Claims 31
From the paper Diversity in neural firing dynamics supports both rigid and learned hippocampal sequences · claims file · checked 2026-11-07: ready
- place-cell-count
Of the 562 well isolated CA1 pyramidal cells recorded in the eight sessions, 491 had place fields on the novel maze, counted separately for each running direction on the linear mazes, which is 54% of pyramidal cells by running direction.
constructed from the firing-rate vectors of place cells ( n = 491 cells) during track running (eight novel exploration sessions) ... In total, 562 well isolated pyramidal cells were included in this study ... Note that 54% of all putative pyramidal cells were considered to have place fields by running direction
Results, paragraph 2; Materials and Methods, Animals, Surgery and Data Collection; Figure S2 legend; Figure S19 legendHow it will be judged
Pyramidal cells as the data set's putative cell types label them. Place fields by the place-field method of the opening comment, per running direction on the linear mazes and once on the circular maze. The denominator of 54% is open; readings (a) pyramidal cells times the number of running directions of their session (2 on linear mazes, 1 on the circular maze); (b) the 562 pyramidal cells. Reproduced when the counts are within tolerance under one reading.
- fig1c-sequence-content
Ripple-event sequence scores (rZ) are above chance in PRE, MAZE and POST, highest on the MAZE, and higher in POST than in PRE sleep (replay).
Cumulative distribution of rZ for PRE, MAZE, and POSTevents ... Sequence content increases from PRE to POST
Figure 1C; Figure S5 legend; Figure S6 (per animal); Results, paragraph 2How it will be judged
Events, decoding and rZ by the ripple-event and Bayesian replay methods of the opening comment. Tests from the Figure 1C legend, a sign-rank test of rZ against zero within each epoch, and a Kruskal-Wallis test across PRE, MAZE and POST followed by Tukey-Kramer tests. Reproduced when rZ is above zero in all three epochs (p < 0.05), POST rZ exceeds PRE rZ and MAZE exceeds both (Tukey-Kramer p < 0.05), under all readings of Y, L, S and D; partially reproduced when POST exceeds PRE but another part fails. The Figure 1C panel values are not available here, so no level is fixed; Figure S6 shows MAZE highest in each of the four animals and POST above PRE in three of them (not in animal B), which the per-animal results are compared with. The time-bin-shuffle rZ of Figure S22B is shown as a variant.
- figs5b-abs-r-above-null
Absolute weighted correlations of ripple events exceed those of the shuffled null in PRE, MAZE and POST, and the POST minus PRE increase is larger in the observed than in the null data.
Note that in all three epochs, a greater degree of sequence content is observed than expected by chance
Figure S5BHow it will be judged
Events and shuffles as in fig1c-sequence-content. Per the Figure S5B legend, observed against null |r| within each epoch by rank-sum test, and the POST - PRE change of mean |r| tested by bootstrap. Reproduced when observed |r| exceeds null |r| in all three epochs at p < 0.05 and the observed POST - PRE change exceeds the null change (bootstrap 95% interval of the difference above zero), under the readings of fig1c-sequence-content.
- figs5e-duration-unchanged
Ripple-event duration did not differ between PRE and POST (median 183 ms against 186 ms).
While ripple-event duration did not differ between the PRE and POST epochs (PRE median duration 183ms, POST median duration 186ms, p < 0.07, Rank-Sum test)
Figure S5E legendHow it will be judged
A claim of absence, judged by the paper's rank-sum test of PRE against POST event durations at p < 0.05. Which events are plotted is open; readings (a) all ripple events of the opening comment; (b) the events used for Bayesian replay (at least 100 ms, enough place cells). Reproduced when the test is not significant and both medians are within tolerance under one reading; the size of the difference is reported either way.
- figs5f-more-participation-post
A greater percentage of place cells participated (fired at least one spike) in POST than in PRE ripple events.
a significantly greater percentage of place cells participated (fired at least one spike in) POST epoch events
Figure S5F legendHow it will be judged
Per event, the percentage of the session's place cells (per running direction template under reading D-a) firing at least one spike, on the events of fig1c-sequence-content. The Figure S5 legend names no test for panel F; the rank-sum test it names for panel E is used. Reproduced when POST exceeds PRE at p < 0.05.
- figs9b-pcc-post-higher
Place cells contribute more to POST than to PRE sequence content (median PCC -0.011 in PRE, 0.183 in POST), and the median change dPCC is 0.159.
Note that consistent with the main Δreplay effect (Fig. 1C, POST > PRE sequence scores (rZ)), cells show significantly higher POST, compared to PRE PCC scores
Figure S9B (values read from the panel titles)How it will be judged
PCC by the method of the opening comment, for all place cells. Panel titles print "p ~ 0", recorded as 0. The tolerance on the PRE median is not 10% of a value near zero; it is about 10% of the gap between the PRE and POST medians (0.194). Reproduced when POST PCC exceeds PRE PCC (rank-sum p < 0.05), dPCC exceeds zero (signed-rank p < 0.05), and the three medians are within tolerance, under readings Y and D; partially reproduced when the tests hold but a median falls outside its tolerance.
- fig2b-strong-contributors
216 place cells contributed significantly to PRE or POST sequence content; 97 in PRE, 190 in POST, 71 in both, 26 in PRE only and 119 in POST only.
Neurons that showed significant PCC scores in either PRE or POST were considered to be strongly contributing to sequence formation ( n = 216 neurons
Results, paragraph 2; Figure 2B; Figure S19A (counts read from the panel)How it will be judged
Counts of place cells (per running direction) with per-event PCC significant at p < 0.05 one-tailed, under readings P-a and P-b of the opening comment. Reproduced when all four counts are within tolerance under one reading; partially reproduced when the total is within tolerance but a subset count is not.
- fig2-pre-contributors-persist
The majority (73%) of neurons strongly contributing to PRE sequences also contributed to POST sequences.
Importantly, the majority of neurons strongly contributing to PRE (73%) also contributed to POST sequences
Results, paragraph 2; Figure S19AHow it will be judged
From the cells of fig2b-strong-contributors, the number significant in both PRE and POST over the number significant in PRE, under readings P-a and P-b. Reproduced when within tolerance under one reading.
- rigid-plastic-counts
Splitting strongly contributing cells at each session's median dPCC gives 105 rigid and 105 plastic cells.
the strongly contributing cells equal to or below the 50th percentile of that session’s ΔPCC were considered rigid cells (n = 105 cells, blue x’s), while the strongly contributing cells above the 50th percentile of that session’s ΔPCC distribution were considered plastic cells (n = 105 cells
Materials and Methods, Rigid and Plastic Cell ClassificationHow it will be judged
Classification as in the opening comment (per session and running direction, the median cell dropped for odd counts). Reproduced when both counts are within tolerance.
- fig2d-high-dpcc-shuffle-abolishes-replay
Shuffling the place fields of the 10 to 20% of cells with the highest dPCC eliminates the POST over PRE increase in sequence content, while it survives shuffling of the cells with the lowest dPCC.
Replay (POST > PRE) was eliminated after shuffling or removal of the top 10 to 20% of cells with the highest ΔPCC scores, whereas it remained after shuffling or removal of the bottom 75%
Figure 2D; Results, paragraph 2How it will be judged
Per the Methods (Per Cell Contribution, last paragraph) and the Figure 2D legend, within each session an increasing percentage of place cells (ranked by dPCC, from the highest or from the lowest) has its maps circularly shuffled (1,000 templates), the replay analysis is rerun, and mean POST rZ - mean PRE rZ is pooled over sessions, with a bootstrapped 95% interval. Percentage steps are not given; 5% steps are a standard choice. The expected value and tolerance stand for the range 10 to 20%, widened by 10% of each bound (9 to 22%). Reproduced when, shuffling from the highest dPCC, the change first falls to zero or below between 9 and 22%, and, shuffling from the lowest dPCC, it stays above zero up to 75%, under readings Y and D.
- figs12-high-dpcc-exclusion
Excluding the 10 to 20% of cells with the highest dPCC turns the POST minus PRE change in sequence content negative, while excluding up to about 75% of the lowest dPCC cells leaves POST above PRE.
Note that even after excluding up to ~75% of ‘low’ ΔPCC cells, POST epoch sequence content exceeds PRE epoch sequence content ... after the exclusion of just 10 to 20% of cells PRE epoch sequence content exceeded POST epoch sequence content
Figure S12How it will be judged
As fig2d-high-dpcc-shuffle-abolishes-replay, with cells excluded from the template and event selection (the 100 ms and 10% or 5 cell criteria applied after exclusion, per the Methods) in place of shuffled. Reproduced when, excluding from the highest dPCC, POST - PRE rZ falls below zero between 9 and 22% excluded, and, excluding from the lowest, the lower bound of its bootstrapped 95% interval stays above zero up to 75%, under readings Y and D.
- figs2-rigid-plastic-along-track
Rigid and plastic cells' place fields are similarly distributed along the track, both enriched near the start and goal ends.
Rigid and plastic neurons were similarly distributed along the track
Results, paragraph 2; Figure S2 legendHow it will be judged
The paper shows this only as rate maps and gives no test, so the test is open. Measure, each cell's peak place-field position (per running direction) as a fraction of track length from the start of the run. Readings (a) a two-sample Kolmogorov-Smirnov test of rigid against plastic positions; (b) a rank-sum test of the distance of each peak from the nearer track end. Reproduced when neither reading finds a difference at p < 0.05; the fractions of fields in the outer fifths of the track are reported for each group.
- fig3a-plastic-more-specific
On the novel maze plastic cells have higher place specificity than rigid cells and fewer place fields.
Plastic neurons had higher place-specific indices and fewer place fields compared with rigid cells
Figure 3A; Results, paragraph 3; Figure S19EHow it will be judged
The Figure 3A panels are not available here; Figure S19E, which the legend says to compare to Figure 3A, shows six measures, within/outside place-field rate ratio, information per spike, information per second, spatial coherence, place-field stability and number of place fields, and marks every one of the six as significant, plastic higher on the first five and lower on the number of fields. Measures as in reference 32 (Mizuseki et al. 2012) per the Methods; the stability split is not given here (readings (a) first against second half of the maze epoch; (b) odd against even laps). Rank-sum test, rigid against plastic (Figure 3A legend). Reproduced when plastic cells are higher on all five specificity measures and have fewer place fields, each at p < 0.05, under one stability reading; partially reproduced when the direction holds for all six but fewer reach p < 0.05.
- fig3b-plastic-specificity-increases
Plastic but not rigid cells steadily increased their spatial specificity during maze running.
plastic but not rigid neurons increased their spatial specificity steadily during learning
Figure 3B; Results, paragraph 3; Figure S19FHow it will be judged
Per lap, each cell's percentage of spikes inside its place field(s) (the measure of Figure S19F, the legend's "infield versus outside field firing ratio"), using only laps with at least 5 spikes from the cell (Figure S13 legend); per cell, the Pearson correlation of this value with lap number. Tests per the Figure S13C legend, a signed-rank test of each group's correlations against zero and a rank-sum test between groups. Reproduced when plastic cells' correlations are above zero (p < 0.05), rigid cells' are not, and plastic exceeds rigid (p < 0.05). The within-field rate gain of Figure S13C is shown as a variant.
- figs13-outfield-rate-falls
Neither group's overall firing rate changed significantly over laps, and neither group's within-field firing rate increased significantly; plastic cells' increase in specificity came from a fall in their out-of-field firing rate, which rigid cells did not show.
Notably, neither cell group showed a significant increase in within-field firing rate over the course of maze running. (E) Conversely, a significant decrease was observed in plastic cells’, but not rigid cells’, outside of field firing rates
Figure S13A, S13D and S13EHow it will be judged
Per cell, the Pearson correlation of lap rate (overall, within field, outside field) with lap number, laps with at least 5 spikes, running above 5 cm/s. Signed-rank test of each group's correlations against zero (Figure S13C legend). Reproduced when the outside-field correlations of plastic cells are below zero (p < 0.05) and those of rigid cells are not, neither group's overall-rate correlations differ from zero at p < 0.05, and neither group's within-field correlations are above zero at p < 0.05.
- fig4a-rigid-fire-faster
Rigid cells fire faster than plastic cells in offline states (session-wide median 0.76 against 0.54 Hz; PRE 0.762 against 0.479 Hz; POST 0.753 against 0.472 Hz), but not during maze running (1.09 against 1.13 Hz).
Session-wide firing rates of rigid neurons were significantly higher compared with plastic neurons
Figure 4A; Results, paragraph 4; Figure S14, firing rate row (values read from the panel titles)How it will be judged
Firing rate as in Figure S14, spikes per second in offline states of each epoch, and during running on the maze for MAZE (the measure for MAZE is not stated; readings (a) all MAZE time; (b) running above 5 cm/s). "Session-wide" is not defined; readings (a) offline time of PRE and POST together; (b) the whole recording. Rank-sum tests. Reproduced when the session-wide, PRE and POST differences are significant (p < 0.05) with rigid higher, the MAZE difference is not (rank-sum p >= 0.05), and the medians are within tolerance under one reading; partially reproduced when the tests hold and the medians do not.
- figs14-plastic-ripple-gain
Plastic cells have a higher within-ripple firing rate gain than rigid cells (POST median 5.75 against 4.52) and a larger PRE to POST increase in gain (0.81 against 0.14).
Ripple-related spiking, bursting, and pair-wise coactivation were higher in plastic versus rigid neurons
Figure 4A; Results, paragraph 4; Figure S14, within-ripple firing rate gain row (values read from the panel titles)How it will be judged
Gain as in the Figure S14 legend, within-ripple rate over offline rate, per epoch. Rank-sum tests, plastic against rigid. The rigid change median is near zero, so its tolerance is set at 0.1, about 10% of the gap between the two groups' changes. Reproduced when plastic exceeds rigid for POST gain and for the change at p < 0.05 and the four medians are within tolerance; partially reproduced when the tests hold and a median does not.
- figs14-plastic-ripple-bursting
Plastic cells burst more within ripples than rigid cells in POST (median 27.3% against 23.3% of ISIs under 6 ms) and increase their within-ripple bursting more from PRE to POST (2.81 against 1.13 percentage points).
How it will be judged
Within-ripple burstiness as the percentage of within-ripple ISIs under 6 ms. Whether ISIs spanning two ripple events count is open; the reading kept is ISIs between spikes of one event (standard), with all ISIs inside concatenated ripple time as a variant. Rank-sum tests. Reproduced when both tests hold with plastic higher and the medians are within tolerance; partially reproduced when the tests hold and a median does not.
- figs16ab-plastic-synchrony-increase
The PRE to POST increase in ripple-triggered firing is present in plastic but not rigid cells, and the PRE to POST increase in pair-wise coactivity is larger among plastic cells.
The PRE to POST increases in ripple-event triggered firing rates seen when all place cells are taken into account (A, left panel) are uniquely present in plastic (A, third-panel) but not rigid (A, second panel) cells ... the PRE to POST increase in pair-wise co-activity observed across all place cells (B, left panel) is more robustly observed in the plastic cell population
Figure S16A and S16B; Figure 4A and Figure S19G (change in pair-wise coactivity)How it will be judged
Ripple-triggered rate from -50 to 100 ms around event onset (the panel's axis), per cell, PRE and POST. The legend gives no test for panel A; readings (a) a signed-rank test of each cell's POST - PRE mean rate in 0 to 100 ms; (b) the same at the peak bin. Coactivity as in the opening comment. Figure S16B shows a significant PRE to POST increase within both rigid and plastic pairs, larger in plastic pairs, with no test between groups; Figure 4A (and Figure S19G, which shows it with p < 0.023) compares the per-cell POST - PRE change in pair-wise coactivity between rigid and plastic cells by rank-sum test. How a cell's change is formed from its pairs is not stated; readings (c) the mean change over the cell's pairs with all other place cells on other shanks; (d) the mean over its pairs within its own group. Reproduced when plastic cells' rate increase is significant (p < 0.05) and rigid cells' is not under one reading of A, and plastic cells' coactivity change exceeds rigid cells' (rank-sum, p < 0.05) under one reading of the per-cell change.
- low-high-rate-split
Splitting place cells into equal halves by offline firing rate gives low-rate cells with median 0.39 Hz and high-rate cells with median 1.12 Hz.
Low-rate neurons (median, 0.39 Hz) contributed more spatial information per spike on the maze and displayed increased within-ripple firing-rate gains from PRE to POST sleep than high-rate cells (median, 1.12 Hz)
Results, paragraph 5How it will be judged
The Figure S16 legend points to the Methods for the split, which the Methods do not give. From the text, place cells are split into equal groups by "off-line" sleep firing rate, and the Methods state that cells removed from an analysis are chosen within each session. Readings (a) median split within each session and running direction; (b) median split within each session; (c) a pooled median split; and for the rate, (i) offline PRE and POST together; (ii) offline PRE only. Reproduced when both medians are within tolerance under one reading. The reading(s) that match fix the split for the claims that depend on this one.
- figs16c-low-rate-coding
Low-rate cells carry more spatial information per spike (0.91 against 0.56 bits/spike), have a higher in/out place-field rate ratio (6.38 against 3.86), fewer place fields (log-mean 1.29 against 1.59), and a larger PRE to POST increase in within-ripple gain (0.71 against 0.06) than high-rate cells.
low-rate cells also contributed more information per spike (Low Rate median ... 0.91 bits/spike, High Rate med ... 0.56 bits/spike, p < 2.5x10-9, Rank-Sum test), had a greater within/outside of place field(s) firing ratio (Low Rate med ... 6.38, High Rate med ... 3.86, p < 7.9x10-14), and had fewer place fields on the maze (Low Rate log-mean ... 1.29, High Rate log-mean ... 1.59, p < 3.94 x10-06). In addition, low rate cells tended to show a more pronounced increase in their within-ripple firing rate gain (Low Rate Δripple-gain med ... 0.71, High Rate med ... 0.06, p < 7.585x10-4
Figure S16C legend; Results, paragraph 5How it will be judged
Groups from low-high-rate-split under the reading that matched there. Measures as in fig3a-plastic-more-specific and figs14-plastic-ripple-gain. "Log-mean" is read as the geometric mean. The high-rate gain change is near zero, so its tolerance is 0.07, about 10% of the gap between the groups. Reproduced when all four rank-sum tests give p < 0.05 in the stated direction and the medians are within tolerance; partially reproduced when the tests hold and a median does not.
- fig4c-coactivation-by-rate
In PRE sleep coactivity is dominated by fast-firing cells, whose correlations stay unchanged into POST, while slow-firing cells show the largest PRE to POST increase in coactivation and the greatest replay of maze coactivity.
In PRE sleep, coactivity was dominated by fast-firing neurons, and these correlations remained unchanged into POST sleep. In contrast, slow-firing neurons showed the strongest increases in coactivation from PRE to POST sleep
Figure 4B and 4C; Results, paragraph 4; Figure S16CHow it will be judged
Per the Figure 4B and 4C legends, place cells sorted by session firing rate and grouped in overlapping windows of 20% of cells stepped by 1%; in each window, the mean coactivity of its pairs (different shanks) in PRE and POST, the POST - PRE change, and the partial correlation of RUN and POST coactivity given PRE. Whether ranks are taken within session or pooled is open (readings (a) percentile of rate within session; (b) pooled); "session firing rate" as in fig4a-rigid-fire-faster. The legends name no test for these panels; the Spearman and signed-rank tests below are standard choices, not the paper's. Reproduced when PRE coactivity rises with the window's rate rank, the POST - PRE change and the partial correlation fall with it (Spearman over windows, both p < 0.05), the change in the top window is not significant (signed-rank over pairs, p > 0.05), and, per Figure S16C, the low-rate half's mean change exceeds the high-rate half's (rank-sum, p < 0.05). Partially reproduced when the slow-firing increase holds and the PRE dominance or the top-window stability does not.
- fig4d-slow-cells-replay
Slow-firing, but not fast-firing, place cells increased their contribution to sequences from PRE to POST sleep.
Slow-firing, but not fast-firing, neurons increased their contribution to neuronal sequences from PRE to POST sleep
Figure 4D; Results, paragraph 5; Figures S20D, S21B and S22EHow it will be judged
The Bayesian replay analysis rerun with templates of only the low-rate or only the high-rate place cells (low-high-rate-split), the event criteria applied to the kept cells. The test of the change is open; readings (a) a rank-sum test of POST against PRE rZ within each group (as in Figure 1D); (b) a bootstrapped 95% interval of mean POST - PRE rZ (as in Figure 2D). Reproduced when the low-rate group's POST - PRE rZ is above zero (p < 0.05) and the high-rate group's is not, under each reading; partially reproduced when the low-rate increase holds and the high-rate group also increases but less. The paired-latency (Figure S20D) and time-bin-shuffle (Figure S22E) versions are shown as variants.
- fig4e-high-gain-cells-replay
The PRE to POST increase in sequence content was limited to cells with high ripple-specific recruitment (within-ripple rate gain).
Conversely, PRE to POST increases in sequence content were limited to cells that showed a high degree of ripple-specific recruitment (gain)
Figure 4E; Results, paragraph 5; Figures S20E, S21C and S22EHow it will be judged
As fig4d-slow-cells-replay, with the place cells split into equal halves by within-ripple firing-rate gain, under the split and epoch readings that matched in low-high-rate-split. Reproduced when the high-gain group's POST - PRE rZ is above zero (p < 0.05) and the low-gain group's is not, under each reading of the test.
- figs16de-more-ripples-post
Ripple events occur more often in POST than in PRE (median inter-event interval 2.47 s against 3 s), as do purely LFP-detected sharp wave-ripples (1.06 s against 1.21 s).
Increased synchrony (B) in the POST compared to PRE epoch may be attributable to both increases in ripple-triggered spiking (A) as well as to an increase in the incidence of ripple events
Figure S16D and S16E (medians read from the panel titles)How it will be judged
Intervals between consecutive events within offline states, intervals under 60 s only (legend); rank-sum test of PRE against POST intervals ("p ~ 0" printed, recorded as 0). Ripple events as in the opening comment; LFP-SWR events under readings L-a and L-b. Reproduced when POST intervals are shorter for both event types at p < 0.05 and the ripple-event medians are within tolerance; the LFP-SWR medians depend on the open detector and are reported, matching under one L reading counts toward reproduced.
- figs17-coactivity-preservation
Pair-wise coactivity is weakly preserved from PRE to RUN (r = 0.045), more strongly from RUN to POST (r = 0.16), and most strongly from PRE to POST (r = 0.34).
the similarity between RUN and POST co-activity patterns (middle panel) was significantly ... greater (corr(RUN, POST) - corr(RUN, PRE) = 0.11, p ≈ 0, Fisher's Z-test). However, the similarity between RUN and POST was likewise far exceeded by the similarity between PRE and POST co-activity patterns (corr(PRE, POST) - corr(RUN, POST) = 0.18, p ≈ 0, Fisher's Z-test)
Figure S17A and S17C (r values read from the panels)How it will be judged
Coactivity as in the opening comment, all place-cell pairs on different shanks; Pearson correlation across pairs between epochs; differences of correlations by Fisher's Z test (as two independent correlations, the test's usual form; a test for dependent correlations is the other reading). Reproduced when both differences are significant (p < 0.05) in the stated direction and the values are within tolerance; partially reproduced when the order RUN-PRE < RUN-POST < PRE-POST holds but values fall outside tolerance.
- figs17d-rigid-plastic-coactivity
Rigid cell pairs preserve PRE coactivity into RUN (r = 0.139) and into POST (r = 0.31) more than plastic pairs (-0.02 and 0.09), while plastic pairs show higher RUN to POST preservation (0.36 against 0.11).
Rigid cell, but not plastic cell, pairs showed preservation of PRE co-activity patterns during the RUN. Conversely, plastic cell pairs showed a significantly higher preservation of co-activity patterns from the RUN to the POST
Figure S17D (r values read from the panels); Results, paragraph 4How it will be judged
As figs17-coactivity-preservation within rigid-rigid and plastic-plastic pairs; groups compared by Fisher's Z test (Figure S17D legend). The plastic PRE-RUN r is near zero, so its tolerance is 0.05, read as consistent with zero. Reproduced when the three group differences are significant (p < 0.05) in the stated direction; the r values are reported against tolerance, and a mismatch in values with the differences holding is partially reproduced.
- figs18-temporal-bias
Pair-wise temporal biases are not preserved from PRE to RUN (r = -0.009) but are preserved from RUN to POST (r = 0.117) and PRE to POST (r = 0.201); the PRE to POST preservation is carried by rigid cell pairs (r = 0.29) and the RUN to POST preservation by plastic cell pairs (r = 0.26).
When all place cells are included in the analysis, no similarity is observed between PRE and RUN (left panel) temporal biases. However, temporal biases are strongly preserved from the RUN to POST epochs (middle panel), as well as from the PRE to POST epochs (right). (C) Limiting this analysis to rigid (blue) or plastic (red) cell pairs, reveals that the preservation of temporal spiking patterns from PRE to POST and from RUN to POST is attributable to the rigid and plastic cell subpopulations, respectively
Figure S18B and S18C (r values read from the panels); Results, paragraph 4How it will be judged
Temporal bias and its median r over 1,000 sign randomisations as in the opening comment. The PRE-RUN value is near zero; its tolerance of 0.05 reads it as consistent with zero. Reproduced when the PRE-RUN r is not significant (p > 0.05), RUN-POST and PRE-POST are (p < 0.05), the rigid PRE-POST and plastic RUN-POST r are significant while the plastic PRE-POST (-0.037) and rigid RUN-POST (0.024) are not, and the values are within tolerance; partially reproduced when the significance pattern holds and a value does not.
- figs11-cluster-quality-similar
Rigid and plastic cells have similar waveforms and similar cluster quality and stability across PRE, MAZE and POST, so clustering errors do not explain their differences.
Note that rigid and plastic cell waveforms are similar to each other and stable from the PRE to the POST epochs
Figure S11; Results, paragraph 3How it will be judged
A claim of absence, judged by the paper's rank-sum tests of rigid against plastic cells at p < 0.05 (Figure S11 legend). Quality is the median Mahalanobis distance of each spike's PCA features to its cluster mean, for spikes in PRE ripple events, MAZE running and POST ripple events; stability is the symmetric Mahalanobis distance between a cell's clusters in two conditions; first spikes of bursts (ISI under 6 ms) only. Reproduced when no comparison is significant; the effect sizes are reported either way.
- fig2e-plastic-removal
The PRE to POST increase in sequence content is attributable to a small number of plastic cells, judged by removing rigid or plastic cells from the replay analysis.
Effect on sequence content of removal of rigid (black) or plastic (gray) neurons. The PRE to POST increase in sequence content is attributable to only a small number of plastic cells.
Figure 2E legendHow it will be judged
The Figure 2E panel is not available here and its legend gives no method beyond removal. Removal as in figs12-high-dpcc-exclusion (cells excluded from template and event selection, the 100 ms and 10% or 5 cell criteria applied after exclusion, within each session, results pooled), excluding all rigid or all plastic cells of rigid-plastic-counts; mean POST rZ - mean PRE rZ with a bootstrapped 95% interval, the annotation of Figure 2D and Figure S12. Reproduced when, after removing the plastic cells, the interval of POST - PRE rZ reaches zero or below, and after removing the rigid cells it stays above zero, under readings Y and D; partially reproduced when both stay above zero but removing plastic cells lowers POST - PRE rZ more than removing rigid cells.
- figs19c-pcc-change-per-cell
Testing each place cell's per-event PCC scores between PRE and POST, 2.2% of cells had significantly higher PRE than POST scores and 21.4% significantly higher POST than PRE scores.
2.2% of cells showed significantly greater PRE compared to POST PCC scores while 21.4% showed higher POST > PRE
Figure S19C legendHow it will be judged
Per the Figure S19C legend, for each place cell (491, per running direction) two one-tailed rank-sum tests of its per-event PCC scores in PRE events against those in POST events, one for each direction, counted at p < 0.05 (the counting threshold of the panel); the denominator is all place cells (the panel's axis, "491 total"). PCC as in figs9b-pcc-post-higher. Reproduced when both percentages are within tolerance under readings Y and D; partially reproduced when the POST > PRE percentage is within tolerance and much larger than the PRE > POST one but a value falls outside tolerance.
Pipeline
- Paper
- waiting (not checked)
- Claims
- done
- Claims check
- done (ready)
- Attempt
- todo
- Questions
- none
- Verified
- none