Datasets / DANDI:000249

DANDI:000249

✓ Verified by an independent agent

Innate and plastic mechanisms for maternal behaviour in auditory cortex

cell population imaging microscopy House mouse 98.0 GB 777 files
  • 32 blocked

Counted separately: 1 claim that also uses data from outside the dataset.

  • 1 blocked

Counted separately: 10 claims on data the paper deposited in other archives.

  • 1 reproduced
  • 1 not reproduced
  • 8 blocked

Counted separately: 1 check that the dataset is the data behind the paper.

  • 1 partially reproduced

Tested on dataset version 0.230423.1416

View on DANDI Paper JSON How to read outcomes

Claims

From the paper Innate and plastic mechanisms for maternal behaviour in auditory cortex

Reproducing Schiavo et al. (2020) from dandiset 000249 and its figshare recordings 12

Notebook reproduce_schiavo_2020.ipynb

Reproduced1

  • ReproducedExtended Data Fig. 9a,brests on an assumptiondata in figshare doi:10.6084/m9.figshare.12613811.v1 (In vitro recordings_oxytocin)

    Oxytocin enhances spiking to repeated 575 ms pulses in auditory cortical slices from stimulus 2 on, while stimulation without oxytocin does not change spike probability.

    evoked spike probability was significantly enhanced 15–30 minutes following the onset of oxytocin wash

    Extended Data Fig. 9a,b

    What the paper reports

    • stim 1: one-way Friedman test with Dunn's correction, p = 0.34 (n = 9 cells)
    • stim 2: one-way Friedman test with Dunn's correction, p = 0.009
    • stim 3: one-way Friedman test with Dunn's correction, p = 0.008
    • stim 4: one-way Friedman test with Dunn's correction, p = 0.001
    • stim 5: one-way Friedman test with Dunn's correction, p = 0.04

    In 9 oxytocin cells spike probability rises from 0.035-0.051 to 0.48-0.56 15 to 30 minutes after wash onset (Friedman chi2 = 51.35, p = 6e-08). Dunn's comparisons, Bonferroni over the five stimuli, give stimuli 2 to 5 adjusted p 0.022, 0.014, 0.0011, 0.020 and stimulus 1 p = 0.071 (paper 0.009, 0.008, 0.001, 0.04 and 0.34), although stimulus 1 also rises (0.044 to 0.48). 5 ACSF controls show no change (Friedman p = 0.56, adjusted p >= 0.79); counting the control window from the first baseline sweep brings in all 7, again with no change (Friedman p = 0.39, adjusted p >= 0.47). Over all 45 pairs only stimulus 4 is significant (p = 0.0098), and Wilcoxon tests find stimulus 1 significant as well (p = 0.0078), so the outcome rests on the comparison family.

    How the test differed from the paper’s method

    • Assumed: the Friedman test runs over ten columns (baseline and post for each of five stimuli), with Dunn's comparisons of baseline against post for each stimulus, Bonferroni over five; the legend gives one p per stimulus, which fits five planned comparisons. All 45 pairs and Wilcoxon tests are variants and change the outcome.
    • A spike is a crossing of 0 mV between a pulse and the next, a standard choice.
    • Baseline is the files the spreadsheet labels baseline or PRE (the first file for one control cell with no such label; the notebook prints each file's label and role); for controls the 15 to 30 minute window starts at the first sweep after the baseline files, a reading of "following baseline acquisition", and from the first baseline sweep as a variant.
    • Under the main reading two of the 7 controls (191202B, 191202C) have no sweeps in the window: each has one 60-sweep baseline file (0 to 11.8 min) and one later file ending 19.3 and 24.8 min after the first baseline sweep.

Partially reproduced1

Not reproduced1

  • Not reproducedFigure 1g; main text, paragraph citing Fig. 1grests on an assumptiondata in figshare doi:10.6084/m9.figshare.12611522.v1 (ACtx slice recordings - Repeated stimulation), in collection doi:10.6084/m9.figshare.c.5043830.v6; the paper's own figshare DOI did not resolve

    In pup-naive auditory cortical slices, IPSCs depress faster than EPSCs at the prototypical stimulation rate (175 ms), while EPSCs and IPSCs depress similarly at 75 or 575 ms.

    whole-cell recordings in pup-naive auditory cortical slices revealed that IPSCs adapted out faster than EPSCs at prototypical repetition rates, while PSCs were similarly depressed at faster or slower stimulation rates

    Figure 1g; main text, paragraph citing Fig. 1g

    At 175 ms (14 EPSC and 13 IPSC cells) IPSCs do not depress more than EPSCs. Per trial, mean Sn/S1 over pulses 2 to 5 is 0.90 (IPSC) against 0.68 (EPSC), p = 0.33, and no pulse differs after Bonferroni (adjusted p 1.0). On per-file mean traces, 0.57 against 0.65, p = 0.39; pulses 2 and 3 adjusted p 0.15 and 0.06. 75 ms does not differ (p 0.52 and 0.72). At 575 ms the unpaired tests do not differ (p 0.82 and 0.08), but on mean traces a paired test finds IPSCs lower (0.55 against 0.66, p = 0.022). Per-trial IPSC ratios are noisy; 35 trials with no first response above baseline were dropped and 17 cell means fall outside 0 to 2. The pulse times read from the protocol fall on the stimulus artefacts (steepest current change a median 0.5 ms after the computed time, within 1 ms for 91.8% of 740 pulses). At 175 ms the per-trial and mean-trace IPSC ratios move in opposite directions (0.90 and 0.57 against 0.68 and 0.65), neither significantly.

    How the test differed from the paper’s method

    • Assumed: the peak is the largest inward (EPSC) or outward (IPSC) current 2 to 30 ms after each pulse (to 2 ms before the next pulse at 50 ms periods) against the mean of the 5 ms before it. The Methods say only "peak amplitude ... normalized to the event evoked by the first stimulation on each trial"; local baselines are the usual choice for trains.
    • Trials with a first response not above baseline give no ratio and are dropped; the mean trace per file is shown as a variant.
    • One-way ANOVA with two groups is run as the equivalent unpaired t-test; a paired test is shown as a variant.
    • 12 of the 13 cells at 575 ms; file 19o11005 (Cell 20) is listed but not on figshare.
    • The 75 ms bin includes trains at 50 ms periods, at the edge of the Methods' 75 plus or minus 25 ms.

Blocked9

  • BlockedMethods, In vivo two-photon calcium imaging· Metadata insufficient

    Retrieval onset ranged from 12 to 109 hours of cohousing in the 13 chronically imaged virgins.

    Retrieval onset ranged from 12–109 hours of cohousing in N=13 virgins

    Methods, In vivo two-photon calcium imaging

    The start of cohousing is not recorded in the files or the metadata, so the hours to retrieval onset cannot be measured. 14 subjects have a session named first retrieval (the paper has 13). The hours from the last session imaged before it (base, 24hrsPRE or 48HrsPRE) are 23.9 to 52.1, and from the last session named base 23.9 to 144.9 (NV50 has only a 24hrsPRE session), but either bounds onset only if cohousing began after that session, which nothing states; they decide nothing.

  • BlockedFigure 1h,i; main text, paragraph citing Fig. 1h,i· Methods underspecifieddata in figshare doi:10.6084/m9.figshare.13056584.v1 (Behavior files, touch times), in collection doi:10.6084/m9.figshare.c.5043830.v6; the paper's own figshare DOI did not resolve

    Of naive virgins able to silence playback with a lever press, only those hearing prototypes significantly increased their press duration by session 8.

    Only virgins listening to prototypes exhibited a significant increase in press duration by session 8 (p=0.03)

    Figure 1h,i; main text, paragraph citing Fig. 1h,i

    What the paper reports

    • prototypes (175 ms): paired two-tailed t-test, p = 0.03 (n = 7 virgins)
    • fast morphs (75 ms): paired two-tailed t-test, p = 0.24 (n = 7 virgins)
    • slow morphs (575 ms): paired two-tailed t-test, p = 0.13 (n = 7 virgins)

    The readings disagree, and the criterion blocks the claim then. Prototypes increase mean press duration under all four (session 1 against 8, p = 0.025 with the median-interval gap and 0.0076 with the 100 ms gap; sessions 1-2 against 7-8, p = 0.032 and 0.021). Fast and slow are not significant under three readings (p 0.081 to 0.73), but with the 100 ms gap and session 1 against 8 slow increases (p = 0.040) and fast reaches p = 0.050. The reading closest to the paper's p-values (0.03, 0.24, 0.13) is the median-interval gap with session 1 against 8 (0.025, 0.081, 0.17). For the paper's authors, which gap ends a press and which sessions were compared. The gap that ends a press under the median-interval rule grows with time in the session (1.5 ms before 100 s, 15 ms to 1000 s, 150 ms after), because a gap must exceed one and a half printed steps.

    How the test differed from the paper’s method

    • Touch times are printed with five significant digits, so a press can be resolved only to 0.1 ms below 10 s, 1 ms below 100 s, 10 ms to 1000 s and 100 ms after; a gap must exceed one and a half local steps (1 ms taken as the step below 100 s) to end a press, under both gap readings, so the median-interval rule ends a press at 1.5, 15 or 150 ms depending on when it falls.
    • The median inter-sample interval (reading a) is the polling period of the task code, taken once, pooled over sessions, from the times printed at 0.1 ms (below 10 s), 0.70 ms; above 10 s the printed intervals are quantized at 1 ms or more. Before revision a mean over the 1 ms part was used (about 0.7 ms); the outcome is unchanged.
    • Two mice miss one session each (NV24 session 5, NV38 session 3); neither is a compared session.
  • BlockedExtended Data Fig. 5b· Methods underspecifieddata in figshare doi:10.6084/m9.figshare.13056584.v1 (Behavior files, touch times)

    The number of lever presses per session did not increase by session 8 in any of the three groups.

    The total number of levers presses in a session (which turns off continuously playing prototypes or morphs) did not increase by session 8, regardless of the stimulus group

    Extended Data Fig. 5b

    What the paper reports

    • 175 ms: paired two-tailed t-test, p = 0.44
    • 75 ms: paired two-tailed t-test, p = 0.19
    • 575 ms: paired two-tailed t-test, p = 0.14

    The readings disagree. No group increases under three readings (all p >= 0.09). With the median-interval gap and sessions 1-2 against 7-8 the fast group increases from 68.5 to 107.4 presses (p = 0.049).

    How the test differed from the paper’s method

    • Touch times are printed with five significant digits, so a press can be resolved only to 0.1 ms below 10 s, 1 ms below 100 s, 10 ms to 1000 s and 100 ms after; a gap must exceed one and a half local steps (1 ms taken as the step below 100 s) to end a press, under both gap readings, so the median-interval rule ends a press at 1.5, 15 or 150 ms depending on when it falls.
    • The median inter-sample interval (reading a) is the polling period of the task code, taken once, pooled over sessions, from the times printed at 0.1 ms (below 10 s), 0.70 ms; above 10 s the printed intervals are quantized at 1 ms or more. Before revision a mean over the 1 ms part was used (about 0.7 ms); the outcome is unchanged.
    • Two mice miss one session each (NV24 session 5, NV38 session 3); neither is a compared session.
  • BlockedExtended Data Fig. 6e,f; main text, paragraph citing Fig. 2g· Otherverifier disagreed at firstdata in figshare doi:10.6084/m9.figshare.12611657.v3 (experienced) and doi:10.6084/m9.figshare.12613496.v2 (naive), in collection doi:10.6084/m9.figshare.c.5043830.v6

    In experienced females, prototype- and morph-evoked EPSCs are similar and negatively correlated with ISI; in naive cortex morph-evoked EPSCs are weaker than prototype-evoked ones and uncorrelated with ISI.

    In experienced females, prototype- and morph-evoked EPSCs were similar in magnitude and negatively correlated with ISI duration

    Extended Data Fig. 6e,f; main text, paragraph citing Fig. 2g

    What the paper reports

    • NV 175 against 125 ms: repeated measures one-way ANOVA, Bonferroni correction, p = 0.006
    • NV 175 against 275 ms: repeated measures one-way ANOVA, Bonferroni correction, p = 0.003
    • NV 175 against 575 ms: repeated measures one-way ANOVA, Bonferroni correction, p = 0.02
    • EV EPSC against ISI: two-tailed Pearson correlation, r = -0.37, p = 0.01 (n = 12 cells)
    • NV EPSC against ISI: two-tailed Pearson correlation, r = -0.13, p = 0.35 (n = 14 cells)

    In the deposited in vivo recordings the current almost never follows the calls, so the claim cannot be tested on the cells it is about. The Methods' evoked measure is in the expected direction at p < 0.05 in 2 of 39 cell recordings, against 3 of 39 for the same measure before the call; no cell has a mean response 0-300 ms after onset at p < 0.05 (mean over cells 0.25 pA for EPSCs); 13 of 110 files have a sweep-locked deflection 0-3 s after onset at p < 0.05, more than the 5.5 expected by chance (binomial p = 0.0034), but only 3 of them, from two cells (naive 190326_e1, both EPSC files, and experienced 190313_e2, one of two), carry a current in the expected direction during the call, first exceeding the null 0.39 to 0.51 s after onset and peaking at 37 to 48 pA; the other 10 peak at scattered times, 9 against the expected direction; and the inclusion rule keeps 1 of 39 (2 s.d., cell 190326_e1) or none (3 s.d.). As computed on every listed cell it would not be reproduced: naive prototype-evoked EPSCs (20.0 pA, largest 5 trials) are not larger than morph-evoked ones (17.4, 18.3, 16.2 pA; adjusted p 0.12, 0.94, 0.20; RM ANOVA p = 0.18), and no variant gives p < 0.05 for all three; experienced EPSCs are not correlated with ISI (r = -0.14, p = 0.35; -0.09 to -0.21 under every variant). No experienced pair differs (smallest adjusted p 0.30 to 1.0) and the naive correlation is not significant (r = -0.07, p = 0.61).

    How the test differed from the paper’s method

    • Call onset, offset and ISI come from the recorded microphone channel. The onset is the start of the sound (envelope above its median by 6 robust s.d.); before revision it was the first loud syllable, which came a median 193 ms (up to 644 ms) late because the first syllable of many calls is quiet, and that rule is run as a variant. In 82 of 83 morph-set files the four call durations form four groups in a fixed order (prototype, 125, 275, 575 ms), and the ISI is assigned by duration rank; one two-sweep file uses the nearest bin of the measured gap. The sets hold the legends' ISIs, not the Methods' 125, 225 and 575 ms.
    • Assumed, for the Methods' inclusion rule ("Only neurons that exhibited at least five significant prototype-evoked E/IPSCs"), which does not define a significant PSC: a trial whose evoked current in the expected direction exceeds k s.d. of the cell's sham measure (k = 2 and 3). It keeps 1 of 39 cell recordings at k = 2 and none at k = 3; the other checks of the data check, which need no such definition, agree that the currents almost never follow the calls (two cells carry an inward current during the call).
    • Every listed cell is used for the computed values; 12 experienced and 14 naive EPSC cells as in the paper, and 7 experienced (paper 6) and 6 naive IPSC cells. One experienced IPSC cell (190410_e1) has 4 or 5 trials per ISI.
    • The largest 5 or 10 trials are ranked by absolute value (the Methods' "absolute magnitudes"), and, as a second reading, by the current in the expected direction.
    • The spreadsheet notes are applied (stop or drop at a sweep keeps the sweeps before it, from sweep N keeps the later ones, drop or block + stop drops the file), only the best call's set is used where a note names one, and sweeps whose holding current contradicts the listed potential are dropped (file 19424013 is listed at -70 mV but carries +560 pA). Using every sweep is a variant.
    • Experienced cells pool dams and EVs; the spreadsheets do not say which are which.
    • One listed file (193260007, an IPSC file of naive cell 190326_e1) is not on figshare; the cell's other IPSC file is used.
    • The Bonferroni family is read as the three comparisons with the prototype that the legend reports; all six pairs are shown.
  • BlockedFigure 2h; main text, paragraph citing Fig. 2h· Otherverifier disagreed at firstdata in figshare doi:10.6084/m9.figshare.12611657.v3 and doi:10.6084/m9.figshare.12613496.v2

    Prototype-evoked EPSCs are similar between groups, but EPSCs evoked by non-prototypical ISIs are larger in experienced than in naive females.

    While prototype-evoked EPSCs were similar, responses to non-prototypical ISIs were enhanced in experienced females

    Figure 2h; main text, paragraph citing Fig. 2h

    What the paper reports

    • 125 ms: unpaired one-tailed t-test, p = 0.001 (n = 12 EV and 14 NV cells)
    • 275 ms: unpaired one-tailed t-test, p = 0.03
    • 575 ms: unpaired one-tailed t-test, p = 0.04

    In the deposited in vivo recordings the current almost never follows the calls, so the claim cannot be tested on the cells it is about. The Methods' evoked measure is in the expected direction at p < 0.05 in 2 of 39 cell recordings, against 3 of 39 for the same measure before the call; no cell has a mean response 0-300 ms after onset at p < 0.05 (mean over cells 0.25 pA for EPSCs); 13 of 110 files have a sweep-locked deflection 0-3 s after onset at p < 0.05, more than the 5.5 expected by chance (binomial p = 0.0034), but only 3 of them, from two cells (naive 190326_e1, both EPSC files, and experienced 190313_e2, one of two), carry a current in the expected direction during the call, first exceeding the null 0.39 to 0.51 s after onset and peaking at 37 to 48 pA; the other 10 peak at scattered times, 9 against the expected direction; and the inclusion rule keeps 1 of 39 (2 s.d., cell 190326_e1) or none (3 s.d.). As computed it would not be reproduced: experienced EPSCs exceed naive ones at 125, 275 and 575 ms (one-tailed p 0.0023, 0.0062, 0.0047) but also at the prototype (33.5 against 20.0 pA, p = 0.013; 0.0051 to 0.041 under every variant). The same measure before the call shows the same difference (EV 32.0 to 35.7 against NV 14.8 to 19.6 pA, one-tailed p 0.0002 to 0.0047), and the baseline s.d. is 11.4 pA in experienced and 5.3 pA in naive recordings.

    How the test differed from the paper’s method

    • Call onset, offset and ISI come from the recorded microphone channel. The onset is the start of the sound (envelope above its median by 6 robust s.d.); before revision it was the first loud syllable, which came a median 193 ms (up to 644 ms) late because the first syllable of many calls is quiet, and that rule is run as a variant. In 82 of 83 morph-set files the four call durations form four groups in a fixed order (prototype, 125, 275, 575 ms), and the ISI is assigned by duration rank; one two-sweep file uses the nearest bin of the measured gap. The sets hold the legends' ISIs, not the Methods' 125, 225 and 575 ms.
    • Assumed, for the Methods' inclusion rule ("Only neurons that exhibited at least five significant prototype-evoked E/IPSCs"), which does not define a significant PSC: a trial whose evoked current in the expected direction exceeds k s.d. of the cell's sham measure (k = 2 and 3). It keeps 1 of 39 cell recordings at k = 2 and none at k = 3; the other checks of the data check, which need no such definition, agree that the currents almost never follow the calls (two cells carry an inward current during the call).
    • Every listed cell is used for the computed values; 12 experienced and 14 naive EPSC cells as in the paper, and 7 experienced (paper 6) and 6 naive IPSC cells. One experienced IPSC cell (190410_e1) has 4 or 5 trials per ISI.
    • The largest 5 or 10 trials are ranked by absolute value (the Methods' "absolute magnitudes"), and, as a second reading, by the current in the expected direction.
    • The spreadsheet notes are applied (stop or drop at a sweep keeps the sweeps before it, from sweep N keeps the later ones, drop or block + stop drops the file), only the best call's set is used where a note names one, and sweeps whose holding current contradicts the listed potential are dropped (file 19424013 is listed at -70 mV but carries +560 pA). Using every sweep is a variant.
    • Experienced cells pool dams and EVs; the spreadsheets do not say which are which.
    • One listed file (193260007, an IPSC file of naive cell 190326_e1) is not on figshare; the cell's other IPSC file is used.
  • BlockedFigure 2i; Extended Data Fig. 6i-k· Otherverifier disagreed at firstdata in figshare doi:10.6084/m9.figshare.12611657.v3 and doi:10.6084/m9.figshare.12613496.v2

    IPSCs are broadly tuned in both groups but weaker in naive virgins at every ISI.

    IPSCs were broadly tuned in both groups ( Extended Data Figure 6i – k ), albeit weaker in naive virgins

    Figure 2i; Extended Data Fig. 6i-k

    What the paper reports

    • 125 ms: unpaired two-tailed t-test, p = 0.047 (n = 6 cells each)
    • 175 ms: unpaired two-tailed t-test, p = 0.009
    • 275 ms: unpaired two-tailed t-test, p = 0.01
    • 575 ms: unpaired two-tailed t-test, p = 0.004
    • EV IPSC against ISI: two-tailed Pearson correlation, r = 0.33, p = 0.11 (n = 6 cells)
    • NV IPSC against ISI: two-tailed Pearson correlation, r = 0.13, p = 0.56 (n = 6 cells)

    In the deposited in vivo recordings the current almost never follows the calls, so the claim cannot be tested on the cells it is about. The Methods' evoked measure is in the expected direction at p < 0.05 in 2 of 39 cell recordings, against 3 of 39 for the same measure before the call; no cell has a mean response 0-300 ms after onset at p < 0.05 (mean over cells 0.25 pA for EPSCs); 13 of 110 files have a sweep-locked deflection 0-3 s after onset at p < 0.05, more than the 5.5 expected by chance (binomial p = 0.0034), but only 3 of them, from two cells (naive 190326_e1, both EPSC files, and experienced 190313_e2, one of two), carry a current in the expected direction during the call, first exceeding the null 0.39 to 0.51 s after onset and peaking at 37 to 48 pA; the other 10 peak at scattered times, 9 against the expected direction; and the inclusion rule keeps 1 of 39 (2 s.d., cell 190326_e1) or none (3 s.d.). As computed the readings disagree: experienced IPSCs exceed naive ones at 125, 275 and 575 ms under every variant (p 0.0018 to 0.024), and at 175 ms with the largest 10 trials (p 0.012 to 0.040) and the expected-direction ranking (p = 0.021) but not the largest 5 absolute values (p = 0.055). Neither correlation with ISI is significant (r -0.19 to 0.13, p >= 0.347). The measure before the call shows the same group difference (two-tailed p 0.0018 to 0.038).

    How the test differed from the paper’s method

    • Call onset, offset and ISI come from the recorded microphone channel. The onset is the start of the sound (envelope above its median by 6 robust s.d.); before revision it was the first loud syllable, which came a median 193 ms (up to 644 ms) late because the first syllable of many calls is quiet, and that rule is run as a variant. In 82 of 83 morph-set files the four call durations form four groups in a fixed order (prototype, 125, 275, 575 ms), and the ISI is assigned by duration rank; one two-sweep file uses the nearest bin of the measured gap. The sets hold the legends' ISIs, not the Methods' 125, 225 and 575 ms.
    • Assumed, for the Methods' inclusion rule ("Only neurons that exhibited at least five significant prototype-evoked E/IPSCs"), which does not define a significant PSC: a trial whose evoked current in the expected direction exceeds k s.d. of the cell's sham measure (k = 2 and 3). It keeps 1 of 39 cell recordings at k = 2 and none at k = 3; the other checks of the data check, which need no such definition, agree that the currents almost never follow the calls (two cells carry an inward current during the call).
    • Every listed cell is used for the computed values; 12 experienced and 14 naive EPSC cells as in the paper, and 7 experienced (paper 6) and 6 naive IPSC cells. One experienced IPSC cell (190410_e1) has 4 or 5 trials per ISI.
    • The largest 5 or 10 trials are ranked by absolute value (the Methods' "absolute magnitudes"), and, as a second reading, by the current in the expected direction.
    • The spreadsheet notes are applied (stop or drop at a sweep keeps the sweeps before it, from sweep N keeps the later ones, drop or block + stop drops the file), only the best call's set is used where a note names one, and sweeps whose holding current contradicts the listed potential are dropped (file 19424013 is listed at -70 mV but carries +560 pA). Using every sweep is a variant.
    • Experienced cells pool dams and EVs; the spreadsheets do not say which are which.
    • One listed file (193260007, an IPSC file of naive cell 190326_e1) is not on figshare; the cell's other IPSC file is used.
  • BlockedFigure 2j; main text, paragraph citing Fig. 2j· Otherverifier disagreed at firstdata in figshare doi:10.6084/m9.figshare.12611657.v3 and doi:10.6084/m9.figshare.12613496.v2

    In experienced females the IPSC-EPSC ratio is positively correlated with ISI, so that the balance shifts toward inhibition for slower ISIs; not so in naive females.

    In experienced females, this ratio was positively correlated with ISI duration (Pearson’s r=0.47, p=0.02

    Figure 2j; main text, paragraph citing Fig. 2j

    What the paper reports

    • experienced: two-tailed Pearson correlation, r = 0.47, p = 0.02 (n = 6 cells)
    • naive: two-tailed Pearson correlation, r = 0.01, p = 0.96 (n = 6 cells)

    In the deposited in vivo recordings the current almost never follows the calls, so the claim cannot be tested on the cells it is about. The Methods' evoked measure is in the expected direction at p < 0.05 in 2 of 39 cell recordings, against 3 of 39 for the same measure before the call; no cell has a mean response 0-300 ms after onset at p < 0.05 (mean over cells 0.25 pA for EPSCs); 13 of 110 files have a sweep-locked deflection 0-3 s after onset at p < 0.05, more than the 5.5 expected by chance (binomial p = 0.0034), but only 3 of them, from two cells (naive 190326_e1, both EPSC files, and experienced 190313_e2, one of two), carry a current in the expected direction during the call, first exceeding the null 0.39 to 0.51 s after onset and peaking at 37 to 48 pA; the other 10 peak at scattered times, 9 against the expected direction; and the inclusion rule keeps 1 of 39 (2 s.d., cell 190326_e1) or none (3 s.d.). As computed it would not be reproduced: in the 7 experienced cells with both currents the ratio is not correlated with ISI (r = -0.055, p = 0.78; -0.12 to 0.087 under every variant) against r = 0.47. The naive correlation is not significant (r = 0.13, p = 0.54).

    How the test differed from the paper’s method

    • Call onset, offset and ISI come from the recorded microphone channel. The onset is the start of the sound (envelope above its median by 6 robust s.d.); before revision it was the first loud syllable, which came a median 193 ms (up to 644 ms) late because the first syllable of many calls is quiet, and that rule is run as a variant. In 82 of 83 morph-set files the four call durations form four groups in a fixed order (prototype, 125, 275, 575 ms), and the ISI is assigned by duration rank; one two-sweep file uses the nearest bin of the measured gap. The sets hold the legends' ISIs, not the Methods' 125, 225 and 575 ms.
    • Assumed, for the Methods' inclusion rule ("Only neurons that exhibited at least five significant prototype-evoked E/IPSCs"), which does not define a significant PSC: a trial whose evoked current in the expected direction exceeds k s.d. of the cell's sham measure (k = 2 and 3). It keeps 1 of 39 cell recordings at k = 2 and none at k = 3; the other checks of the data check, which need no such definition, agree that the currents almost never follow the calls (two cells carry an inward current during the call).
    • Every listed cell is used for the computed values; 12 experienced and 14 naive EPSC cells as in the paper, and 7 experienced (paper 6) and 6 naive IPSC cells. One experienced IPSC cell (190410_e1) has 4 or 5 trials per ISI.
    • The largest 5 or 10 trials are ranked by absolute value (the Methods' "absolute magnitudes"), and, as a second reading, by the current in the expected direction.
    • The spreadsheet notes are applied (stop or drop at a sweep keeps the sweeps before it, from sweep N keeps the later ones, drop or block + stop drops the file), only the best call's set is used where a note names one, and sweeps whose holding current contradicts the listed potential are dropped (file 19424013 is listed at -70 mV but carries +560 pA). Using every sweep is a variant.
    • Experienced cells pool dams and EVs; the spreadsheets do not say which are which.
    • One listed file (193260007, an IPSC file of naive cell 190326_e1) is not on figshare; the cell's other IPSC file is used.
  • BlockedFigure 2k; Extended Data Fig. 6l· Otherverifier disagreed at firstdata in figshare doi:10.6084/m9.figshare.12611657.v3

    Within cells of experienced cortex, 575 ms-evoked IPSCs are stronger than EPSCs, while the two do not differ at 125, 175 or 275 ms.

    575-evoked IPSCs were significantly stronger than EPSCs (p=0.03

    Figure 2k; Extended Data Fig. 6l

    What the paper reports

    • 575 ms: paired two-tailed t-test, p = 0.03 (n = 6 cells)
    • 125 ms: two-tailed paired t-test, p = 0.47
    • 175 ms: two-tailed paired t-test, p = 0.26
    • 275 ms: two-tailed paired t-test, p = 0.44

    In the deposited in vivo recordings the current almost never follows the calls, so the claim cannot be tested on the cells it is about. The Methods' evoked measure is in the expected direction at p < 0.05 in 2 of 39 cell recordings, against 3 of 39 for the same measure before the call; no cell has a mean response 0-300 ms after onset at p < 0.05 (mean over cells 0.25 pA for EPSCs); 13 of 110 files have a sweep-locked deflection 0-3 s after onset at p < 0.05, more than the 5.5 expected by chance (binomial p = 0.0034), but only 3 of them, from two cells (naive 190326_e1, both EPSC files, and experienced 190313_e2, one of two), carry a current in the expected direction during the call, first exceeding the null 0.39 to 0.51 s after onset and peaking at 37 to 48 pA; the other 10 peak at scattered times, 9 against the expected direction; and the inclusion rule keeps 1 of 39 (2 s.d., cell 190326_e1) or none (3 s.d.). As computed it would not be reproduced: at 575 ms IPSCs exceed EPSCs on average (38.5 against 32.4 pA, 4 of 7 cells) but not significantly (paired p = 0.40; 0.17 to 0.99 under every variant); 125, 175 and 275 ms do not differ (p 0.17 to 0.96).

    How the test differed from the paper’s method

    • Call onset, offset and ISI come from the recorded microphone channel. The onset is the start of the sound (envelope above its median by 6 robust s.d.); before revision it was the first loud syllable, which came a median 193 ms (up to 644 ms) late because the first syllable of many calls is quiet, and that rule is run as a variant. In 82 of 83 morph-set files the four call durations form four groups in a fixed order (prototype, 125, 275, 575 ms), and the ISI is assigned by duration rank; one two-sweep file uses the nearest bin of the measured gap. The sets hold the legends' ISIs, not the Methods' 125, 225 and 575 ms.
    • Assumed, for the Methods' inclusion rule ("Only neurons that exhibited at least five significant prototype-evoked E/IPSCs"), which does not define a significant PSC: a trial whose evoked current in the expected direction exceeds k s.d. of the cell's sham measure (k = 2 and 3). It keeps 1 of 39 cell recordings at k = 2 and none at k = 3; the other checks of the data check, which need no such definition, agree that the currents almost never follow the calls (two cells carry an inward current during the call).
    • Every listed cell is used for the computed values; 12 experienced and 14 naive EPSC cells as in the paper, and 7 experienced (paper 6) and 6 naive IPSC cells. One experienced IPSC cell (190410_e1) has 4 or 5 trials per ISI.
    • The largest 5 or 10 trials are ranked by absolute value (the Methods' "absolute magnitudes"), and, as a second reading, by the current in the expected direction.
    • The spreadsheet notes are applied (stop or drop at a sweep keeps the sweeps before it, from sweep N keeps the later ones, drop or block + stop drops the file), only the best call's set is used where a note names one, and sweeps whose holding current contradicts the listed potential are dropped (file 19424013 is listed at -70 mV but carries +560 pA). Using every sweep is a variant.
    • Experienced cells pool dams and EVs; the spreadsheets do not say which are which.
    • One listed file (193260007, an IPSC file of naive cell 190326_e1) is not on figshare; the cell's other IPSC file is used.
  • BlockedExtended Data Fig. 6g,h· Metadata insufficientdata in figshare doi:10.6084/m9.figshare.12611657.v3

    Prototype- and morph-evoked EPSCs do not differ between lactating dams and experienced virgins, justifying their pooling.

    Prototype and morph-evoked EPSCs did not differ between dams (n=4 cells) and EVs (n=8)

    Extended Data Fig. 6g,h

    What the paper reports

    • 125 ms: unpaired two-tailed t-test, p = 0.11
    • 175 ms: unpaired two-tailed t-test, p = 0.8
    • 275 ms: unpaired two-tailed t-test, p = 0.85
    • 575 ms: unpaired two-tailed t-test, p = 0.85

    The experienced in vivo whole-cell article (figshare 12611657 v3) lists 12 cells by date and number, with no field saying which 4 come from dams and which 8 from experienced virgins, and the ABF files carry no subject information.

Not tested in a notebook 32

12 data missing from the dataset, 20 metadata insufficient

Blocked32

  • BlockedFigure 1a; Methods, Stimulus library and vocalization analysis· Data missing from the dataset

    Pup distress call ISIs (5,355 intervals from 37 recordings of pups at postnatal days 1-8) follow a lognormal distribution; calls with ISIs around the median (175 +/- 25 ms) serve as prototypes, and ISIs of 575 ms or more essentially never occur.

    5,355 inter-syllable intervals were measured from 37 audio recordings ... and a lognormal frequency distribution was generated

    Figure 1a; Methods, Stimulus library and vocalization analysis

    The 37 pup vocalization recordings and the 5,355 measured ISIs are not in the dandiset or in the figshare collection the dandiset links; the paper's figshare DOI does not resolve.

  • BlockedFigure 1c; main text, paragraph citing Fig. 1c· Data missing from the dataset

    Experienced females (dams and experienced virgins pooled) retrieve cold, silent pups on fewer trials than warm, vocalizing pups (32.0% against 80.0%).

    Experienced females retrieved cold, silent pups on fewer trials than warm, vocalizing pups (retrieval (cold) : 32.0% of trials, retrieval (warm) : 80.0%

    Figure 1c; main text, paragraph citing Fig. 1c

    What the paper reports

    • warm against cold, single syllable and 575 ms: two-tailed Fisher's test, FDR correction, p = 3 × 10⁻⁴, FDR (Benjamini-Hochberg)-corrected (n = warm 55 trials, cold 25, single 20, 575 ms 22)

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedFigure 1c; main text, paragraph citing Fig. 1c· Data missing from the dataset

    Cold pups dubbed over with prototypes (76.7%) or with morphs of ISIs 75 to 375 ms (73.7-81.0%) are retrieved at rates similar to warm pups.

    Cold pups dubbed over with prototypes (76.7%), as well as morphs with ISIs between 75:375 ms (73.7–81.0%), were retrieved at similar rates to warm pups

    Figure 1c; main text, paragraph citing Fig. 1c

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedFigure 1c; main text, paragraph citing Fig. 1c· Data missing from the dataset

    ISIs slower than 575 ms (27.3%) and single syllables (15.0%) do not effectively elicit retrieval, being retrieved less than warm pups.

    In contrast, ISIs slower than 575 ms (27.3%) and single syllables (15.0%) did not effectively elicit retrieval

    Figure 1c; main text, paragraph citing Fig. 1c

    What the paper reports

    • warm against cold, single syllable and 575 ms: two-tailed Fisher's test, FDR correction, p = 3 × 10⁻⁴, FDR (Benjamini-Hochberg)-corrected (n = warm 55 trials, single 20, 575 ms 22)

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedExtended Data Fig. 2a· Data missing from the dataset

    Dams and experienced virgins do not differ in retrieval of warm pups, prototype-dubbed pups or cold pups.

    Dams and EVs did not differ in their retrieval behavior

    Extended Data Fig. 2a

    What the paper reports

    • warm, dams 85.2% (27 trials) against EVs 77.0% (26): two-tailed Fisher's test, p = 0.5
    • prototypes, dams 91.7% (12) against EVs 66.7% (18): two-tailed Fisher's test, p = 0.19
    • cold, dams 25.0% (12) against EVs 38.5% (13): two-tailed Fisher's test, p = 0.67

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedExtended Data Fig. 2b· Data missing from the dataset

    On successful trials, latencies to retrieve cold pups and pups dubbed with slow stimuli (single syllables and 575 ms) are longer than for warm pups and for pups dubbed with 75 to 375 ms ISIs.

    Latencies to retrieve cold pups and cold pups dubbed over with slow morphs (single syllables (‘SS’) and 575 ms ISIs) were longer than for warm pups (p=0.02) or cold pups dubbed over with morphs containing ISIs between 75:375 ms (p=0.001)

    Extended Data Fig. 2b

    What the paper reports

    • slow and cold against warm: Kruskal-Wallis H-test with Dunn's correction, p = 0.02
    • slow and cold against 75:375 ms morphs: Kruskal-Wallis H-test with Dunn's correction, p = 0.001

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedFigure 4c; main text, paragraph citing Fig. 4c· Data missing from the dataset

    Virgins exposed to slow morphs during cohousing retrieve pups dubbed with slow morphs more reliably than standard cohoused virgins (75% against 14.3% of trials).

    virgins in the CH+575 group retrieved cold pups dubbed over with slow morphs (ISI:575 ms; 75% of trials) more reliably than controls (14.3%

    Figure 4c; main text, paragraph citing Fig. 4c

    What the paper reports

    • two-tailed Fisher's test, p = 0.001 (n = CH 14-26 trials, CH+575 12-16 trials)

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedFigure 4d; Extended Data Fig. 8g· Data missing from the dataset

    Photoinhibiting left auditory cortex only during slow morph playback blocks the learning of slow ISIs (18.8% against 66.7% in shams) without affecting retrieval of prototypes or latencies.

    Photoinhibiting left auditory cortex (ACtx) exclusively during slow morph playback selectively blocked the learning of slow ISIs compared to shams

    Figure 4d; Extended Data Fig. 8g

    What the paper reports

    • slow morphs, sham against opto-ACtx: two-tailed Fisher's test, p = 0.02 (n = sham 3 mice 12 trials, opto 4 mice 16 trials)
    • latencies: unpaired two-tailed Mann-Whitney test, p = 0.19 (n = 26 and 23 trials)

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedExtended Data Fig. 8h,i; main text, paragraph citing it· Data missing from the dataset

    Virgins hearing single syllables during cohousing do not learn to retrieve pups dubbed with single syllables.

    virgins hearing single syllables (ISI:30s) during cohousing were unable to associate these calls with isolated pups

    Extended Data Fig. 8h,i; main text, paragraph citing it

    What the paper reports

    • single syllable, CH 22.2% (9 trials) against CH+SS 25.0% (16): two-tailed Fisher's exact test, p > 0.99
    • warm, CH 84.6% (13) against CH+SS 85.7% (14): two-tailed Fisher's exact test, p > 0.99
    • latencies: unpaired two-tailed Mann-Whitney test, p = 0.98 (n = 17 trials each)

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedFigure 4e; Extended Data Fig. 9d· Data missing from the dataset

    Inhibiting oxytocin neurons during slow morph playback lowers retrieval of warm pups (43.8% against 92.3%), prototype-dubbed pups (36.8% against 83.3%) and slow-morph-dubbed pups (33.3% against 83.3%), and changes retrieval latencies.

    we observed deficits in retrieval rates for warm pups (CH+575 sham : 92.3% vs. CH+575 opto-OT : 43.8% of trials), as well as cold pups dubbed over with prototypes (83.3% vs. 36.8%) and slow morphs (83.3% vs. 33.3%

    Figure 4e; Extended Data Fig. 9d

    What the paper reports

    • warm: two-tailed Fisher's test, p = 0.008 (n = sham 3 mice 12-13 trials, opto-OT 5 mice 16-19 trials)
    • prototypes: two-tailed Fisher's test, p = 0.02
    • slow: two-tailed Fisher's test, p = 0.01
    • latencies: unpaired two-tailed Mann-Whitney test, p = 3 × 10⁻⁴ (n = 32 and 20 trials)

    The retrieval assays (warm, cold and dubbed pups, latencies) are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links (doi:10.6084/m9.figshare.c.5043830.v6), whose behaviour item holds only the operant touch times. The paper's own figshare DOI (doi:10.35092/yhjc.c.5037737) does not resolve, and the source data sheet was not found.

  • BlockedExtended Data Fig. 2d; main text, paragraph citing Extended Data Fig. 2c,d· Data missing from the dataset

    In a y-maze, experienced females show no preference between a prototype and a morph with ISIs 25 to 350 ms, but approach prototypes more than morphs slower than 350 ms (425 ms 25.0%, 575+ ms 15.8% of trials approaching the morph).

    Mice showed a significant preference to approach prototypes when ISIs were slower than 350 ms

    Extended Data Fig. 2d; main text, paragraph citing Extended Data Fig. 2c,d

    What the paper reports

    • 62.5 ms bin, 44.2%, 52 trials: two-tailed binomial test against 0.50, p = 0.49
    • 275 ms bin, 55.6%, 45 trials: two-tailed binomial test against 0.50, p = 0.46
    • 425 ms bin, 25.0%, 20 trials: two-tailed binomial test against 0.50, p = 0.04
    • 575+ ms bin, 15.8%, 19 trials: two-tailed binomial test against 0.50, p = 0.005

    The y-maze trials are not in the dandiset, which holds only two-photon movies, nor in the figshare collection the dandiset links. The paper's figshare DOI does not resolve.

  • BlockedExtended Data Fig. 3e; main text, paragraph citing it· Metadata insufficient

    About 16% of excitatory neurons in left auditory cortex respond to prototypes, regardless of maternal experience.

    We observed robust responses to prototypes in ~16% of excitatory neurons regardless of maternal experience

    Extended Data Fig. 3e; main text, paragraph citing it

    What the paper reports

    • two-tailed unpaired t-test, p = 0.74 (n = 9 experienced and 12 naive virgins)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedFigure 1f; main text, paragraph citing Fig. 1e,f· Metadata insufficient

    Excitatory neuronal tuning to temporal morphs is broader in experienced than in naive virgins across the ISIs that elicit retrieval; naive cortex responds less to morphs than to prototypes.

    Compared to naive virgins, temporal tuning was broader in experienced virgins across ISIs that reliably elicited pup retrieval

    Figure 1f; main text, paragraph citing Fig. 1e,f

    What the paper reports

    • EV against NV at 375 ms: one-way ANOVA, Bonferroni correction, p = 0.05, Bonferroni-corrected (n = 9 experienced and 12 naive virgins)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedExtended Data Fig. 6c· Metadata insufficient

    Tuning width over 75 to 375 ms is broader for excitatory neurons in experienced than in naive virgins; in naive cortex interneurons are more broadly tuned than excitatory neurons; and interneuron tuning width does not differ between experienced and naive virgins.

    Excitatory neuronal tuning curves were significantly broader in EVs (N=9 mice) than in NVs (N=12; p=0.001), and interneuron tuning was broader than excitatory tuning in naive cortex (Inh: N=6; p=0.04). Interneuron tuning width did not differ between experienced (N=5) and naive virgins (p>0.99; one-way ANOVA, Bonferroni correction)

    Extended Data Fig. 6c

    What the paper reports

    • EV against NV excitatory: one-way ANOVA, Bonferroni correction, p = 0.001, Bonferroni-corrected (n = 9 EV and 12 NV mice)
    • NV interneurons against NV excitatory: one-way ANOVA, Bonferroni correction, p = 0.04, Bonferroni-corrected (n = 6 and 12 mice)
    • EV against NV interneurons: one-way ANOVA, Bonferroni correction, p > 0.99, Bonferroni-corrected (n = 5 and 6 mice)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedExtended Data Fig. 3k; main text, paragraph citing Extended Data Fig. 3i-m· Metadata insufficient

    The narrow tuning in naive cortex does not extend to temporally modulated pure tones; tuning width to tone sequences does not differ between experienced and naive virgins.

    the narrow tuning in naive cortex did not extend to temporally-modulated pure tones

    Extended Data Fig. 3k; main text, paragraph citing Extended Data Fig. 3i-m

    What the paper reports

    • unpaired two-tailed t-test, p = 0.97 (n = experienced 3 mice, 94 neurons; naive 4 mice, 45 neurons)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions. The tone sequence files (names with kmtf) give the tone frequency and the ISIs but not their order either.

  • BlockedExtended Data Fig. 3m· Metadata insufficient

    In naive auditory cortex, normalized responses to temporally modulated tones are higher than to pup call morphs at 75, 375 and 575 ms.

    We observed higher normalized ΔF/Fs to temporally-modulated tones (N=5 mice) than to pup call morphs (n=11–12) in naive auditory cortex

    Extended Data Fig. 3m

    What the paper reports

    • 75 ms: unpaired two-tailed Mann-Whitney test, p = 0.009
    • 375 ms: unpaired two-tailed Mann-Whitney test, p = 0.04
    • 575 ms: unpaired two-tailed Mann-Whitney test, p = 0.004

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedExtended Data Fig. 4a-c· Metadata insufficient

    Broad temporal tuning in experienced virgins is left-lateralized; tuning width (75 to 375 ms) is larger in left than in right auditory cortex.

    broad temporal tuning was left-lateralized in experienced females

    Extended Data Fig. 4a-c

    What the paper reports

    • unpaired two-tailed Mann-Whitney test, p = 0.003 (n = 9 left and 4 right experienced virgins)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedExtended Data Fig. 4e; main text, paragraph citing Extended Data Fig. 4d,e· Metadata insufficient

    Qualitatively, the left auditory cortex of a lactating dam shows broad temporal tuning, similar to that of experienced virgins.

    Qualitatively, we observed broad temporal tuning in the left auditory cortex of a lactating dam (N=1 dam, n=10 single-cell tuning curves)

    Extended Data Fig. 4e; main text, paragraph citing Extended Data Fig. 4d,e

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedExtended Data Fig. 4f,g· Metadata insufficient

    In virgins retrieving on 10-30% of trials, temporal tuning was sharper before the onset of reliable retrieval; more neurons are broadly tuned once they retrieve on 100% of trials.

    A larger proportion of neurons were more broadly tuned (higher normalized ΔF/Fs) when EVs retrieved on 100% of trials (p<0.0001; two-tailed unpaired Kolmogorov-Smirnov test)

    Extended Data Fig. 4f,g

    What the paper reports

    • two-tailed unpaired Kolmogorov-Smirnov test, p < 10⁻⁴ (n = 2 mice, 78 neurons before and 52 after)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedFigure 2c,d; main text, paragraph citing Fig. 2· Metadata insufficient

    Interneurons are broadly tuned regardless of experience, but their tuning curve slopes at the transition between retrieved (75:375 ms) and ignored (575:975 ms) ISIs are sharper in experienced virgins.

    tuning curve slopes at the transition between ISIs eliciting retrieval (75:375 ms, Fig. 1c ) and ignored ISIs (575:975 ms) were significantly sharper in experienced virgins (p=0.03

    Figure 2c,d; main text, paragraph citing Fig. 2

    What the paper reports

    • unpaired two-tailed t-test, p = 0.03 (n = 5 experienced and 6 naive virgins)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedExtended Data Fig. 6d; main text, paragraph citing Fig. 2e· Metadata insufficient

    Population tuning slopes at the behavioural transition (375:975 ms) are significantly negative in experienced virgins, for excitatory neurons and interneurons, and not in naive virgins.

    Slopes in EVs were significantly negative

    Extended Data Fig. 6d; main text, paragraph citing Fig. 2e

    What the paper reports

    • EV excitatory: one-sample t-test to 0.0, p = 0.02 (n = 7 mice)
    • EV interneurons: one-sample t-test to 0.0, p = 0.007 (n = 5 mice)
    • NV excitatory: one-sample t-test to 0.0, p = 0.84 (n = 11 mice)
    • NV interneurons: one-sample t-test to 0.0, p = 0.29 (n = 6 mice)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedFigure 2e· Metadata insufficient

    In experienced virgins, excitatory and inhibitory single-cell tuning are matched across behaviourally salient ISIs and differ at 575 and 975 ms.

    Excitatory and inhibitory neuronal tuning were also matched in experienced virgins, responding similarly across behaviorally-salient ISIs and with negative slopes between 375:975 ms

    Figure 2e

    What the paper reports

    • excitatory against inhibitory at 575 ms: one-way ANOVA, Bonferroni correction, p = 0.009, Bonferroni-corrected (n = 366 excitatory and 260 inhibitory single-cell tuning curves)
    • excitatory against inhibitory at 975 ms: one-way ANOVA, Bonferroni correction, p = 0.048, Bonferroni-corrected (n = 366 excitatory and 260 inhibitory single-cell tuning curves)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedFigure 2f; main text, paragraph citing Fig. 2f· Metadata insufficient

    In naive cortex, interneurons are more broadly tuned than excitatory neurons but lack a sharp boundary at the behavioural transition.

    while interneurons in naive cortex were more broadly tuned than excitatory neurons, the lack of a sharp boundary suggests interneurons were still relatively untuned to relevant statistics

    Figure 2f; main text, paragraph citing Fig. 2f

    What the paper reports

    • NV interneuron against NV excitatory tuning width (Extended Data Fig. 6c): one-way ANOVA, Bonferroni correction, p = 0.04, Bonferroni-corrected (n = 6 and 12 mice)
    • NV interneuron population slope against zero (Extended Data Fig. 6d): one-sample t-test to 0.0, p = 0.29 (n = 6 mice)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions. It also depends on ed6c-tuning-width-groups and ed6d-slopes-negative-in-ev, blocked for the same reason.

  • BlockedExtended Data Fig. 7a-c; main text, paragraph citing Fig. 3b· Metadata insufficient

    During cohousing prototype-responsive interneurons are relatively stable while the set of prototype-responsive excitatory neurons changes; about 18.8% of excitatory and 72.3% of inhibitory neurons responsive at baseline remain responsive in the final session.

    Of the neurons that were prototype-responsive at baseline and successfully tracked, ~18.8% of excitatory neurons and ~72.3% of inhibitory neurons remained responsive in the final imaging session

    Extended Data Fig. 7a-c; main text, paragraph citing Fig. 3b

    What the paper reports

    • excitatory, baseline against post-onset prototype dF/F: two-tailed Spearman correlation, r = -0.40, p < 10⁻⁴ (n = 104 neurons)
    • inhibitory: two-tailed Spearman correlation, r = 0.46, p < 10⁻⁴ (n = 122 neurons)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions. Tracking cells across days also needs the ROIs, which are not deposited.

  • BlockedFigure 3e-g; main text, paragraph citing Fig. 3e· Metadata insufficient

    During cohousing, excitatory single-cell tuning broadens between 75 and 375 ms after retrieval onset, while interneuron tuning does not broaden; excitatory and inhibitory tuning, mismatched at first, become similar 24 hours after first retrieval.

    Excitatory population tuning broadened between 75:375 ms ( Fig. 3e , f ), elaborating on the initial preference for prototypes. Interneurons did not broaden their tuning

    Figure 3e-g; main text, paragraph citing Fig. 3e

    What the paper reports

    • excitatory widths before against after onset: unpaired two-tailed Kolmogorov-Smirnov test, p < 10⁻⁴ (n = 165 naive and 70 (+24 h) excitatory curves)
    • interneuron widths before against after onset: unpaired two-tailed Kolmogorov-Smirnov test, p = 0.16 (n = 128 naive and 86 (+24 h) interneuron curves)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedFigure 3h· Metadata insufficient

    Both excitatory and inhibitory population tuning sharpen at the transition 375 to 975 ms after virgins start retrieving.

    both excitatory and inhibitory tuning selectively sharpened between 375:975 ms

    Figure 3h

    What the paper reports

    • excitatory: one-tailed paired t-test, p = 0.02 (n = 6 mice)
    • inhibitory: one-tailed paired t-test, p = 0.02 (n = 6 mice)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedExtended Data Fig. 8c,d· Metadata insufficient

    Pup-naive virgins imaged on three consecutive days without cohousing show no systematic change in temporal tuning.

    non-cohoused controls only exposed to calls during consecutive imaging sessions showed no systematic tuning changes

    Extended Data Fig. 8c,d

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedFigure 4b; main text, paragraph citing Fig. 4b· Metadata insufficient

    Virgins that heard slow morphs (575 ms) during cohousing have excitatory tuning broader at 575 ms than standard cohoused virgins.

    prototype-responsive neurons in virgins that heard slow exemplars were more broadly tuned to 575 ms ISIs

    Figure 4b; main text, paragraph citing Fig. 4b

    What the paper reports

    • one-way ANOVA, Bonferroni correction, p = 0.007, Bonferroni-corrected (n = 6 CH+575 and 5 CH virgins)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedFigure 4f,g; Extended Data Fig. 9e· Metadata insufficient

    Virgins with oxytocin neurons inhibited during cohousing have uncharacteristically narrow tuning 24 hours after first retrieval, narrower than cohoused virgins' excitatory and inhibitory tuning.

    These behavioral deficits were associated with uncharacteristically narrow temporal tuning in auditory cortex 24 hours after first retrieval

    Figure 4f,g; Extended Data Fig. 9e

    What the paper reports

    • opto-OT against CH excitatory: Kruskal-Wallis test with Dunn's correction, p = 0.008 (n = 4 opto-OT and 14 CH mice)
    • opto-OT against CH inhibitory: Kruskal-Wallis test with Dunn's correction, p = 0.03 (n = 4 opto-OT and 10 CH mice)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions.

  • BlockedFigure 4h; main text, paragraph citing Fig. 4h· Metadata insufficient

    Single-cell tuning in opto-OT virgins is narrower than in reliably retrieving virgins but similar to virgins retrieving on 10-30% of trials.

    Single-cell tuning curves in CH+575 opto-OT virgins were significantly narrower compared to reliably retrieving virgins

    Figure 4h; main text, paragraph citing Fig. 4h

    What the paper reports

    • against CH 100%: Kruskal-Wallis test with Dunn's correction, p < 10⁻⁴ (n = 40 opto-OT and 52 CH 100% curves)
    • against CH 10-30%: Kruskal-Wallis test with Dunn's correction, p > 0.99 (n = 40 opto-OT and 78 CH 10-30% curves)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions. It also depends on ed4g-broadening-with-reliable-retrieval, blocked for the same reason.

  • BlockedExtended Data Fig. 3a-c; Methods, Image processing and analysis· Metadata insufficient

    Neuropil correction has no significant effect on prototype-evoked dF/F in excitatory or inhibitory neurons.

    Neuropil correction had no significant effect on call-evoked ΔF/Fs in excitatory or inhibitory neurons

    Extended Data Fig. 3a-c; Methods, Image processing and analysis

    What the paper reports

    • excitatory: two-tailed unpaired Kolmogorov-Smirnov test, p = 0.57 (n = 52 neurons, one region)
    • inhibitory: two-tailed unpaired Kolmogorov-Smirnov test, p = 0.14 (n = 64 neurons, one region)

    The dandiset's NWB files hold only raw, unaligned two-photon movies. They have no ROIs or fluorescence traces (the paper drew ROIs by hand), no stimulus series, no trial table and no times of the calls; the session description names the three to five calls or morphs of a recording but not the order in which they were played (pseudorandom, Methods). Folding the movies recovers the trial period (40 frames, 10.24 s), and in 31 of 38 three-call movies a cycle of three trials, but nothing says which call is played at which position of the cycle (data check). Responses cannot be assigned to named calls, so the claim cannot be tested. The figshare imaging items hold the same movies as TIFF files with the same descriptions. The neuropil and blood vessel regions and the two corrected data sets are not identified either.

  • BlockedExtended Data Fig. 4d· Data missing from the datasetuses data from outside the dataset

    In experienced virgins, call-evoked dF/F across ISI bins is correlated with the probability of each ISI in natural pup calls.

    In EVs (N=9 mice), evoked ΔF/Fs were correlated with ISI probability from the distribution in Fig. 1a (Pearson’s r=0.86, p=0.007; two-tailed)

    Extended Data Fig. 4d

    What the paper reports

    • two-tailed Pearson correlation, r = 0.86, p = 0.007 (n = 9 experienced virgins; 8 ISI bins)

    It needs the ISI distribution of Fig. 1a (fig1a-isi-distribution, blocked as data missing) and call-evoked dF/F per ISI, which the movies cannot give without stimulus times and order (metadata insufficient, as for the other imaging claims).

Findings about the data

What attempts found about the dataset itself, for its owner and the archive.

  • Data missingThe NWB files hold only raw, unaligned two-photon movies. There are no ROIs or fluorescence traces, no stimulus times, no trial tables and no record of which call each trial played (pseudorandom order, Methods), so none of the paper's 20 imaging findings can be tested from the dandiset. The figshare imaging items hold the same movies.
  • Data missingThe behavioural data of Figs. 1c, 4c-e and Extended Data Figs. 2 and 8-9 (pup retrieval, latencies, y-maze) and the vocalization recordings of Fig. 1a are in neither the dandiset nor the figshare collection it links. The paper's figshare DOI 10.35092/yhjc.c.5037737 does not resolve.
  • Description missingGroups (experienced or naive, dam, CH+575, opto-OT, cell type, hemisphere) are given only in free-text experiment descriptions; the indicator and location fields are "unknown" in every file, and the start of cohousing, needed for retrieval onset, is not recorded.
  • Metadata errorThree recordings (sessions of 2018-03-08 named plus24hrs_postexp1) are deposited under both sub-NV58 and sub-NV63, with the same start times and frames; nothing says to which mouse they belong.
  • Metadata errorSubjects EV56 and EV62 are described as NV56 and NV62 in their own experiment descriptions, and the five tones-only mice carry no experience group in theirs; the figshare item titles give it (EV56_Experienced_Exc and so on).
  • Description missingThe in vivo whole-cell spreadsheet of experienced females does not say which cells come from dams, so the dams-against-EVs comparison of Extended Data Fig. 6g,h cannot be made.
  • Metadata errorIn the in vivo whole-cell article of experienced females, file 19424013 is listed at -70 mV but carries about +560 pA throughout, as at a depolarized holding potential.
  • Data missingFiles listed in the figshare spreadsheets but not deposited, 193260007 (naive in vivo, listed number with an extra digit) and 19o11005 (slice, Cell 20 at 575 ms).
  • File format errorThe operant touch-time csv files print times with five significant digits, so after 1000 s of a session they step by 0.1 s and 99% of samples repeat the previous value; press durations can be resolved only to that step.
  • Paper inconsistentThe Methods give the in vivo morph set as ISIs of 125, 225 and 575 ms; the microphone channel of the recordings shows 125, 275 and 575 ms with the prototype, as the figure legends say, played in a fixed order, not pseudorandom.
  • OtherIn the in vivo whole-cell recordings the current almost never follows the played calls: the evoked measure is in the expected direction at p < 0.05 in 2 of 39 cell recordings (3 of 39 before the call), no cell responds 0-300 ms after onset, 13 of 110 files have a sweep-locked deflection (5.5 expected by chance) of which 3, from two cells (190326_e1 and 190313_e2), carry an inward current during the call from about 0.4 s after onset, and the Methods' inclusion rule keeps at most 1 of 39. The experienced recordings are about twice as noisy as the naive ones (baseline s.d. 11.4 against 5.3 pA), and the largest-trial measure gives the same experienced-naive difference before the call as during it.
  • Methods not availableFor the paper's authors (fig1i-operant-prototypes, ed5b-press-count-unchanged), which gap between touch samples ends a lever press, and whether session 1 was compared with session 8 or the first with the last sessions. The readings give different outcomes for the fast and slow groups.
  • Methods not availableFor the paper's authors (the in vivo claims and ed9b-oxytocin-enhances-spiking), how a significant PSC was defined for the inclusion rule, how many of the "largest 5-10 evoked trials" were averaged per cell and whether they were ranked by absolute value, and the Dunn comparison family of Extended Data Fig. 9b.
  • OtherThe imaging Methods say the calls were played in a pseudorandom order at 0.1 Hz. The movies recur every 40 frames (10.24 s, 0.098 Hz) and the tone files every 20 frames (5.12 s); in 31 of 38 three-call movies (one per subject) folding at three trial periods stands above two and four, as a fixed cycle of three calls would. No file says which call is played at which position of the cycle.
  • OtherThe first syllable of many played calls is much quieter than the rest on the microphone channel of the in vivo recordings; an onset taken at the first loud syllable comes a median 193 ms late.

Notebooks

Every registered entry that uses this dataset. An entry can hold several notebooks, and can use more than one dataset.

Reproduction Maternal behaviour and temporal tuning in auditory cortex (Schiavo et al. 2020)

Tests the findings of Schiavo, Valtcheva, Bair-Marshall, Song, Martin and Froemke (2020) on dandiset 000249, which holds the raw two-photon movies of 54 mice, and on the figshare collection the dandiset links, which holds the in vivo and slice whole-cell recordings and the operant touch times. The movies carry no ROIs, stimulus times or stimulus order, so the imaging findings cannot be tested; the whole-cell, slice and operant findings are tested on the figshare recordings. All files

  • Reproducing Schiavo et al. (2020) from dandiset 000249 and its figshare recordings
    reproduce_schiavo_2020.ipynb    Executed, with outputs · Source

Verification

  • Round 1
    2026-10-29    Agreed with 39 of 44 outcomes. Independent agent run by CatalystNeuro (same harness, separate context, did not write the notebook). Reviewed: The handover of the first run (the executed notebook with the conclusions and summary removed, the helper module and the claims file), with the paper and the archives.
    • ed6ef-epsc-tuning: the verifier’s outcome was “Blocked”. As computed it would be not reproduced, but the notebook's own control shows the measure picks up no call-evoked current: only 1 of 12 EV and 2 of 14 NV EPSC cells and 1 of 7 and 0 of 6 IPSC cells show a call-evoked current of the expected sign, about the rate the sham window gives, 46-48% of trials have the expected sign, and the Methods say every analysed cell had at least five significant prototype-evoked PSCs. The notebook never shows a trial-averaged current aligned to the microphone onset. EV has 7 IPSC cells against 6, and the inclusion rule is not applied.
    • fig2h-epsc-enhanced-ev: the verifier’s outcome was “Blocked”. As computed it would be not reproduced, because the prototype comparison is significant in every variant. The pre-call sham measure shows the same EV>NV differences, so what is being measured is recording noise, not evoked EPSCs. Trials are ranked by absolute value, so wrong-sign noise trials get selected; the largest trials in the expected direction were not run, and with EV baseline noise about twice NV's the top-k magnitudes favour EV whatever the response.
    • fig2i-ipsc-weaker-nv: the verifier’s outcome was “Blocked”. The 5- and 10-trial readings disagree at 175 ms. In any case the sham IPSC shows the same EV>NV difference, and EV has 7 IPSC cells against the paper's 6.
    • fig2j-ie-ratio-tracks-isi: the verifier’s outcome was “Blocked”. As computed it would be not reproduced (EV r about 0.06-0.09, p about 0.7, against 0.47). It rests on the same unvalidated measure, with 7 cells, not 6.
    • fig2k-ipsc-exceeds-epsc-575: the verifier’s outcome was “Blocked”. As computed it would be not reproduced: the paired p at 575 ms is 0.17-0.48. Same caveat as ed6ef.
  • Round 2
    2026-10-29    Agreed with 44 of 44 outcomes. Independent agent run by CatalystNeuro (same harness, separate context, did not write the notebook). Reviewed: The full executed notebook of the first revision, with its conclusions and summary, the helper module, the claims file and the attempt record, with the paper and the archives.

Method differences

How the tests differed from the paper’s method or data, and the claims each applies to.

  • Call onset, offset and ISI come from the recorded microphone channel. The onset is the start of the sound (envelope above its median by 6 robust s.d.); before revision it was the first loud syllable, which came a median 193 ms (up to 644 ms) late because the first syllable of many calls is quiet, and that rule is run as a variant. In 82 of 83 morph-set files the four call durations form four groups in a fixed order (prototype, 125, 275, 575 ms), and the ISI is assigned by duration rank; one two-sweep file uses the nearest bin of the measured gap. The sets hold the legends' ISIs, not the Methods' 125, 225 and 575 ms.ed6ef-epsc-tuning fig2h-epsc-enhanced-ev fig2i-ipsc-weaker-nv fig2j-ie-ratio-tracks-isi fig2k-ipsc-exceeds-epsc-575
  • Assumed, for the Methods' inclusion rule ("Only neurons that exhibited at least five significant prototype-evoked E/IPSCs"), which does not define a significant PSC: a trial whose evoked current in the expected direction exceeds k s.d. of the cell's sham measure (k = 2 and 3). It keeps 1 of 39 cell recordings at k = 2 and none at k = 3; the other checks of the data check, which need no such definition, agree that the currents almost never follow the calls (two cells carry an inward current during the call).ed6ef-epsc-tuning fig2h-epsc-enhanced-ev fig2i-ipsc-weaker-nv fig2j-ie-ratio-tracks-isi fig2k-ipsc-exceeds-epsc-575
  • Every listed cell is used for the computed values; 12 experienced and 14 naive EPSC cells as in the paper, and 7 experienced (paper 6) and 6 naive IPSC cells. One experienced IPSC cell (190410_e1) has 4 or 5 trials per ISI.ed6ef-epsc-tuning fig2h-epsc-enhanced-ev fig2i-ipsc-weaker-nv fig2j-ie-ratio-tracks-isi fig2k-ipsc-exceeds-epsc-575
  • The largest 5 or 10 trials are ranked by absolute value (the Methods' "absolute magnitudes"), and, as a second reading, by the current in the expected direction.ed6ef-epsc-tuning fig2h-epsc-enhanced-ev fig2i-ipsc-weaker-nv fig2j-ie-ratio-tracks-isi fig2k-ipsc-exceeds-epsc-575
  • The spreadsheet notes are applied (stop or drop at a sweep keeps the sweeps before it, from sweep N keeps the later ones, drop or block + stop drops the file), only the best call's set is used where a note names one, and sweeps whose holding current contradicts the listed potential are dropped (file 19424013 is listed at -70 mV but carries +560 pA). Using every sweep is a variant.ed6ef-epsc-tuning fig2h-epsc-enhanced-ev fig2i-ipsc-weaker-nv fig2j-ie-ratio-tracks-isi fig2k-ipsc-exceeds-epsc-575
  • Experienced cells pool dams and EVs; the spreadsheets do not say which are which.ed6ef-epsc-tuning fig2h-epsc-enhanced-ev fig2i-ipsc-weaker-nv fig2j-ie-ratio-tracks-isi fig2k-ipsc-exceeds-epsc-575
  • One listed file (193260007, an IPSC file of naive cell 190326_e1) is not on figshare; the cell's other IPSC file is used.ed6ef-epsc-tuning fig2h-epsc-enhanced-ev fig2i-ipsc-weaker-nv fig2j-ie-ratio-tracks-isi fig2k-ipsc-exceeds-epsc-575
  • Touch times are printed with five significant digits, so a press can be resolved only to 0.1 ms below 10 s, 1 ms below 100 s, 10 ms to 1000 s and 100 ms after; a gap must exceed one and a half local steps (1 ms taken as the step below 100 s) to end a press, under both gap readings, so the median-interval rule ends a press at 1.5, 15 or 150 ms depending on when it falls.fig1i-operant-prototypes ed5b-press-count-unchanged
  • The median inter-sample interval (reading a) is the polling period of the task code, taken once, pooled over sessions, from the times printed at 0.1 ms (below 10 s), 0.70 ms; above 10 s the printed intervals are quantized at 1 ms or more. Before revision a mean over the 1 ms part was used (about 0.7 ms); the outcome is unchanged.fig1i-operant-prototypes ed5b-press-count-unchanged
  • Two mice miss one session each (NV24 session 5, NV38 session 3); neither is a compared session.fig1i-operant-prototypes ed5b-press-count-unchanged
  • Groups are read from the NWB experiment descriptions and, where these give none, from the figshare item titles of the same mouse id (collection doi:10.6084/m9.figshare.c.5043830.v6), as the criterion allows.subject-group-link
  • Assumed: the peak is the largest inward (EPSC) or outward (IPSC) current 2 to 30 ms after each pulse (to 2 ms before the next pulse at 50 ms periods) against the mean of the 5 ms before it. The Methods say only "peak amplitude ... normalized to the event evoked by the first stimulation on each trial"; local baselines are the usual choice for trains.fig1g-ipsc-depress-faster
  • Trials with a first response not above baseline give no ratio and are dropped; the mean trace per file is shown as a variant.fig1g-ipsc-depress-faster
  • One-way ANOVA with two groups is run as the equivalent unpaired t-test; a paired test is shown as a variant.fig1g-ipsc-depress-faster
  • 12 of the 13 cells at 575 ms; file 19o11005 (Cell 20) is listed but not on figshare.fig1g-ipsc-depress-faster
  • The 75 ms bin includes trains at 50 ms periods, at the edge of the Methods' 75 plus or minus 25 ms.fig1g-ipsc-depress-faster
  • The Bonferroni family is read as the three comparisons with the prototype that the legend reports; all six pairs are shown.ed6ef-epsc-tuning
  • Assumed: the Friedman test runs over ten columns (baseline and post for each of five stimuli), with Dunn's comparisons of baseline against post for each stimulus, Bonferroni over five; the legend gives one p per stimulus, which fits five planned comparisons. All 45 pairs and Wilcoxon tests are variants and change the outcome.ed9b-oxytocin-enhances-spiking
  • A spike is a crossing of 0 mV between a pulse and the next, a standard choice.ed9b-oxytocin-enhances-spiking
  • Baseline is the files the spreadsheet labels baseline or PRE (the first file for one control cell with no such label; the notebook prints each file's label and role); for controls the 15 to 30 minute window starts at the first sweep after the baseline files, a reading of "following baseline acquisition", and from the first baseline sweep as a variant.ed9b-oxytocin-enhances-spiking
  • Under the main reading two of the 7 controls (191202B, 191202C) have no sweeps in the window: each has one 60-sweep baseline file (0 to 11.8 min) and one later file ending 19.3 and 24.8 min after the first baseline sweep.ed9b-oxytocin-enhances-spiking

Provenance

Attempt record
000249-schiavo-2020-01
Paper
Innate and plastic mechanisms for maternal behaviour in auditory cortex
Date
2026-10-29
Made by
An agent, run by CatalystNeuro
Dataset
DANDI:000249
Version
0.230423.1416 (doi:10.48324/dandi.000249/0.230423.1416)
Run on
GitHub Actions
Harness
Claude Code
Model
claude-opus-5-5
Protocol
pilot-12

Transcripts of the agent runs

Each run’s prompt, messages and tool calls, with long output cut and the text of the paper left out, and its final report.

← DANDI:000248DANDI:000252 →