DANDI:000010

Reproducing Li et al. (2015) from dandiset 000010 18

Notebook reproduce_li_2015.ipynb

• ✕Not reproducedResults, ALM contains neurons with bilateral selectivity, paragraph 1; Figure 2d

  Among selective neurons, 219 of 912 had selective preparatory activity only, 389 peri-movement activity only, and 304 both.

  subsets of neurons showed selective preparatory activity (Fig. 2b, left column; Fig. 2d, 219/912 neurons), peri-movement activity during the response epoch (Fig. 2b, right column; 389/912), or both

  Results, ALM contains neurons with bilateral selectivity, paragraph 1; Figure 2d

  Under the main reading, 231 selective neurons are preparatory only (paper 219, within), 284 peri-movement only (paper 389) and 397 both (paper 304). Over the eight readings (trials, response epoch, light trials kept or left out), preparatory only is 228 to 233 (always within), peri-movement only 272 to 294 (never within) and both 289 to 397 (within in three readings). No reading puts all three counts within tolerance. The answer on light trials named two readings, (t1) all trials and (t2) trials without the light, either of which may decide; the outcome is the same under both, so it is unchanged. Counted by class, 628 neurons are preparatory (sample or delay) against the paper's 523 (219 + 304) and 681 peri-movement (response) against 693 (389 + 304), so the mismatch lies in the preparatory class.

  How the test differed from the paper’s method

  • Trials with tagging light are found from the laser traces; readings (t1) and (t2) are both run and give the same outcome.
• ✕Not reproducedMethods, Electrophysiology data analyses, paragraph 1; Extended Data Figure 1b to everifier disagreed at first

  Fast-spiking neurons have higher baseline firing rates and are less selective than putative pyramidal neurons.

  Fast-spiking neurons had higher baseline firing rates and were less selective than putative pyramidal neurons (Extended Data Fig. 1)

  Methods, Electrophysiology data analyses, paragraph 1; Extended Data Figure 1b to e

  Fast-spiking neurons have higher baseline rates under both readings (0.5 s before the sample: median 4.51 against 1.50 spikes/s, mean 6.90 against 2.92; Mann-Whitney P = 1.3e-17). They are not less selective: 0.750 of them are selective against 0.733 of putative pyramidal neurons (Fisher P = 0.75; 0.710 and 0.699 without light trials), and their mean absolute rate difference is larger (median 1.46 against 0.62 spikes/s). The baseline part holds and the selectivity part goes the other way. The criterion requires both parts under every reading, so the outcome, earlier partially_reproduced, is not_reproduced (first verification). The answer on light trials named two readings, (t1) all trials and (t2) trials without the light, either of which may decide; the outcome is the same under both.

  How the test differed from the paper’s method

  • Baseline reading (b) covers the 0.57 s of recorded time before the sample event, as spikes are stored almost only in trial windows (at least 99% of them in 97 of 99 sessions), so it is close to reading (a).
  • Trials with tagging light are found from the laser traces; readings (t1) and (t2) are both run and give the same outcome.
• ✕Not reproducedResults, A movement command emerges in the pyramidal tract neurons, paragraph 2; Figure 3f, j

  Contralateral population selectivity emerges in PT neurons from 570 ms before the response epoch (34% spike rate modulation, P < 0.05), but not in IT neurons (-19%, P = 0.40).

  Robust contralateral selectivity (the spike count differenceforlickrighttrials and licklefttrials) emerged in pyramidal tract neuron responses starting 570 ms before the response epoch (Fig. 3f; 34% spike rate modulation; P , 0.05, two-tailed t-test against0inspikecounts),butnotforintratelencephalicneurons(219%

  Results, A movement command emerges in the pyramidal tract neurons, paragraph 2; Figure 3f, j

  What the paper reports

  • PT contra-selectivity against 0: two-tailed t-test against 0 in spike counts, p < 0.05 (n = 45 PT neurons)
  • IT contra-selectivity against 0: two-tailed t-test against 0 in spike counts, p = 0.4 (n = 27 IT neurons)

  PT contra-selectivity is positive before the go cue but not significant in either test window under any reading (P = 0.17 to 0.36 with the light trials, 0.12 to 0.22 without), and the windowed selectivity is not significant through to the go cue, so no onset is found (without light trials 8 to 16 single 200 ms windows reach P < 0.05, not continuously). PT modulation is 13.9% to 27.0% with light trials and 18.9% to 38.5% without, against 34%. IT is not significant (P = 0.06 to 0.13), with modulations of -15.9% to -29.7%, some within the tolerance of -19%. The answer on light trials named two readings, (t1) all trials and (t2) trials without the light, either of which may decide; the outcome is the same under both, so it is unchanged.

  How the test differed from the paper’s method

  • Trials with tagging light are found from the laser traces; readings (t1) and (t2) are both run and give the same outcome.
• ✕Not reproducedMethods, Calcium imaging data analyses, paragraphs 2 and 4; Results, A movement command emerges in the pyramidal tract neurons, paragraph 3verifier disagreed at first

  In two-photon imaging of ALM (59 fields of view, 4 mice), 58% of cells (2,740 of 4,706) were task-related and 57% of those (1,575) trial-type selective, with a slight majority of contra-preferring cells (61%, 971 of 1,575).

  This criterion identified 58% (2,740/4,706) cells as task related ... 57% (1575/2740) of task related cell are trial-type selective ... Overall, imaging detected a slight majority of contra-preferring neurons (61%, 971/1,575)

  Methods, Calcium imaging data analyses, paragraphs 2 and 4; Results, A movement command emerges in the pyramidal tract neurons, paragraph 3

  Changed from partially reproduced after verification round 3 agreed. The criterion sets only a reproduced condition, all values within tolerance under one reading, and no reading meets it (0 of 16); four values are within tolerance in 12 readings, and the selective count only in the 4 readings with span (b) and the mean over epochs, where the task-related count is above it. The selective count is measured and misses under every reading, unlike the intermediate units of units-classification, a class the dandiset lacks. The measured values are from the reading (1.3 s response epoch, span a, mean over epochs, correct trials). The dandiset holds 4,706 included cells in 59 planes, the paper's numbers. Over 16 readings, the task-related count is 2,551 to 2,772 with span (a), within 2,740 +/- 274, and 3,104 to 3,134 with span (b), above it; the contra-preferring fraction is 0.547 to 0.590, within tolerance in 15 readings. The selective count is 1,315 to 1,417 where the task-related count fits (paper 1,575, lower limit 1,417.5) and 1,469 to 2,138 elsewhere. No reading puts all five values within tolerance; the closest misses the selective count by 0.5 (1,417; 1.5 s response epoch, span a, mean over epochs, all non-early trials). The answer on the response epoch made (r1) 1.3 s and (r2) 1.5 s readings of the criterion; both were already run, so the outcome is unchanged. The selective count is below tolerance with the mean over epochs (1,315 to 1,417 where the task count fits) and above it with each epoch separately (1,844 to 1,957); the paper's 1,575 lies between. Baselines near 0 are rare (F0 at or below 0 in 145 of 305,349 included ROI-trials, below a tenth of the ROI's median F in 622); leaving them out, a check chosen here and not a reading, gives 1,416 (cut at 0) or 1,419 (cut at 0.1, then all five values within tolerance) in the closest reading, so the selective count sits at the edge of the tolerance; the cut is not a reading and does not decide the outcome. Early licks cannot be left out, as the imaging files hold only sample, delay and go events and flag no early lick. The neuropil-corrected F of 3,403 included ROIs reaches 0 in at least one frame; the longest such run is one frame in 2,034 of them, two to seven frames in 1,199 and eight frames (0.5 s) or more in 170.

  How the test differed from the paper’s method

  • The files' `included` flag stands for the Methods' brightness rule; 424 included ROIs have a median F below 1.05 times their neuropil.
  • Early licks cannot be excluded, as the imaging files flag none.
• ✕Not reproducedResults, A movement command emerges in the pyramidal tract neurons, paragraph 3; Figure 4f; Methods, Calcium imaging data analyses, paragraph 3

  Contra- and ipsi-preferring neurons are found in all fields of view and are spatially intermingled; nearest-neighbour selectivity and same-type distances do not differ from shuffled data (P = 0.34 and P = 0.61).

  Contra-preferring and ipsi-preferring neurons were observed in all fields of view (Fig. 4f; Methods). The probability that nearby neurons share the same selectivity was not different from randomized data (P 5 0.34, bootstrap, Methods). The distance between nearest neurons with the same selectivity was also not different from chance (P 5 0.61, bootstrap)

  Results, A movement command emerges in the pyramidal tract neurons, paragraph 3; Figure 4f; Methods, Calcium imaging data analyses, paragraph 3

  What the paper reports

  • probability that the nearest neighbour has the same preference: bootstrap, shuffled labels, p = 0.34
  • distance between nearest neurons of the same preference: bootstrap, shuffled labels, p = 0.61

  In 57 of 64 combinations of selection reading, neighbour set and shuffle (10^4 shuffles each), at least one test gives P < 0.05: nearby cells share their preference more often, and same-preference cells are closer, than with shuffled labels (for example 0.539 against 0.517 for shuffles). Every combination with all cells as neighbours finds it. In 9 of 16 selection readings, some plane lacks one of the two preferences. Neighbouring ROIs share neuropil, which may contribute. The answer on the response epoch, (r1) 1.3 s and (r2) 1.5 s, names the two lengths already run in the selection readings, so the outcome is unchanged. The shuffle across planes mixes planes with different contra fractions and is not a spatial null; with the shuffle within planes, 28 of 32 rows still give P < 0.05.

  How the test differed from the paper’s method

  • Distances are in pixels between ROI centroids; the files give no spatial scale.
• ◐Partially reproducedMethods, Electrophysiology data analyses, paragraph 1; Extended Data Figure 1a

  1,408 single units were recorded from left ALM in 19 mice over 99 sessions; by spike width 124 were fast-spiking, 1,245 putative pyramidal and 39 intermediate.

  A total of 1,408 single units were recorded across 99 recording sessions ... Units with spike width ,0.35 ms were defined as fast-spiking neurons (124/1,408) and units with spike width .0.45 ms as putative pyramidal neurons (1,245/1,408)

  Methods, Electrophysiology data analyses, paragraph 1; Extended Data Figure 1a

  The files hold 1,441 units: 1,245 labelled Pyr, 124 FS, 45 PT and 27 IT. Every PT and IT unit has a spike train identical to a Pyr unit in the same file, so there are 1,369 distinct units (within 1,408 +/- 140.8; 1,441 is also within). Sessions 99 and mice 19 match. By trough-to-peak width of the mean waveform, 124 units are fast-spiking (below 0.35 ms) and 1,245 putative pyramidal (above 0.45 ms), as in the paper, and none is intermediate, against 39. Five of six counts match; the intermediate units are not in the dandiset. The waveforms are 29 samples with the trough at sample 15, so widths cap at 0.666 ms (746 units), which does not affect the thresholds.

  How the test differed from the paper’s method

  • Duplicated PT/IT units are counted once.
• ◐Partially reproducedMethods, Electrophysiology data analyses, paragraph 4; Extended Data Figure 5f

  Of 73 neurons antidromically activated from pyramidal tract axons, 45 passed the collision test (PT neurons), 22 failed and 6 could not be tested; of 134 activated from IT axons, 27 passed (IT neurons), 106 failed and 1 could not be tested.

  Neurons that passed the collision test were classified as pyramidal tract neurons (45/73) or intratelencephalic neurons (27/134). Neurons that failed the test were classified as pyramidal-tract-activated (22/73) or intratelencephalic-activated (106/134)

  Methods, Electrophysiology data analyses, paragraph 4; Extended Data Figure 5f

  The files label 45 units PT and 27 IT, the paper's counts. The 22 PT-activated and 106 IT-activated neurons that failed the collision test, and the 7 that could not be tested, carry no label, so two of the four counts cannot be checked. The 8 IT units of the two Tlx x Ai32 mice have no light record in their files.

  How the test differed from the paper’s method

  • Only the labelled classes can be counted.
• ◐Partially reproducedMethods, Electrophysiology data analyses, paragraph 4; Results, A movement command emerges in the pyramidal tract neurons, paragraph 1; Extended Data Figure 5e

  PT neurons had shorter antidromic latencies (mean 2.7 ms) than IT neurons (4.6 ms), and both had smaller jitter (0.1 and 0.2 ms) than activated neurons that failed collision (latency 8.3 and 9.8 ms, jitter 2.8 and 2.4 ms).

  Pyramidal tract neurons and intratelencephalic neurons had shorter mean antidromic spike latencies (pyramidal tract neurons, 2.7ms; intratelencephalic neurons, 4.6 ms) and smaller temporal jitter (mean s.d. of the antidromic spike latency, pyramidal tract neurons, 0.1 ms; intratelencephalic neurons, 0.2 ms) than the pyramidal-tract-activated (mean latency, 8.3ms; jitter, 2.8ms) and intratelencephalic-activated neurons

  Methods, Electrophysiology data analyses, paragraph 4; Results, A movement command emerges in the pyramidal tract neurons, paragraph 1; Extended Data Figure 5e

  PT latency is below IT latency in all 18 combinations of reading and parameter (windows of 5, 8 and 10 ms; collision intervals of 5, 10 and 20 ms for (c); 0.5 and 1 ms around the evoked peak for (r)). PT latency is within tolerance in all (2.65 to 2.78 ms, paper 2.7); IT latency in 10 (4.63 to 5.00 ms with the 8 and 10 ms windows under (w) and (c), 4.90 to 4.91 ms under (r) with 8 ms; 3.61 to 3.66 ms with 5 ms and 5.09 ms under (r) with 10 ms are outside). The jitter matches the printed precision only under (r), 0.115 to 0.129 ms for PT and 0.188 to 0.211 ms for IT with 0.5 ms around the peak (0.31 to 0.33 ms with 1 ms); under (w) and (c) it is 0.30 to 0.64 ms and 0.75 to 0.96 ms, which include spontaneous spikes. All four measurable values are within tolerance in one combination ((r), 8 ms window, 0.5 ms; PT 2.674 and 0.115 ms, IT 4.905 and 0.197 ms); keeping spikes within 0.5 ms of the peak bounds the s.d., so that agreement is partly set by the reading, and the window and peak width are open parameters chosen here (1 of 18 combinations). The 22 PT-activated and 106 IT-activated neurons that failed collision carry no label, so their four values, and the comparison of jitter with them, cannot be tested. 45 PT and 19 of 27 IT units have a light record. The answer named readings (w), (c) and (r) with the window and intervals open; the claim, earlier blocked as methods_underspecified, is tested under all three and is partially reproduced under each, as the order holds and values fall outside or cannot be measured.

  How the test differed from the paper’s method

  • The open parameters are run at several values (window 5, 8 and 10 ms; collision interval 5, 10 and 20 ms; peak width 0.5 and 1 ms around the 0.1 ms histogram bin with most spikes).
  • Only the neurons that passed collision are labelled, so the activated neurons cannot be measured.
  • The 8 IT units of the two Tlx x Ai32 mice have no light record and are not included.
• ✓ReproducedResults, ALM is required for movement planning, paragraph 1

  Mice in the object location discrimination task performed at 78.4% correct on average, and fewer than 13% of responses came before the go cue.

  (mean per cent correct, 78.4%; responses before the go cue, ,13%)

  Results, ALM is required for movement planning, paragraph 1

  Without the trials with tagging light, in the 99 extracellular sessions, per cent correct is 76.5% pooled, 75.8% over sessions and 75.2% over mice with no-lick trials counted as not correct (reading x), and 77.2%, 76.5% and 75.9% with them left out (reading y), all within 78.4 +/- 7.84%. Early licks are 11.6% of trials over sessions, below 14.3%, but 14.5% pooled and 14.7% over mice, so the claim holds under reading (b) only, with either no-lick reading. Keeping the light trials gives 75.3%, 75.0% and 74.4% correct under (x) (76.6%, 76.3% and 75.7% under y) and 13.0%, 11.7% and 14.7% early licks. The imaging sessions, which flag no early licks, give 82.3% (x) and 83.5% (y) correct pooled. The answer on no-lick trials made (x) and (y) readings of the criterion, either of which may decide; both were already run, so the outcome is unchanged, and the measured value is now that of (x), the first reading listed (75.8%, earlier 76.5% under y). Dropping also the 103 trials without an early lick whose go cue is missing or not 2.6 s after the sample event, as the other claims' trial masks do, gives 75.8% (x) and 76.5% (y) over sessions and 11.7% early licks over sessions (14.5% pooled, 14.8% over mice), the same outcome.

  How the test differed from the paper’s method

  • The dandiset holds 23 of the paper's 42 task mice (the recording and imaging mice).
  • Trials with tagging light are found from the laser traces, as the trial tables do not flag them.
  • 1,540 extracellular trials with a lick instruction but no events (1,447 of them marked right, hit, no early) are left out; counted as given they would raise performance.
  • Only the extracellular sessions are used, as the imaging files flag no early licks.
• ✓ReproducedResults, ALM contains neurons with bilateral selectivity, paragraph 1; Methods, Electrophysiology data analyses, paragraph 2

  73% (912 of 1,245) of putative pyramidal neurons distinguished the trial types in at least one trial epoch.

  A large fraction(73%,912/1,245)ofneuronsdistinguishedtrialtypes(P , 0.05, two-tailed t-test, not corrected for multiple comparisons) (Fig. 2b)

  Results, ALM contains neurons with bilateral selectivity, paragraph 1; Methods, Electrophysiology data analyses, paragraph 2

  Correct trials, a 1.3 s response epoch and all trials (t1) give 912 of 1,245 putative pyramidal neurons selective (0.733). The other readings give 910, 846 and 851; without the trials with tagging light (t2), 870 (0.699), 868, 795 and 792. All are within tolerance. The answer on light trials wrote (t1) and (t2) into this criterion for every electrophysiology claim of Figs 2 and 3, either of which may decide; both were already run and agree here, so the outcome is unchanged.

  How the test differed from the paper’s method

  • Trials with tagging light are found from the laser traces, as the trial tables do not flag them.
  • 66 extracellular trials without an early lick whose go cue is not 2.6 s after the sample event, those with no go cue (75 extracellular, 2 imaging) and trials without events are dropped.
• ✓ReproducedResults, ALM contains neurons with bilateral selectivity, paragraph 1; Figure 2c; Extended Data Figure 1d, f

  Selectivity emerges in the sample epoch, increases through the delay epoch and peaks at the beginning of the response epoch.

  Selectivityemergedinthesampleepoch,increasedthroughoutthedelay epoch, and reached a maximum at the beginning of the response epoch

  Results, ALM contains neurons with bilateral selectivity, paragraph 1; Figure 2c; Extended Data Figure 1d, f

  With 200 ms and 50 ms steps, pre-sample windows hold 52 to 65 selective neurons (binomial limit 71 to 74), sample-epoch windows reach 272 and 280, the count grows through the delay (326 to 408, 321 to 408), and the maximum (574, 582) falls 0.2 s and 0.35 s after the go cue. The first window after the go cue holds 500, so reading (3a) fails and (3b) holds under both steps. Without the trials with light the same holds (maximum 525 and 538 at 0.2 s and 0.25 s). The answer on light trials named two readings, (t1) all trials and (t2) trials without the light, either of which may decide; the outcome is the same under both, so it is unchanged.

  How the test differed from the paper’s method

  • Spikes are stored only from 0.57 s before the sample event, so there are two (200 ms step) or eight (50 ms step) pre-sample windows.
  • Trials with tagging light are found from the laser traces; readings (t1) and (t2) are both run and give the same outcome.
• ✓ReproducedResults, ALM contains neurons with bilateral selectivity, paragraph 1; Figure 2c, bottom; Methods, Electrophysiology data analyses, paragraph 2

  Individual neuron responses are dynamic although selectivity is stable at the level of the population (Fig. 2c, bottom, population selectivity and the correlation of the population response with that at the go cue).

  Individualneuronresponseswere dynamic, despite stable selectivity at the level of the population (Fig. 2c, bottompanelandExtended Data Fig.2a)

  Results, ALM contains neurons with bilateral selectivity, paragraph 1; Figure 2c, bottom; Methods, Electrophysiology data analyses, paragraph 2

  With the light trials kept (t1), all 24 correlation readings hold (neurons a or b, normalization v or s, concatenated or difference vector, go-cue bin gs, gc or ge). The mean correlation with the go-cue vector is lower over the sample epoch (-0.12 to 0.24) than over the delay epoch (0.04 to 0.78), below 1, and the population selectivity is above 0 in every delay bin under the four selectivity readings, n10 and n20 (minimum 0.62 spikes/s). Without the light trials (t2), 22 of 24 correlation readings hold; two fail (all neurons, variance, difference vector, bin starting at or centred on the go cue, sample 0.078 and 0.126 against delay 0.054 and 0.103). The answers on the three details named readings (v)/(s), (n10)/(n20) and (gs)/(gc)/(ge); the centred bin (gc) is new and holds under t1 (sample -0.04 to 0.24, delay 0.11 to 0.67). The answer on light trials lets either of (t1) and (t2) decide, so the claim, earlier blocked because one t2 reading failed, is reproduced under t1. Under (ge) the go-cue bin lies inside the delay epoch, so its correlation of 1 enters the delay mean; (gs) avoids this and holds in all 8 t1 rows (sample -0.123 to 0.148, delay 0.036 to 0.478).

  How the test differed from the paper’s method

  • The bin centred on the go cue comes from a second grid of 200 ms bins shifted by 0.1 s, normalized over that grid.
  • Trials with tagging light are found from the laser traces; under (t2) two readings fail.
• ✓ReproducedResults, ALM contains neurons with bilateral selectivity, paragraph 1; Extended Data Figure 2c

  On error trials ALM activity predicts the direction the mouse licked, so most selective neurons switch their trial-type preference relative to correct trials.

  On error trials, when mice licked in the opposite direction to the instruction provided by object location (Fig. 1a), a majority of ALM neurons switched their trial type preference to predict the licking direction, as indicated by the negative correlations

  Results, ALM contains neurons with bilateral selectivity, paragraph 1; Extended Data Figure 2c

  Correlation across selective neurons between correct-trial and error-trial selectivity: delay r = -0.30 (897 neurons, at least one error trial of each type) and -0.25 (833, at least five); response -0.38 and -0.34; all P < 1e-12. Without the light trials: delay -0.29 and -0.32, response -0.33 and -0.30. Sample epoch near 0 (-0.025 to 0.045). Per mouse the delay r ranges from -0.76 to 0.49. The answer on light trials named two readings, (t1) all trials and (t2) trials without the light, either of which may decide; the outcome is the same under both, so it is unchanged.

  How the test differed from the paper’s method

  • Trials with tagging light are found from the laser traces; readings (t1) and (t2) are both run and give the same outcome.
• ✓ReproducedResults, ALM contains neurons with bilateral selectivity, paragraph 2; Figure 2e; Results, A movement command emerges in the pyramidal tract neurons, paragraph 2

  Contra-preferring and ipsi-preferring neurons are present in equal proportions in left ALM, and the overall spike count shows no significant contralateral bias (P = 0.13; 8% spike rate modulation across all task-selective neurons).

  Significant contralateral bias also did not appear in the overall spike count across the recorded population of neurons (Extended Data Fig. 1; P 5 0.13, two-tailed t-test against 0) ... The contra-preferring and ipsi-preferring neurons were present in equal proportions

  Results, ALM contains neurons with bilateral selectivity, paragraph 2; Figure 2e; Results, A movement command emerges in the pyramidal tract neurons, paragraph 2

  What the paper reports

  • overall spike count difference, lick right minus lick left: two-tailed t-test against 0, p = 0.13

  On the paper's test as written (overall spike count over the three epochs, all putative pyramidal neurons), the lick right minus lick left difference is not different from 0 (P = 0.82; 0.52 without light trials). The bootstrap finds no contralateral excess (419 of 912 selective neurons contra-preferring, 0.459; 408 of 870 without light trials). The statement's equal proportions do not hold strictly: ipsi-preferring neurons are the majority (two-sided binomial P = 0.016, or 0.072 without light trials). In the 570 ms before the go cue the difference reaches P = 0.022 to 0.050 across neuron and light readings, and the delay epoch gives P = 0.051 to 0.12. The 8% modulation is matched in some readings (all neurons, 570 ms, over the mean rate, 8.5%; selective neurons, delay, light trials left out, 8.0% and 8.3%) and not in others (5.8% to 10.7%). Only one reading meets both the absence test and the 8% modulation, selective neurons, delay epoch, light trials left out (P = 0.060, 7.97%), and its value is the measured one (earlier 8.5%, from all neurons in the 570 ms window, where the test gives P = 0.036 with a contralateral excess). The answer on light trials named two readings, (t1) all trials and (t2) trials without the light, either of which may decide; the outcome is the same under both, so it is unchanged. The outcome rests on reading (a)(i) because the protocol judges a claim of absence on the paper's test as written ("the overall spike count across the recorded population"). It is fragile: in the 570 ms window, where the Results report the same P = 0.13 for all task-selective neurons, all four tests give P = 0.022 to 0.050 with a contralateral excess. The criterion sets no rule for deciding a claim of absence across its readings, so the outcome depends on the reading taken: on the paper's second P = 0.13 test, for all task-selective neurons, the 570 ms window gives P = 0.050 and 0.038 with a contralateral excess and the delay epoch P = 0.12 and 0.060.

  How the test differed from the paper’s method

  • The claim of absence is judged on the paper's test as written ('overall spike count across the recorded population'), reading (a)(i); the other window and neuron readings of the criterion are shown, and the 570 ms window finds a contralateral excess.
  • Trials with tagging light are found from the laser traces; readings (t1) and (t2) are both run and give the same outcome.
• ✓ReproducedResults, A movement command emerges in the pyramidal tract neurons, paragraph 2; Figure 3e, i

  Both PT and IT neurons show preparatory and movement-related activity.

  Both intratelencephalic and pyramidal tract neurons displayed preparatory and movement-related activity (Fig. 3e, i)

  Results, A movement command emerges in the pyramidal tract neurons, paragraph 2; Figure 3e, i

  Under the main reading, 17 of 45 PT neurons have preparatory and 24 peri-movement selectivity, 13 of 27 IT neurons each. Every reading, with or without the light trials, gives at least 11 neurons of each kind in each type. The answer on light trials named two readings, (t1) all trials and (t2) trials without the light, either of which may decide; the outcome is the same under both, so it is unchanged.

  How the test differed from the paper’s method

  • PT/IT neurons are the Pyr twins of the labelled duplicates.
  • Trials with tagging light are found from the laser traces; readings (t1) and (t2) are both run and give the same outcome.
• ✓ReproducedResults, A movement command emerges in the pyramidal tract neurons, paragraph 2; Figure 3f, j insets

  Contra-preferring neurons outnumber ipsi-preferring neurons among PT neurons (P < 0.05, bootstrap) but not among IT neurons (P = 0.74).

  Contra-preferring neurons outnumbered ipsi-preferring neurons for the pyramidal tract neurons (Fig. 3f; P , 0.05, bootstrap, Methods), but not the intratelencephalic neurons (Fig. 3j, P 5 0.74)

  Results, A movement command emerges in the pyramidal tract neurons, paragraph 2; Figure 3f, j insets

  What the paper reports

  • PT neurons, more contra- than ipsi-preferring: one-tailed bootstrap, p < 0.05 (n = 45 PT neurons)
  • IT neurons, more contra- than ipsi-preferring: one-tailed bootstrap, p = 0.74 (n = 27 IT neurons)

  With the light trials kept (t1), 30 of 45 PT neurons are contra-preferring (bootstrap P = 0.010) and 13 of 27 IT neurons (P = 0.58), so the claim holds under reading (a); selective PT neurons give 18 of 29 (P = 0.09). Without the trials with tagging light (t2), PT gives 26 of 45 (P = 0.14) and 17 of 32 selective (P = 0.30), and the claim holds under no reading. The answer on light trials made (t1), the literal reading of the Methods, and (t2) readings of the criterion, either of which may decide, so the claim, earlier blocked because the two disagreed, is reproduced under (t1). The PT result depends on keeping the trials in which the PT neurons were driven by the tagging light. The light is balanced between trial types (0.295 of lick right and 0.307 of lick left correct trials; 4 of 75 sessions with Fisher P < 0.05, 3.8 expected; weighted by PT neurons, 0.005 less light on lick right trials, which would favour ipsi preference). PT neurons are contra-preferring in 26 of 45 on trials without light and 26 of 45 on light trials alone (P = 0.14 and 0.15), and 29 of 45 (P = 0.022) on the average of the two differences; light trials hold 1.27 times the spikes (median). So the t1 result is not a bias from the light; it reaches P < 0.05 only with the trials pooled.

  How the test differed from the paper’s method

  • Trials with tagging light are found from the laser traces; the claim holds under (t1) only.
• ⊘BlockedResults, A movement command emerges in the pyramidal tract neurons, paragraph 4; Figure 4g, h· Metadata insufficient

  Among layer 5 neurons, about a quarter of labelled PT (182 of 534) and IT (174 of 542) neurons were selective; 67% of selective PT neurons were contra-preferring (P < 0.001, binomial test) but only 48% of IT neurons, and PT neurons were more often contra-preferring than IT neurons (P < 0.001, chi-squared test).

  we observed more contra-preferring than ipsi-preferring neurons in the pyramidal tract neuron population (67%, 122/182, Fig. 4g, h; P , 0.001, binomial test), but not in the intratelencephalic neurons (fraction of contra-preferring neurons:48%,83/174;Fig.4g,h)

  Results, A movement command emerges in the pyramidal tract neurons, paragraph 4; Figure 4g, h

  What the paper reports

  • PT, contra- against ipsi-preferring: binomial test, p < 0.001 (n = 182 selective PT neurons)
  • IT, contra- against ipsi-preferring: binomial test, p > 0.05 (n = 174 selective IT neurons)
  • PT against IT fraction contra-preferring: chi-squared test, p < 0.001

  No imaging plane gives its depth, so the layer 5 population (450 to 700 um) cannot be selected. The answer found nothing else in the paper that marks the layer 5 planes, and the criterion now says that all labelled cells is a substitute population, not a reading of the paper, whose results are a variant; the paper also reports IT cells in superficial layers. The claim, earlier not_reproduced on all labelled cells, is therefore blocked. On the variant, 530 included cells are labelled PT (paper 534) and 843 IT (paper 542); selective PT cells are contra-preferring in a majority under every reading (0.577 to 0.649; binomial P < 0.001 in 5 of 16 readings) and IT cells are not (0.476 to 0.532; P > 0.05 in all), but the chi-squared test of PT against IT never reaches P < 0.001 (smallest 0.0016), so on this population the three tests never come out as in the paper. A search of every text field and attribute of the 59 imaging files (58 each, including the ImagingPlane group, session and image descriptions) finds no mention of depth, microns or layer.

  How the test differed from the paper’s method

  • All included cells labelled PT or IT are shown as a substitute for the layer 5 cells, which cannot be selected; the IT counts are not expected to match on it.
• ⊘BlockedResults, Pyramidal tract neurons drive upcoming movements, paragraph 1· Methods underspecified

  The reaction time after the go cue was 61.1 +/- 10.5 ms (mean +/- s.d. across mice).

  reactiontimepostcueonsetacrossmice:61.1 6 10.5 ms,mean 6 s.d.

  Results, Pyramidal tract neurons drive upcoming movements, paragraph 1

  The answer confirmed that the paper gives no rule for reading a reaction time from the photodiode signal and named readings (k) a fixed raw threshold and (z) the baseline mean plus a number of s.d., with the threshold, the number and the baseline open. Both are now run on the `lick_trace` of all 99 sessions (24,023 non-early trials, 113 without a stored window). Under (z) the mean across mice is 40.2 to 129.1 ms and rises with the number of s.d. (last 0.5 s as baseline, 40.2 to 44.3 ms at 2 s.d., 59.2 to 61.9 at 3, 71.5 to 75.5 at 4, 76.7 to 83.8 at 5, 92.8 to 106.3 at 10); it is within 61.1 +/- 10.5 ms in all 8 rows at 3 s.d. with the last 0.5 s, 6 of 8 at 2 s.d. and 2 of 8 at 3 s.d. with the delay epoch, 1 at 4 s.d. and none at 5 or 10. Under (k) no single raw threshold suits all sessions, whose baselines run from 0.069 to 4.83 a.u.; 36% to 90% of trials have no crossing and the mean is 77.9 to 166.1 ms, never within tolerance. Since the number of s.d., which the paper does not give, can be chosen to give any mean from about 40 to 130 ms, the claim stays blocked, now for that reason and not for want of a rule. Where the mean matches, the s.d. across mice is 31.5 to 38.5 ms against the paper's 10.5; the signal falls after the go cue in 15 of 99 sessions from 6 mice, and one of them, 219248, gives 178.9 ms; reading (z) counts only upward crossings, so the fall is not detected there.

  How the test differed from the paper’s method

  • The `lick_trace` series is taken as the photodiode signal; the files do not say it is.
  • The open parameters are run at several values (2, 3, 4, 5 and 10 s.d.; raw thresholds 0.1 to 4 a.u.).