged
Per mouse and day, the novel trial value minus the familiar trial value of fig4e-block6-novel-activity-lower-cre and fig4e-block6-familiar-activity-lower-cre; days 1 to 5, mixed effects ANOVA. Reproduced when the virus main effect has p < 0.05 with control larger.
- fig4-block6-activity-increase-pattern
In block 6 population activity increased above baseline on familiar and novel trials in both groups, more on the arms than on the stem, and more on novel than on familiar trials in both groups.
During this block of trials, we observed an increase in population activity for both familiar and novel arm trials. This increase in activity rate was larger on the arms of the Y-Maze than the stem and greater for novel compared to familiar arm trials in both control and Cre animals
Results, CA1 Stx3 is necessary for increased neural activity in novel environments, paragraph 1; Figure 4CHow it will be judged
The paper reports no test. Measure of fig4e-block6-familiar-activity-lower-cre, averaged per mouse over days 1 to 5; group values are across-animal means. Holds when, in each group, (i) the familiar and the novel values are above 0, (ii) the value over arm bins exceeds the value over stem bins for each trial type, (iii) novel exceeds familiar. Where the stem ends is open; readings, (a) the first 150 cm (stem and arms are of equal length, Methods), (b) the first 100 cm (trajectories diverge two thirds of the way down the stem). Reproduced when all three hold in both groups; partially when some do.
- fig4e-novel-increase-grows-over-days
The increase in activity on novel trials did not dissipate over days and grew stronger over the five days.
The increase in activity rate for novel trials did not dissipate over days, and in fact grew stronger over the five days of the experiments (Fig. 4C & E)
Figures 4C and 4E; Results, CA1 Stx3 is necessary for increased neural activity in novel environments, paragraph 2How it will be judged
Novel trial values of fig4e-block6-novel-activity-lower-cre. Reproduced when the day main effect has p < 0.05 and the across-animal mean on day 5 exceeds that on day 1 (both groups pooled, as the day effect is).
- fig4d-activity-increase-spread-across-cells
On day 5 the larger activity increase of control mice is spread over a large part of the population, roughly half of the cells on novel trials, and is not due to a small population of highly active cells.
we observed that roughly half of the cells displayed increased activity rates in control animals as compared to Cre animals. Thus, the effect is driven by a large proportion of the population and is not due to a small population of the highly active neurons (Fig 4D)
Figure 4D; Results, CA1 Stx3 is necessary for increased neural activity in novel environments, paragraph 2How it will be judged
The paper gives no test or measure for this panel (the legend refers to the tests of Fig 4C). Per cell, the block 6 novel trial activity of day 5, each position-binned rate divided by the cell's mean over the 10 trials preceding block 6 and averaged over bins and trials (Methods, Novel arm population activity rate; Fig 4D legend, "trial averaged activity rates across cells"); per mouse, the cumulative distribution over cells (a log transform does not change it). "Roughly half of the cells" is not defined by the paper; the reading used is that the across-animal mean control quantile exceeds the Cre quantile at every decile from the median (0.5) to 0.9, so that the shift is not confined to the top tail. Familiar trials are shown as a variant. Not a test; the notes give both groups' quantiles.
- figs6b-block4-activity-no-difference
In the fourth (familiar) block, normalized population activity did not differ between control and Cre mice.
These results were not driven by a general increase in activity rate over the course of imaging or increases in running speed during the last block of trials (Fig. S6 & S2F-G)
Figure S6B; Results, CA1 Stx3 is necessary for increased neural activity in novel environments, paragraph 1How it will be judged
A claim of absence. Measure of fig4e-block6-familiar-activity-lower-cre applied to block 4, each cell normalized by its mean over the 10 trials preceding block 4 (Methods, "the 10 trials preceding the block of trials to be analyzed"). Days 1 to 5, mixed effects ANOVA. Reproduced when the virus main effect has p >= 0.05.
- figs6e-familiar-activity-trials-since-novel
Familiar trial activity in block 6 depends on the number of trials since the last novel trial, in control and in Cre mice.
The magnitude of the increase in activity rate during familiar trials also depended on the time since encountering a novel trial (Fig. S6)
Figure S6E; Results, CA1 Stx3 is necessary for increased neural activity in novel environments, paragraph 1How it will be judged
Per familiar trial of block 6, the population value of fig4e-block6-familiar-activity-lower-cre, grouped by the number of trials since the last novel trial; per mouse, day and group, the mean; a two-way repeated measures ANOVA (day and trials since novel, within) per virus group. The legend names 4 conditions and 5 days but prints F(2,.) and F(5,.), which fit 3 levels and 6 days; the Cre df fit 4 mice with complete data. Readings, levels (a) 1, 2, 3 trials since, (b) 1, 2, 3 or more; days (i) 1 to 5, (ii) 1 to 6. Reproduced when in each group the trials since novel effect has p < 0.05 under a reading. The paper states no direction for the effect; the notes give it.
- figs7b-reward-overrepresentation
Control mice allocate a greater proportion of place cells to the locations just before the reward zone than Cre mice, on familiar and novel trials.
we find that control animals allocate a greater proportion of place cells to rewarded locations than Cre animals (Fig. 5A-D & S7)
Figures S7A-B; Results, CA1 Stx3 is necessary for place cell reward representations and memory, paragraph 1How it will be judged
Among the place cells of a trial type and day, the proportion whose peak of the trial-averaged rate map lies in the 50 cm preceding the front of that arm's reward zone (Methods, Reward overrepresentation analysis); logit transform; days 1 to 6, mixed effects ANOVA per trial type. The peak is taken from all trials of the type (the Methods' split halves are for the sequence plots). Proportions of 0 or 1, which logit cannot take, are clipped to half a cell (a standard choice, said if used). Reproduced when the virus main effect has p < 0.05 with control higher for both trial types; partially for one.
- fig5d-reward-cells-fewer-in-cre
Control mice have a larger proportion of reward cells, active before the reward zone on both arms, than Cre mice, and the proportion changes over days.
Proportion of the population defined as “reward cells” shown for each mouse on each day. Statistical tests performed on logit transformed proportions
Figure 5D; Results, CA1 Stx3 is necessary for place cell reward representations and memory, paragraph 1How it will be judged
A reward cell has significant spatial information on both familiar and novel trials and its peak lies near the front of the reward zone on both (Methods, Reward overrepresentation analysis). Readings, peak window (a) the 50 cm preceding the front of the zone (as in the same section's place cell rule), (b) within 50 cm of the front on either side ("within 50 cm of the front"); denominator, (i) all cells of the session ("proportion of the population"), (ii) cells significant on both trial types. Logit transform; days 1 to 6, mixed effects ANOVA. Reproduced when the virus main effect has p < 0.05 with control higher.
- fig5d-reward-cells-increase-over-days-control
The proportion of reward cells increased over days in control mice.
The proportion of reward cells also increased over days for the control animals indicating accumulating plasticity in the place code (Fig. 5D)
Figure 5D; Results, CA1 Stx3 is necessary for place cell reward representations and memory, paragraph 1How it will be judged
Proportions of fig5d-reward-cells-fewer-in-cre. The paper's test is the day main effect of that ANOVA, over both groups. Reproduced when it has p < 0.05 and the control across-animal mean on day 6 exceeds that on day 1 under a reading.
- fig5d-reward-cells-above-chance-control-only
Averaged across days, the fraction of reward cells exceeds the level expected from random remapping in control mice but not in Cre mice.
Across-day average fraction of reward cells is greater than chance in control but not Cre animals
Figure 5D legend; Results, CA1 Stx3 is necessary for place cell reward representations and memory, paragraph 1How it will be judged
Per mouse, the mean over days 1 to 6 of the reward cell fraction of fig5d-reward-cells-fewer-in-cre, against chance by a one sample t-test. Chance is "the proportion of reward coding neurons expected by random remapping between novel and familiar arms with constant coding for the stem of the maze" (Fig 5D legend), computed in a way the paper does not give. Readings, (a) per mouse and day, the reward cell fraction after randomly permuting across cells the novel trial peak locations of cells whose peaks lie on the arms, stem peaks kept (1,000 permutations, mean); (b) analytically, with arm peaks uniform over arm bins and independent between trial types, stem peaks equal. Reproduced when the control test has p < 0.05 above chance and the Cre test p >= 0.05 under a reading; partially when control holds and Cre is also above chance. The Discussion says Cre mice have more reward coding neurons than chance, which the legend's p=.080 does not support.
- fig5h-extinction-group-difference
After the reward locations were switched, control mice kept licking at the old reward zone longer than Cre mice; a binomial mixed model of licking fits better than with shuffled group labels.
The true model fit exceeded those from a shuffled distribution in which animal identities were scrambled, indicating that the Cre and control animals displayed significantly different licking behavior (Fig. 5H-I)
Figures 5E-I; Results, CA1 Stx3 is necessary for place cell reward representations and memory, paragraph 2How it will be judged
Left arm trials after the reversal (day 7 from block 3, and day 8). Outcome per trial is whether the mouse licked more than once in the previous reward anticipatory zone. The paper names that zone two ways, "the previous reward anticipatory zone" (Fig 5H-I legends) and "the old reward zone" (Results). Readings, zone (A) the 30 cm preceding the old left reward zone (the anticipatory window of Methods, VR behavioral analyses), (B) the old 25 cm left reward zone itself. Binomial generalized linear mixed model in statsmodels (BinomialBayesMixedGLM, statsmodels' binomial mixed model) with trial number, virus and their interaction as fixed effects (the paper compares intercepts and slopes between groups) and a random intercept per mouse. Model fit is the mean log loss of the fitted probabilities; the null refits with virus labels permuted across mice (1,000 permutations, a standard choice); p is the fraction of shuffles with log loss at or below the true one. Open and run as readings, trial number (a) counted on from day 7 into day 8, (b) day 7 only; random effects (i) intercept, (ii) intercept and slope. Reproduced when p < 0.05 with control predicted to lick more.
- fig5i-extinction-intercept-not-slope
The groups differ in the intercept of the extinction model but not its slope; control mice lick at the old zone for longer but then extinguish at a similar rate.
we observed a group difference in the intercepts of the binomial regression (shuffled p=.007) but not the slope (shuffled p=.751)
Figure 5I; Results, CA1 Stx3 is necessary for place cell reward representations and memory, paragraph 2How it will be judged
Model and shuffles of fig5h-extinction-group-difference; the shuffle p for each parameter is the fraction of shuffles whose virus intercept term (or trial by virus slope term) is at least as large in absolute value as the true one. Reproduced when the intercept p < 0.05 and the slope p >= 0.05 under a reading.
- fig6c-pv-backward-shift-day1-control
Control mice show a population-level backward shift of place fields over the first novel arm trials of day 1, larger than in Cre mice; it is not detectable on later days.
control animals displayed widespread population-level backward shifting of place fields in these initial novel arm trials60,83,85 (Fig. 6A-C & S8A-C). This effect was not detectable on subsequent days
Figure 6C; Results, CA1 Stx3 contributes to experience dependent place field shifts, paragraph 2How it will be judged
Initial PV shift per mouse and day (header, Population vector shift) over the first five novel trials, days 1 to 5. Open and run as readings, cells (a) all cells, (b) novel trial place cells of that day; COM over lags (i) of the raw cross-correlation, (ii) with negative values set to zero. Mixed effects ANOVA; posthoc unpaired t-tests per day and one sample t-tests against 0 per group and day, Holm corrected. Reproduced when the virus main effect has p < 0.05, the day 1 control against Cre test has p < 0.05 with control more negative, the control day 1 shift is below 0 with p < 0.05 and control days 2 to 5 have p >= 0.05; partially when day 1 holds but a later day is also significant.
- fig6c-no-pv-shift-cre
In Cre mice there is no significant population-wide backward shift within a day.
at a population level, there was no evidence of significant population wide backward shifting within a day or across days (Fig. 6B-C & S8)
Figures 6C and S8B; Results, CA1 Stx3 contributes to experience dependent place field shifts, paragraph 2How it will be judged
A claim of absence, on the paper's test, the Holm-corrected one sample t-tests of the Cre initial PV shift against 0 on each of days 1 to 5 (Fig 6C legend), with the readings of fig6c-pv-backward-shift-day1-control. Reproduced when all have p >= 0.05.
- figs8h-across-day-pv-shift
Across days, novel arm place codes shift backward gradually in control mice but not in Cre mice.
a slower backward shift of the population place code was also detectable across days during novel arm trials from control animals (Fig. S8G-H)
Figures S8G-H; Results, CA1 Stx3 contributes to experience dependent place field shifts, paragraph 2How it will be judged
The PV shift method applied to the trial-averaged novel arm rate maps of each day in place of single trials (Fig S8G legend); per mouse the mean of the upper triangle of the day by day matrix. Cells are those tracked across days. Readings, days (a) 1 to 6, (b) 1 to 5; with the COM readings of fig6c. Reproduced when the control shift is below 0 with p < 0.05, the Cre shift has p >= 0.05, and control is more negative than Cre with p < 0.05.
- figs8d-pv-shift-not-explained-by-speed
The day 1 PV shift is not correlated with running speed on the first five novel trials.
There is no significant correlation between PV shift and running speed in the first 5 novel trials of day 1
Figure S8D; Results, CA1 Stx3 contributes to experience dependent place field shifts, paragraph 2How it will be judged
A claim of absence. Per mouse, the day 1 initial PV shift of fig6c and the mean running speed over the first five novel trials of day 1; Pearson correlation over the 16 mice. Reproduced when p >= 0.05.
- fig6d-field-com-shift-larger-control
Single place fields shift backward abruptly after their formation lap on novel trials, more in control than in Cre mice, after controlling for running speed on the formation lap; the difference holds on days 1, 2 and 3.
In agreement with the population vector analysis, the magnitude of the initial place field shift was greater in control animals than Cre animals. This difference remained significant after controlling for running speed through the place field on the formation lap
Figures 6D and S8E-F; Results, CA1 Stx3 contributes to experience dependent place field shifts, paragraph 3How it will be judged
Novel trial fields with a formation lap (header, Single place fields). COM of the cell's activity within the field on each lap from unbinned position and activity (Methods), relative to the field's mean COM over laps from the formation lap on. Formation lap speed is the mean running speed through the field on that lap. statsmodels MixedLM, COM shift ~ speed + virus (Fig S8F legend), random intercept per mouse; Wald test of the virus coefficient. Readings, "trial 1" is (a) the formation lap, (b) the lap after it. Reproduced when on day 1 the virus effect has p < 0.05 with control more negative (backward); partially when day 1 holds but not days 2 and 3.
- fig6e-width-speed-correlation
Place field width correlates with running speed through the field on the formation lap in both groups, and neither the intercept nor the slope differs between control and Cre.
The place field width is correlated with running speed through the place field on the formation lap in both groups ... Neither the intercept nor the slope of the regression differs between Cre and control (Linear mixed effects regression, Wald test p>.05)
Figure 6E; Results, CA1 Stx3 contributes to experience dependent place field shifts, paragraph 3How it will be judged
Fields and formation lap speed as in fig6d; width as in fig3f. The Fig 6E legend does not name the day or trial type. Readings, (a) day 1 novel trials (the data of Fig 6D and S8F), (b) novel trials of days 1 to 6 pooled. statsmodels MixedLM, width ~ speed * virus, random intercept per mouse. The correlation part holds when the speed slope is positive with Wald p < 0.05 within each group (the paper gives no p for it); the absence part, on the paper's Wald tests, holds when the virus and speed by virus terms have p >= 0.05. Reproduced when both parts hold, partially when one does.
- figs9a-hse-trial-proportion-no-difference
The proportion of trials containing a highly synchronous event does not differ between control and Cre mice.
Proportion of trials that contain a significant HSE does not differ between control animals (black) and Cre animals (red)
Figure S9A; Results, CA1 Stx3 is necessary for offline consolidation of place codes, paragraph 2How it will be judged
A claim of absence. HSEs as in the header; per mouse and day, the fraction of trials with at least one HSE in a post-reward stopping period; days 1 to 6, mixed effects ANOVA. Reproduced when the virus main effect has p >= 0.05.
- figs9b-hse-size-no-difference
The proportion of the population participating in HSEs does not differ between control and Cre mice.
Proportion of the neural population that participates in HSEs does not differ between control (black) and Cre (red) animals
Figure S9B; Results, CA1 Stx3 is necessary for offline consolidation of place codes, paragraph 2How it will be judged
A claim of absence. Per HSE, the number of participating cells over the session's cell count; per mouse and day the mean over HSEs; days 1 to 6, mixed effects ANOVA. Reproduced when the virus main effect has p >= 0.05.
- fig7e-reactivated-cells-more-stable-across-days-control
In control mice, cells that participated in at least one HSE have higher across-day spatial rate map correlation than cells that did not, on familiar and novel trials.
in control animals, cells that participated in at least one HSE had higher across day spatial rate map correlation and higher trial x trial rate map correlation than cells that did not participate in HSEs (Fig. 7C-H)
Figure 7E; Results, CA1 Stx3 is necessary for offline consolidation of place codes, paragraph 2How it will be judged
Cells tracked across days; HSE participation as in the header. Per mouse and trial type, the mean across-day rate map correlation of reactivated cells minus that of non-reactivated cells; one sample t-tests against 0 for each group and trial type, Holm corrected over the four tests (Fig 7E legend). The paper does not define the across-day correlation or the timing of participation. Readings, correlation (a) mean Pearson correlation of the cell's trial-averaged rate maps over all pairs of days, (b) over consecutive days; reactivated (i) in an HSE on any day, (ii) on the earlier day of the pair. Reproduced when both control tests have p < 0.05 with positive differences.
- fig7e-cre-reactivation-boost-novel-only
In Cre mice, reactivated cells have a weakly higher across-day correlation on novel trials but not on familiar trials.
Cre animals did not show this boost in spatial coding reliability for cells that participated in HSEs except for a weak increase in across day spatial rate map correlation for novel trials (Fig. 7D-E)
Figure 7E; Results, CA1 Stx3 is necessary for offline consolidation of place codes, paragraph 2How it will be judged
Measure, tests and readings of fig7e-reactivated-cells-more-stable-across-days-control, for Cre mice. Reproduced when the novel test has p < 0.05 with a positive difference and the familiar test has p >= 0.05 (the absence part judged on the paper's test).
- fig7e-reactivation-effect-virus-and-novelty
The across-day stability gain of reactivated cells is larger in control than in Cre mice and larger on novel than familiar trials, and the novelty effect is larger in control mice.
The increase in across day correlation for reactivated cells is greater for control animals than Cre animals. This increase in stability of the neural representation is larger for novel trials than familiar trials
Figure 7E; Results, CA1 Stx3 is necessary for offline consolidation of place codes, paragraph 2How it will be judged
Per mouse differences of fig7e-reactivated-cells-more-stable-across-days-control; mixed effects ANOVA with virus between and trial type within; unpaired t-test of novel minus familiar between groups. Reproduced when the virus and trial type main effects have p < 0.05 (control higher, novel higher) and the novelty effect is larger in control with p < 0.05; partially when the main effects hold but not the novelty comparison.
- fig7h-reactivated-cells-higher-trial-correlation-control
In control mice, cells that participated in HSEs have higher trial by trial rate map correlation than cells that did not, on familiar and novel trials.
Reactivated cells in control animals have higher trial x trial correlation than non-reactivated cells for both familiar
Figure 7H; Results, CA1 Stx3 is necessary for offline consolidation of place codes, paragraph 2How it will be judged
Per cell, per day and trial type, the mean Pearson correlation between its single trial rate maps over all pairs of trials; per mouse, the mean over cells and days of reactivated minus non-reactivated cells. Readings, reactivated (i) on that day, (ii) on any day; cells (a) all cells, (b) tracked cells. One sample t-tests against 0, Holm corrected over the four tests. Reproduced when both control tests have p < 0.05 with positive differences.
- fig7h-no-trial-correlation-boost-cre
In Cre mice, reactivated cells do not have higher trial by trial correlation than non-reactivated cells, on either trial type.
In Cre animals, reactivated cells do not have higher trial x trial correlation than non-reactivated cells for either trial type (Holm-corrected p>.05)
Figure 7HHow it will be judged
A claim of absence, on the paper's tests (one sample t-tests, Holm corrected) with the measure and readings of fig7h-reactivated-cells-higher-trial-correlation-control. Reproduced when both Cre tests have p >= 0.05.
- fig7h-reactivation-effect-virus-and-novelty
The trial by trial correlation gain of reactivated cells is larger in control than in Cre mice and larger on novel than familiar trials.
The increase in trial-by-trial correlation for reactivated cells is greater for control animals than Cre animals. This increase in stability of the neural representation is larger for novel trials than familiar trials
Figure 7HHow it will be judged
Per mouse differences of fig7h-reactivated-cells-higher-trial-correlation-control; mixed effects ANOVA with virus between and trial type within. Reproduced when both main effects have p < 0.05 with control higher and novel higher.
Pipeline
- Paper
- done (1 of 1)
- Claims
- done
- Claims check
- done (ready)
- Attempt
- todo
- Questions
- none
- Verified
- none