Claims 46
From the paper Padilla-Coreano, N., Batra, K., Patarino, M. et al. (2022). Cortical ensembles orchestrate social competition through hypothalamic outputs. Nature 603, 667-671. · claims file · checked 2026-10-21: ready
- fig1b-dominants-more-rewards
In the unimplanted cohort, relative dominants collected more rewards than relative subordinates when competing in dyads.
Mice with higher relative ranks (dominant; Dom ) collected more rewards than relative subordinates ( Sub ) when competing in dyads
Figure 1b; Results, first paragraphHow it will be judged
Per dyad, total rewards won by the relative dominant and the subordinate over the session, as in the legend; Wilcoxon signed-rank test across the 12 dyads. Reproduced when p < 0.05 with dominants higher. Percent competitions won by absolute rank (6 per rank, Figure 1b right) is reported and not judged, as the text states no result for it.
- fig1c-port-pushing-displacement
In the unimplanted cohort, relative dominants had higher port occupation, higher pushing success, and higher time displaced from the port (as the legend words it) than relative subordinates.
Port occupation, pushing success (pushing that resulted in displacement of competitor), and time displaced from port was higher for relative dominants
Figure 1cHow it will be judged
Per dyad, dominant against subordinate, Wilcoxon signed-rank tests on port occupation time, pushing success (percent of pushes that displaced the competitor, Methods 6, Reward competition behavioral analysis) and time displaced. The legend leaves open which way the displacement measure runs; readings (a) time spent displacing the competitor, as the main text says dominants "were more successful at displacing competitors", (b) time being displaced, as printed. Reproduced when all three tests give p < 0.05 with dominants higher under one reading.
- ed1d-pushing-success
In the unimplanted cohort, relative dominants have higher pushing success during the tone.
Relative dominant mice have higher pushing success during the tone (n=12 per group, paired t-test p=0.025)
Extended Data Fig. 1dHow it will be judged
Pushing success (percent of pushes that displaced the competitor, Methods 6) during the 10 s tone, per mouse, paired across the 12 dyads. Reproduced when the paired t-test gives p < 0.05 with dominants higher.
- ed1g-distance-by-outcome-not-rank
In the unimplanted cohort, the distance to the reward port around the tone differs by trial outcome (win against lose) but not by relative rank.
Distance to reward port across time by trial type and relative rank
Extended Data Fig. 1gHow it will be judged
Per trial, the mean AlphaTracker distance to the human-marked port location over the windows the legend gives (-5 s to tone onset; -5 s to +10 s); 2-way ANOVA of trial type by relative rank on trials. Reproduced when trial type has p < 0.05 and rank has p >= 0.05 in both windows; the legend gives no direction for trial type, which is reported. The rank result is a claim of absence judged on this test.
- fig1e-alphatracker-precision
AlphaTracker tracks two unmarked mice with higher precision than human annotators.
AlphaTracker performance surpasses human accuracy when tracking 2 or 4 mice
Figure 1e; Extended Data Fig. 2How it will be judged
Needs the annotated evaluation frames of Methods 15 (800 frames from 4 videos, 200 double-annotated), which the paper does not deposit. Reproduced when AlphaTracker's average precision over body parts exceeds the between-human precision on those frames.
- fig1i-hmmglm-decodes-behaviour
A 6-state HMM-GLM decodes the 9 behaviour labels from mPFC multi-unit activity above chance and better than a GLM or an ensemble of 9 GLMs, while the GLM and the 9 GLMs do not differ.
An HMM-GLM model with 6 hidden states decoded behavioral labels from neural activity with superior cross-validated performance to static models
Figure 1i; Extended Data Fig. 5j; Methods 17How it will be judged
Input, from Methods 17 (Dataset), multi-unit rates on the 32 channels after common average reference, spikes at 4 SD from the mean, in 70 bins of 200 ms per trial (aligned at -2 to +12 s from tone onset, from the dandiset description; the paper gives no alignment), and one of the 9 labels per bin. HMM-GLM as fixed by run_model.m and model_HMMGLM.m (see the opening comment). FT, GLM (multinomial logit, 9-output softmax), SVM (Gaussian kernel, one against all) and 9 GLMs (one logit GLM per label) as Methods 17 describe. Per behaviour, the AUC is the mean over the 10 folds. Tests as the legend gives them; the printed p-values match a two-sided sign test of 9 of 9 AUCs above 0.5 (0.0039) and an exact rank-sum with no overlap between 9 and 9 (4.1e-5). Reproduced when HMM-GLM AUCs exceed 0.5 by the sign test (p < 0.05), exceed the GLM and 9-GLM AUCs by rank-sum (p < 0.05), and the GLM against 9 GLMs rank-sum gives p >= 0.05 (a claim of absence, judged on that test). Open, with readings; spike threshold (a) negative crossings below mean minus 4 SD, (b) crossings in either direction; rates (a) counts per bin, (b) z-scored per channel; AUC (a) with the one-bin backward shift the Methods describe (auc_ROC_delay), (b) without (auc_ROC); folds (a) contiguous in session order, as the code forms them, (b) random. The processed input in the code repository is a variant input, not the basis of the outcome.
- ed5j-six-states
Held-out log-likelihood of the HMM-GLM increases with the number of hidden states and plateaus at 6 states.
We found that as you increase the number of hidden states, the log likelihood increases, reaching a plateau at the 6-state model
Extended Data Fig. 5j, left; Methods 17, Cross-validation and Model SelectionHow it will be judged
Per number of states, the trial-averaged held-out log-likelihood over 10 folds, with SEM across folds (Methods 17), other settings as in fig1i-hmmglm-decodes-behaviour. The number of states tried is not given; read as 1 to 10. "Plateau" is not defined; readings (a) 6 is the smallest number whose mean lies within one SEM of the best mean, (b) 6 has the highest mean. Reproduced when one reading selects 6, partially reproduced when it selects 5 or 7.
- ed5k-hmmglm-generalizes-across-rank
The 6-state HMM-GLM predicts behaviour above chance whichever relative rank group is used for training and for testing.
The model performed equally well when training for one relative rank and testing on the other
Extended Data Fig. 5k; Results, mPFC neurons encode competition behaviorHow it will be judged
The model of fig1i-hmmglm-decodes-behaviour trained on dom trials and tested on sub trials, and the reverse, plus within-group 10-fold runs. Per behaviour the AUC on the test trials. Reproduced when every train and test combination gives a sign test of the 9 AUCs against 0.5 at p < 0.05.
- ed5b-rank1-wins-most
In the recording cohort, rank 1 mice won the highest percentage of competitions.
Percent competitions won by absolute rank is highest for rank 1 mice in dataset used for mPFC recordings
Extended Data Fig. 5bHow it will be judged
A competition is won by the mouse that won more of its trials (the meaning ED Fig. 8d gives to a winner). Ties are not covered by the paper; readings (a) a tie counts as not won, (b) ties are left out. Reproduced when rank 1 has the highest percentage under one reading.
- ed5c-rewards-not-different
Among mice wearing loggers, relative dominants and subordinates obtained similar numbers of rewards, while the body weight difference between competitors correlated with rewards won.
number of rewards obtained by relative dominants ( dom ) and subordinates ( sub ) during the reward competitions between animals wearing loggers
Extended Data Fig. 5cHow it will be judged
The rewards part is a claim of absence judged on the paired t-test across the 22 pairs (reproduced when p >= 0.05). The correlation holds when Pearson p < 0.05, using only mice with a same-day weight.
- ed5d-subordinates-slower-latency
In win trials, relative subordinates had longer latencies to pick up the reward than relative dominants.
Subordinates had longer latencies to pick up the reward during win trials
Extended Data Fig. 5dHow it will be judged
Latency of the winner's port entry to pick up the reward in win trials. No section of the Methods defines this latency; Methods 5 and ED Fig. 1a measure reward latency from tone onset. Readings (a) from tone onset, (b) from delivery at 2 s, which differ by a constant and give the same tests. Reproduced when the rank-sum gives p < 0.05 with subordinates slower. The RM-ANOVA is recorded as printed; its F and p do not agree (F(1,24) = 2.06 gives p near 0.16) and its factors are not given, so it is not judged.
- ed5e-port-occupancy-not-different
Percent port occupancy during the tone did not differ by relative rank in the recording cohort.
Percent port occupancy during tone across relative rank (n=31 sessions, paired t-test p=.90)
Extended Data Fig. 5eHow it will be judged
A claim of absence, judged on a paired t-test of percent of the 10 s tone spent at the port, dom against sub per session. Reproduced when p >= 0.05; the difference is reported.
- ed5f-dominants-displace-more
Relative dominants were more successful at displacing subordinates from the port, over the whole competition and during the tone.
Relative dominants were more successful displacing subordinates from the reward port throughout the competition
Extended Data Fig. 5fHow it will be judged
Displacement success as Methods 6 define pushing success (percent of pushes that displaced the competitor), from the hand-scored labels, dom against sub per session, paired t-test. Reproduced when both give p < 0.05 with dominants higher.
- ed5g-time-displaced-not-different
Total time being displaced from the port did not differ by relative rank in the recording cohort.
Total time being displaced from reward port by relative rank in dataset used for mPFC recordings (n=31 sessions; paired t-test p=0.15)
Extended Data Fig. 5gHow it will be judged
A claim of absence, judged on a paired t-test of total time labelled displaced, dom against sub per session. Reproduced when p >= 0.05.
- fig2a-win-lose-segregated-before-cue
mPFC trajectories for win and lose trials, and for self and other port entries, occupy segregated subspaces, already before cue onset.
Neural trajectories during the cue and port entries of the self or other (competitor) for win or lose trials occupied segregated neural activity subspaces even before the cue onset
Figure 2a; Results, mPFC reflects rank and winningHow it will be judged
The paper gives no test and no measure of segregation. With the shared trajectory method, the bin-by-bin distance between the win and lose trajectories over -5 to 0 s is compared with the distance between trajectories of two random halves of the data over 50 iterations, the only reference distance the paper reports (ED Fig. 7g-h, right). The paper gives no threshold for this comparison; readings (a) the win against lose distance exceeds the mean random-half distance, (b) it exceeds the 95th percentile of the random-half distances. Reproduced when, for dom and for sub, both readings agree that it exceeds; when they disagree the claim is methods_underspecified.
- ed6d-alignment-lower-dominants
The alignment of win and lose trajectories is lower for relative dominants than for relative subordinates.
alignment of win and lose trajectories was significantly lower for dominant mice
Extended Data Fig. 6dHow it will be judged
Alignment index of Methods 14 (after Miri et al.), Tr(P1' Cov(D2) P1) / Tr(P2' Cov(D2) P2), with D1 and D2 the neurons by time-bin mean rates for win and lose trials (-5 to 5 s, 100 ms) and P1, P2 their principal components, per leave-one-animal-out iteration (13 per group). The number of PCs is not given; readings (a) as many as reach 90% of each condition's variance, (b) the first 8, as in the SVM, (c) the first 3, as plotted. Reproduced when the rank-sum gives p < 0.05 with dominants lower.
- fig2b-subordinates-longer-trajectories
Relative subordinates have longer neural trajectories than relative dominants for win and for lose trials.
Relative subordinates had longer neural trajectories compared to relative dominants
Figure 2b; Results, mPFC reflects rank and winningHow it will be judged
Shared trajectory method, tone-aligned win and lose trials, groups by relative rank in the session. Per leave-one-animal-out iteration (14 dom, 13 sub, giving 25 error df), length over baseline (-5 to 0 s) and over event (0 to 5 s); 2-way RM-ANOVA, rank between and event within. Reproduced when, for win and for lose, the rank effect has p < 0.05 with sub longer, under the shared readings.
- fig2b-subordinates-larger-variance
Firing rate variance is higher for relative subordinates, and firing rates do not change faster for them.
our analyses revealed larger firing rate variance, but not faster firing rate changes, for relative subordinate mice
Figure 2b, right; Results, mPFC reflects rank and winningHow it will be judged
Per neuron, dom against sub, KS test and Wilcoxon rank-sum as in the legend. Variance is not defined; readings (a) variance over 100 ms bins of the trial-averaged rate from -5 to 5 s, (b) the same over 0 to 5 s, (c) variance across trials of the 0 to 5 s rate. Reproduced when, under one reading, lose trials give p < 0.05 in both tests and win trials give KS p < 0.05, with sub higher.
- ed6c-rate-of-change-dominants-win
Firing rate rate of change is higher for relative dominants in win trials only.
Firing rate rate of change is higher for relative dominants only in win trials
Extended Data Fig. 6cHow it will be judged
Per neuron, dom against sub. Rate of change is not defined; readings (a) mean absolute difference between adjacent 100 ms bins of the trial-averaged rate over -5 to 5 s, (b) the same over 0 to 5 s. Reproduced when, under one reading, win trials give p < 0.05 in both tests with dom higher and lose trials give p >= 0.05 in both (absence judged on those tests).
- ed6b-intermediates-relative-rank
Within intermediate-rank mice (ranks 2 and 3), trajectories for win and lose trials are longest when the mouse is the relative subordinate.
Neural trajectory lengths (using principal components that captured 90% of variance) for win and lose trials are highest for relative subordinates in intermediate (ranks 2 or 3) mice
Extended Data Fig. 6bHow it will be judged
Shared trajectory method restricted to sessions of rank 2 and 3 mice, grouped by relative rank. Reproduced when both rank effects have p < 0.05 with sub longer.
- ed6ef-port-entry-trajectories
Trajectory lengths for self and other port entries, during the tone and during the ITI, are higher for relative subordinates.
trajectory lengths (using PCs that captured 90% of variance) for self entry (top) and other entry (bottom) during the tone are higher for relative subordinates
Extended Data Fig. 6e-fHow it will be judged
Shared trajectory method on port entries of the subject (self) and of the competitor (other), during the tone or outside it (ITI). Reproduced when all four rank effects have p < 0.05 with sub longer; partially reproduced for two or three.
- ed7b-distance-by-trial-type-not-rank
In the recording cohort, distance to the reward port differed by trial type but not by rank.
Distance to reward port differed by trial-type but not by rank
Extended Data Fig. 7bHow it will be judged
Per trial, mean AlphaTracker distance to port; the window is not given, readings (a) -5 to 0 s, (b) 0 to 5 s, (c) -5 to 5 s from tone onset, from ED Fig. 7a. Reproduced when, under one reading, trial type has p < 0.05 and rank has p >= 0.05.
- ed7c-distance-correlation-not-rank
The correlation of single-unit firing rate with distance to port does not differ by rank.
Distribution of the correlation coefficients for firing rate and distance to port for the population of mPFC single units did not differ by rank
Extended Data Fig. 7cHow it will be judged
A claim of absence. Per unit, the correlation of binned firing rate with distance to port over the session; bin width not given, 100 ms (the PCA bin) is used as a standard choice, and Pearson as the default correlation. Reproduced when the KS test gives p >= 0.05.
- ed7de-trajectories-close-far
Trajectories are longer for relative subordinates in win and lose trials whether the mice were close to or far from the port at tone onset.
Neural trajectory lengths were higher for relative subordinates during win trials in which mice were close or far to the reward port during tone onset
Extended Data Fig. 7d-eHow it will be judged
Shared trajectory method on sessions with matched video conditions (dom 19, sub 18 sessions, leaving out one session per iteration). The split between close and far is not given; readings (a) median split of distance at tone onset, (b) within or outside the port contact range. Reproduced when all four rank effects have p < 0.05 with sub longer under one reading.
- ed7f-trajectories-without-place-cells
Trajectories remain longer for relative subordinates after removing cells correlated with distance to port.
neural trajectories for self entry during the tone are highest for relative subordinates without the distance correlated cells
Extended Data Fig. 7fHow it will be judged
Shared trajectory method on 18 dom and 18 sub sessions after removing cells whose correlation with distance (ed7c) is significant; the level is not given, readings (a) p < 0.05, (b) p < 0.01. The second panel's event is not named; it is read as win trials. Reproduced when both rank effects have p < 0.05 with sub longer.
- ed7hij-random-halves-no-group-effect
Trajectory lengths from two random halves of the data show no group effect, so the rank effect is not due to chance.
trajectory lengths for win and lose trials when data is divided randomly show no effect of group indicating that the effect of rank is not due to chance
Extended Data Fig. 7g-jHow it will be judged
A claim of absence. 50 random splits of the data from 49 sessions of 20 mice into two halves; per split, trajectory lengths of each half for baseline and event; 2-way ANOVA of event by group on the 50 iterations (F(1,196) fits 50 x 2 x 2 values). What is split is not given; readings (a) sessions, (b) neurons. The legend says "bootstrapping"; halves are drawn without replacement, as "halves" implies. Reproduced when all six group effects have p >= 0.05 under the paper's test.
- fig2c-svm-decodes-rank-and-outcome
An SVM decodes competitive success and relative rank from single-trial mPFC population activity above chance, and rank more accurately than outcome.
An SVM was able to decode both competitive success and relative rank
Figure 2c; Results, mPFC predicts future winsHow it will be judged
Shared SVM method, labels win against lose, or dom against sub. Which window forms the features of Figure 2c is not fixed; readings (a) -5 to 5 s, (b) -30 to 30 s. Reproduced when, under one reading, both AUC sets exceed shuffled by rank-sum at p < 0.05 and rank AUCs exceed outcome AUCs at p < 0.05.
- fig2c-predicts-before-trial
mPFC activity predicts whether the next trial will be won or lost about 30 s before the trial begins, with decoding above chance across the trial.
PFC neural activity could predict whether the next trial would be a win or a loss ~30 seconds before the competition trial began
Results, mPFC predicts future wins; Figure 2cHow it will be judged
Shared SVM method on win against lose with features from -30 to 0 s before the cue only. The time resolution of Figure 2c is not given; readings (a) one model on all of -30 to 0 s, (b) separate models on 5 s windows from -30 s, as the -5 to 5 s feature window suggests. Reproduced when AUCs exceed shuffled by rank-sum at p < 0.05 for reading (a), or for the -30 to -25 s window and every later window under reading (b).
- ed8bc-outcome-generalizes-dom-to-sub
Win and lose decoding trained on relative dominants generalizes to subordinates, but decoding trained on subordinates does not generalize to dominants.
mPFC population encoding of win/lose in relative dominants generalizes to relative subordinates
Extended Data Fig. 8b-cHow it will be judged
Shared SVM method on win against lose, trained on one group's trials and tested on the other's (10 models from the folds of the training group). Reproduced when dom/dom, dom/sub and sub/sub exceed shuffled at p < 0.05 and sub/dom gives p >= 0.05 (absence judged on that test).
- ed8d-outcome-winner-loser
Competition outcome decoders trained on winner mice and tested on loser mice, and the reverse, perform above chance during the cue, and better during the cue than at baseline.
Decoder performance for classifying competition outcome using training data from winner data
Extended Data Fig. 8dHow it will be judged
Winner and loser mice are those that won or lost the majority of trials. Baseline is read as -5 to 0 s and cue as 0 to 5 s (the -5 to 5 s window of the Methods). The legend does not say which panel is which; the order of the sentence is used (left winner to loser). Reproduced when both cue decoders beat shuffle at p < 0.05 and both sign-rank tests on the 10 folds give p < 0.05 with cue higher.
- ed8e-relative-rank-intermediates
Relative rank can be decoded in intermediate-rank mice (ranks 2 and 3).
SVM performance for decoding relative rank specifically for intermediate (ranks 2 or 3) mice
Extended Data Fig. 8e; Results, mPFC predicts future winsHow it will be judged
Shared SVM method on trials of rank 2 and 3 mice, labelled by relative rank. Reproduced when AUCs exceed shuffled at p < 0.05.
- ed8fgh-absolute-rank
Absolute rank can be decoded from mPFC population activity during competition (one model per rank, and rank 1 against rank 4) and while mice perform the task alone.
Notably, we can decode the absolute social rank of individuals from mPFC activity, even when alone
Extended Data Fig. 8f-h; Results, mPFC predicts future winsHow it will be judged
Shared SVM method, one against all for each absolute rank (Methods 14). "Mean performance across ranks" is read as the per-fold AUC averaged over the four models, against the same for shuffled labels. Reproduced when all three comparisons give p < 0.05.
- ed8ij-alone-vs-competition
mPFC activity decodes alone tone trials from win trials of the same session, but does not decode early from late trials within a task.
mPFC population activity can decode between alone tone presentations and win trials during the competition trials
Extended Data Fig. 8i-jHow it will be judged
Shared SVM method on the sessions with alone trials followed by competition trials. Early and late are not defined; read as the first and second half of the trials of each task. Reproduced when alone against win exceeds shuffled at p < 0.05 and both early against late tests give p >= 0.05 (absence judged on them).
- fig2e-alone-vs-competition-trajectories
Tone trajectories are longer for subordinates during competition, and the distance between alone and competition tone trajectories increases from rank 1 to rank 4.
Subordinate mice (rank 4) had larger changes induced by competition with longer tone trajectory lengths during competition
Figure 2d-e; Results, mPFC predicts future winsHow it will be judged
Shared trajectory method, groups by absolute rank (1, 2 and 3, 4) and task (alone, competition), tone-aligned; for competition all trials are read as used. How the common subspace is built is not given; readings (a) one PCA on alone and competition tone rates together, (b) the competition global PCA with alone rates projected. The distance is over all pairs of alone and competition leave-one-out trajectories. Reproduced when rank 4 has the longest competition trajectory with the interaction p < 0.05, and the distance rises from rank 1 to rank 4 with the 1-way ANOVA p < 0.05, under one reading.
- fig3a-functional-clusters
mPFC single units have diverse responses to tone and port-entry events, form functional clusters, and the distribution across clusters differs between relative dominants and subordinates.
mPFC single units showed diverse responses to the tone for win or lose trials and to port entries performed by self or the other
Figure 3a; Extended Data Fig. 9a; Methods 14How it will be judged
As Methods 14 describe, the concatenated mean rates of the six events, cells with baseline rate at or above 0.05 Hz, agglomerative clustering on Euclidean distance cut at 0.4. The linkage is not given; readings (a) MATLAB linkage default (single), (b) Ward. Strong responders have a mean z-score of at least 2 or at most -1 for at least one event (Methods 14, ED Fig. 9a). The paper reports no test of the rank difference in cluster distribution, only the percent enrichment per cluster (Figure 3a right); that part is shown, under each linkage, and not judged. The outcome rests on the counts against the expected values.
- fig3b-win-trials-larger-sub-excitation
In win trials, excitation magnitude differs by relative rank (larger in subordinates) with no difference in the number of responsive cells; lose trials differ in neither.
subordinates had larger responses to win trials and port entries of the other
Figure 3b; Results, Rank-dependent mPFC representationsHow it will be judged
Shared responsive-cell method, tone-aligned win and lose trials, groups by relative rank, 20 mice. Reproduced when win excited magnitude gives p < 0.05 with sub higher, and the four lose tests and the win Fisher test give p >= 0.05 (absence judged on them).
- fig3c-tone-port-entries
Relative dominants have more cells responsive to their own port entries during the tone; for the competitor's port entries the number does not differ but excitation magnitude is larger in subordinates.
number of mPFC self tone port entry responsive cells was higher for relative dominants
Figure 3cHow it will be judged
Shared responsive-cell method on port entries during the tone, by self or by the competitor. Reproduced when self Fisher gives p < 0.05 with dom more and other excitation gives p < 0.05 with sub larger, and other Fisher gives p >= 0.05.
- fig3d-iti-port-entries
During the ITI, relative dominants have more cells responsive to self and to other port entries, while subordinates have larger excitation magnitudes.
relative dominants had more mPFC responsive cells to self port entries during the ITI while relative subordinates had larger excitation magnitude
Figure 3dHow it will be judged
Shared responsive-cell method on port entries outside the tone. Reproduced when both Fisher tests give p < 0.05 with dom more and both excitation tests give p < 0.05 with sub larger.
- ed9b-alone-no-rank-difference
While mice perform the task alone, rank 1 and rank 4 mice do not differ in the number of responsive cells or response magnitude to the tone, tone port entries or ITI port entries.
During competition, but not while alone, relative dominants had more cells that were responsive to self port entries
Extended Data Fig. 9b; Results, Rank-dependent mPFC representationsHow it will be judged
A claim of absence. Shared responsive-cell method on alone sessions of rank 1 and rank 4 mice. Reproduced when the three Fisher tests and the rank-sum tests on magnitudes all give p >= 0.05; counts are reported against the legend's.
- ed9cd-self-entry-cells-competition
During competition, dominants have more cells responsive to self port entries than subordinates by absolute rank (1 against 4, during the tone and the ITI), and within intermediate mice by relative rank during the tone (not significant during the ITI, p = 0.06).
Total responsive cells and response magnitude to task-relevant event during social competition for absolute rank 1 vs rank 4
Extended Data Fig. 9c-dHow it will be judged
Shared responsive-cell method. Reproduced when the rank 1 against 4 self entry tests (tone and ITI) and the intermediate tone test give p < 0.05 with dominants having more. The intermediate ITI test is reported and not judged.
- fig4a-ex-vivo-latency
Ex vivo, ChrimsonR-expressing mPFC-BLA and mPFC-LH cells respond to red light with shorter latencies than neighbouring cells, which sets the phototagging latency threshold.
average photo-latency of action potentials (AP) or excitatory postsynaptic potentials (EPSPs) for mPFC-BLA (left plot) and mPFC-LH (right plot) ChrimsonR+ cells compared to neighboring cells
Figure 4aHow it will be judged
Latency from light onset to the peak AP or EPSP, averaged over 10 pulses (Methods 16). Reproduced when both t-tests give p < 0.05 with ChrimsonR+ shorter.
- fig4b-phototagged-counts
Phototagging identified mPFC-LH and mPFC-BLA neurons in 3 of 12 and 3 of 8 mice.
3 out of 8 mice had mPFC-BLA phototagged cells and 3 out of 12 mice had mPFC-LH phototagged cells
Methods 12; Figure 4b; Extended Data Fig. 10cHow it will be judged
Phototagging as the shared method gives. ED Fig. 10c gives 43 mPFC-LH neurons and Figure 4b 42; both readings are accepted. Mouse counts are compared exactly.
- fig4b-lh-stronger-reward-excitation
mPFC-LH neurons fire more than mPFC-BLA neurons during reward delivery in win trials of the competition, but not when the task is performed alone; mPFC-LH tone responses differ between alone and competition.
We found that mPFC-LH neurons had stronger excitation to reward delivery than mPFC-BLA neurons during reward competition, but not when performing the task alone
Figure 4b; Results, mPFC-LH neurons modulate dominanceHow it will be judged
Per tagged neuron, the mean response over win trials (competition) or tone trials (alone). The legend's "reward delivery (0-2 s)" is read (a) as 0 to 2 s after delivery (2 to 4 s after tone onset), (b) as 0 to 2 s after tone onset; the measure is (a) z-scored rate, as the legend's alone comparison uses, (b) raw rate. The alone against competition test is not named; it is read as a rank-sum (different neuron counts), over the tone read as (a) 0 to 10 s, (b) 0 to 5 s. Reproduced when, under one reading, LH exceeds BLA at p < 0.05 in competition, the alone comparison gives p >= 0.05, and mPFC-LH alone against competition gives p < 0.05.
- ed10c-lh-responsive-proportions
Proportions of cells responsive to tones and port entries differ between mPFC-LH and non-phototagged cells.
Responsive cells to tones and port entries while performing the reward task alone vs in social competition
Extended Data Fig. 10cHow it will be judged
Shared responsive-cell method. The legend does not say which event the Fisher test is for; a Fisher test on the printed counts gives 0.006 for alone entries during tone and 0.06 for competition self entries during ITI, neither 0.011. Readings are each event of the legend. Reproduced when, under one reading, the responsive fraction of mPFC-LH cells exceeds that of non-phototagged cells at p < 0.05, and the counts are within tolerance.
- fig4d-lh-stimulation-dominance
Stimulating mPFC-LH neurons in subordinate mice increased trials won and reward port occupation and decreased time being displaced.
ChR2-expressing mice won more rewards during the entire competition, had greater reward port occupation, and spent less time being displaced when they received optical stimulation
Figure 4c-d; Results, mPFC-LH neurons modulate dominanceHow it will be judged
Light OFF session against light ON session, per mouse, whole session (Methods 5). Reproduced when the two interactions give p < 0.05, the ChR2 post hoc tests give p < 0.05 with ON higher, and the ChR2 displacement test gives p < 0.05 with ON lower.
- ed10ek-no-other-behaviour-effects
mPFC-LH stimulation did not change reward latency alone, home-cage feeding, social preference, open-field anxiety, place preference or effort-based choice.
stimulating mPFC-LH neurons did not affect reward-seeking behavior while performing the reward competition assay alone, feeding in the home cage, anxiety, sociability, place preference or general effort
Extended Data Fig. 10e-k; Results, mPFC-LH neurons modulate dominanceHow it will be judged
A claim of absence, judged on the paper's tests; paired t-tests for e and g, and 2-way RM ANOVA (light, virus, interaction) for f, h, i and k. Reproduced when none finds an effect at p < 0.05.
Pipeline
- Paper
- done (1 of 1)
- Claims
- done
- Claims check
- done (ready)
- Attempt
- todo
- Questions
- none
- Verified
- none