DANDI:000015

Reproducing Chen et al. (2017) from dandiset 000015 23

Notebook reproduce_chen_2017.ipynb

• ✕Not reproducedResults, Mapping Activity Using Two-Photon Calcium Imaging, paragraph 3

  A large fraction of neurons distinguished between trial types, 23% in MM (1,523 of 6,594) and 31% in ALM (2,754 of 8,837).

  the activity of a large proportion of neurons distinguished between trial types in MM (23%, 1,523/6,594) ... and ALM (31%, 2,754/8,837

  Results, Mapping Activity Using Two-Photon Calcium Imaging, paragraph 3

  No reading puts all four values within tolerance. The measured values are those of the closest reading, R2, W3, T1, where ALM is 2,853 (32.3%, within tolerance) and MM 1,736 (26.3%), above the tolerance (at most 1,675 and 25.3%); R2, W1, T1 is next, with ALM 2,877 (32.6%) and MM 1,755 (26.6%). Until the second verification the measured values were those of W1, which were given as closest by mistake. Rule R1 marks 78% to 91% of ROIs selective, and 72% to 85% with permuted trial types, so it counts chance bins; R2 with permuted labels gives 10.5% to 14.3%. With T2 bins R2 gives ALM 45% and MM 38% (W1). ALM has the higher fraction under every reading. Without the 1,262 ROIs with a non-positive baseline, MM is 1,675 of 5,854 (28.6%). Among the ROIs the files mark included, ALM 43.6% and MM 34.4%. Per mouse (R2, W1, T1) ALM 25.8% to 40.1% and MM 19.7% to 32.7%. The answer on baselines F0 <= 0 (the paper does not say) made dF/F as computed (V1) the reading that decides, as run before, and V2, without the ROIs with F0 <= 0 on any trial, a check; V2 is now reported for every reading, and none puts all four values within tolerance (R2, T1 gives ALM 2,762 of 8,315, 33.2%, and MM 1,656 of 5,854, 28.3%, in W3, and ALM 2,786, 33.5%, and MM 1,675, 28.6%, in W1). The outcome is unchanged.

  How the test differed from the paper’s method

  • Reading A2 not run (no coordinates).
• ✕Not reproducedResults, Analysis of Behavior-Related Activity, paragraph 5; Figure 4C

  Outcome-selective neurons typically became selective after the answer lick; some preferred correct and others error trials, and some kept the outcome long after the trial ended.

  These neurons typically became selective after the answer lick. A subset of the outcome-selective neurons responded preferentially during correct trials ... whereas others responded strongly during error trials

  Results, Analysis of Behavior-Related Activity, paragraph 5; Figure 4C

  Of 9,431 ROIs purely outcome selective in any bin (T1), 41.3% have their first such bin after the go cue (T2 41.8% of 8,892), below the half the criterion requires. With permuted trial types 8,070 ROIs qualify, 29.6% first after the go cue, so most of the 9,431 are chance bins. Requiring two consecutive bins, a variant that does not decide, leaves 1,463 ROIs, 78.5% first after the go cue. Both signs of gamma occur (4,157 correct-preferring and 5,274 error-preferring at the first bin), and outcome selectivity reaches the end of the window (9.17 s after trial start).

• ✕Not reproducedResults, Spatial Map of Behavior-Related Activity, paragraph 2; Figure 6G to 6I

  Outcome-selective neurons were seen only during the response epoch, scattered roughly uniformly through ALM and MM; mixed neurons were also seen mainly in the response epoch in both areas.

  Outcome-selective neurons, which were observed only during the response epoch, were scattered approximately uniformly throughout ALM and MM

  Results, Spatial Map of Behavior-Related Activity, paragraph 2; Figure 6G to 6I

  Part (b) holds; averaged over the response epoch the ALM and MM outcome fractions differ by a factor of 1.46 (T1) and 1.45 (T2). Part (a) fails in MM; its largest outcome fraction before the go cue (0.0112, T1) is not below a third of the largest after it (0.0268); in ALM 0.0123 against 0.0373 just passes. T2 gives the same. The pre-cue maxima are near the range of permuted-label fractions (0.0067 to 0.0105), so the test turns on chance fluctuations against a small response peak in MM; part (a) makes no allowance for the false-positive floor of about 1% per bin. Mixed ROIs (reported only) are 2.8 times as frequent in ALM as in MM over the response epoch.

• ✕Not reproducedResults, Mapping Activity Using Two-Photon Calcium Imaging, paragraph 3; Figure S4 (not read)

  In two-photon imaging, the peak of activity shifted from MM to ALM over the delay.

  the peak of activity shifted from MM to ALM over the course of the delay epoch (Figure S4)

  Results, Mapping Activity Using Two-Photon Calcium Imaging, paragraph 3; Figure S4 (not read)

  Over the whole window the mean dF/F of MM ROIs peaks after that of ALM (6.09 against 4.52 s, T1; 6.14 against 4.54, T2). The ALM mean at the end of the delay (0.212) is above its sample-epoch mean (0.112) and maximum (0.137). The MM mean is dominated by ROIs with a non-positive baseline (range -1.15 to 0.71); without them (V2) MM peaks at 4.80 s, still after ALM. The peak is taken over the whole extracted window (the bins of convention T). Searching only before the go cue, a variant that does not decide, MM peaks before ALM under every bin and dF/F reading (V1 2.08 against 4.16 s with T1 bins and 0.80 against 4.14 s with T2; V2 1.51 against 4.09 s and 1.53 against 4.14 s), and within the delay epoch alone V1 gives 2.08 against 4.16 s (T1) and 2.60 against 4.14 s (T2). The outcome therefore turns on the window of the peak, which the criterion does not name; a question was asked after the first verification (questions/000015.yaml) and is not yet answered, and the outcome is kept on the whole window until it is. The answer on the criterion's terms (the paper does not say) named readings L1 and L2 for the sample-epoch level and E1 (last bin) and E2 (mean over the last 0.5 s) for the end of the delay; E2 is new here. The ALM mean over the last 0.5 s is 0.205 (T1), also above both levels, so the second part holds in all 8 combinations of bins and readings and the first fails in all; the answer on F0 <= 0 made V1 decide and V2 a check, as run before. The outcome is unchanged.

• ✕Not reproducedResults, Projection-Specific Coding, paragraphs 1 and 2; Figure 8D

  During the delay epoch, ALM PT neurons showed a bias toward contralateral (lick right) selectivity.

  A contra bias during the delay epoch was ... observed in PT neurons (Figure 8D), consistent with previous electrophysiological recordings

  Results, Projection-Specific Coding, paragraphs 1 and 2; Figure 8D

  The mean beta of 385 ALM PT ROIs over the delay is positive (lick right), 0.0284 with T1 bins, 1.71 hierarchical bootstrap SDs above zero (T2 0.0291, 1.56 SD), short of the 2 SD bound. Without ROIs with a non-positive baseline it is 1.97 (T1) and 1.86 (T2) SD. It is positive in each of the three mice (0.0093 to 0.0341). The direction of the claim holds; its bound does not. The answer on F0 <= 0 (the paper does not say) made V1, the dF/F as computed, decide and V2, without those ROIs, a check; both were run before and the outcome is unchanged.

• ✕Not reproducedDiscussion, paragraph beginning 'MM shows little preparatory activity'; Figure 8D

  MM neurons were more active in posterior trials, when the pole was within reach of the whiskers, than in anterior trials.

  Neurons were more active in posterior trials, when the pole was within reach of the whiskers, than in the anterior location, when the pole was out of reach (Figure 8D

  Discussion, paragraph beginning 'MM shows little preparatory activity'; Figure 8D

  The mean alpha of the 6,594 MM ROIs over the sample epoch is positive (posterior preferring), 0.0232 with T1 bins, 1.87 hierarchical bootstrap SDs above zero (T2 0.0242, 1.94), short of the 2 SD bound (1.70 and 1.63 without ROIs with a non-positive baseline). It is positive in all five mice with MM planes, and significant MM ROI-bins in the sample epoch prefer the posterior pole 1,912 times against 747. The direction holds; the bound does not. The answer on F0 <= 0 (the paper does not say) made V1 decide and V2 a check; both were run before and the outcome is unchanged.

• ✕Not reproducedDiscussion, paragraph beginning 'We also observed neurons representing trial outcome'; Figure 8Dverifier disagreed at first

  Overall, neurons preferring error trials outnumbered those preferring correct trials.

  Overall, neurons preferring error trials outnumbered neurons preferring correct trials (Figure 8D

  Discussion, paragraph beginning 'We also observed neurons representing trial outcome'; Figure 8D

  Reading (a) finds error preference; among ROIs purely outcome selective from the go cue to 2.5 s after it, 2,825 prefer error and 2,004 correct trials (T1, binomial p = 2.8e-32; T2 2,730 and 1,933). Reading (b) does not reach its bound; mean gamma over all ROIs is -0.0082 (T1), negative in every mouse, but 1.81 bootstrap SDs from zero (T2 1.70). The criterion requires both readings. The answer on F0 <= 0 (the paper does not say) made V1 decide, as run before, and V2 a check; V2 is now run in full. Under V2 (a) holds (2,722 against 1,927) and (b) reaches 2.20 SD with T1 bins and 2.00 SD with T2, which the printed test puts just short of the bound. The outcome, decided by V1, was partially_reproduced before the first verification and is now not_reproduced; the criterion requires both readings and names no partial outcome, and the other bootstrap claims whose direction holds but whose bound does not (fig8d-pt-contra-bias, fig8d-mm-posterior-preferring) are not_reproduced too.

• ◐Partially reproducedResults, Mapping Activity Using Two-Photon Calcium Imaging, paragraph 1; Figure 2E, 2F

  Two-photon imaging covered 12 locations in 6 mice with 8.8 +- 2.6 planes per location, 105 planes in all, each imaged for 92 +- 26 trials at 73% +- 9% correct, 3 +- 1 planes per session.

  For each mouse (n = 6), we collected data from 1 to 3 image locations (n = 12 locations in total, Figure 2E), each containing multiple (8.8 ± 2.6) image planes ... Each image plane (n = 105, Figure 2F and S2) was sampled at 14 Hz while mice performed the task for 92 ± 26 trials, with performance at 73% ± 9% trials correct ... allowing sequential imaging of 3 ± 1 planes per session (range 1–6

  Results, Mapping Activity Using Two-Photon Calcium Imaging, paragraph 1; Figure 2E, 2F

  Six mice and 105 distinct planes, as stated. Reading (a), the four analysed trial types, gives 92.15 +- 25.98 trials per plane at 73.83% +- 9.03% correct, both within tolerance. Reading (b), all trials, gives 107.37 trials, within tolerance, at 63.89% correct, just outside. Planes per session 3.09 +- 1.11, range 1 to 6, over 34 sessions. The 12 locations and the 8.8 planes per location cannot be counted, as the files carry no location identifier; subject and area give 10 combinations, which fits 1 to 3 locations per mouse but does not count them. Every value that could be measured is within tolerance; the outcome is partial because two of the seven could not be tested.

  How the test differed from the paper’s method

  • The location counts are not tested, for lack of a location identifier.
• ◐Partially reproducedResults, Spatial Map of Behavior-Related Activity, paragraph 1; Figures 6F to 6I and S7 (not read)

  Lick-direction selectivity, in number of neurons and strength, kept increasing and spread across ALM layers and to a minor extent into MM, peaking after the go cue.

  Lickdirection selectivity, including number of neurons and the strength of coding, continued to increase and spread across ALM layers and to a minor extent into MM (Figure S7), reaching a maximum after the go cue

  Results, Spatial Map of Behavior-Related Activity, paragraph 1; Figures 6F to 6I and S7 (not read)

  Part (a) holds; in ALM the fraction of purely lick-direction-selective ROIs and mean |beta| peak after the go cue (4.44 and 4.66 s, T1; 4.67 and 4.67 s, T2). Part (c) holds; the MM fraction peaks at 0.026 (T1), above its delay mean 0.008 and delay maximum 0.015, and below the ALM peak 0.062, under both readings of its delay value and both bin readings. Part (b), the spread across ALM layers, cannot be tested, as the files hold no depth. The outcome is partial because part (b) is untested, not because it failed. The answer on the MM delay value (the paper does not say) made the mean and the maximum over the delay bins readings of the criterion, (c) holding under one; both were already run and both hold, so nothing changed.

  How the test differed from the paper’s method

  • Part (b) not tested (no plane depth).
• ◐Partially reproducedResults, Projection-Specific Coding, paragraph 1; Figure 8C

  IT neurons were found in both superficial and deep layers, PT neurons only in deep layers; 719 PT and 2,631 IT neurons were imaged in three mice.

  In three mice, we combined retrograde labeling of pyramidal tract (PT) (n = 719) and intratelencephalic (IT) (n = 2,631) neurons ... IT neurons were found in both superficial and deep layers, whereas PT neurons were exclusively in deep layers

  Results, Projection-Specific Coding, paragraph 1; Figure 8C

  719 PT and 2,631 IT ROIs in three mice (an063, an064, an066), exactly the paper's counts. The depth part (PT only deep, IT in both) cannot be tested, as the files hold no plane depth. The outcome is partial because that part is untested.

  How the test differed from the paper’s method

  • Depth not tested.
• ✓ReproducedResults, paragraph 2 (four trial types)

  Mice licked early, during the delay epoch, on 13.3% +- 10% of trials, and these trials were not analysed.

  Trials in which the mice licked early during the delay epoch (13.3% ± 10%) were not analyzed

  Results, paragraph 2 (four trial types)

  Mean over the 105 planes of the percentage of trials with an early lick, 13.24 +- 9.14 with all trials as denominator (reading a). Every reading is within 13.3 +- 10, from 13.24 (planes) to 15.08 (mice, trials with a lick). Per mouse 8.8% to 21.9%.

• ✓ReproducedResults, Mapping Activity Using Two-Photon Calcium Imaging, paragraph 3

  Each imaged plane had on average 34 +- 13 CR, 36 +- 14 CL, 12 +- 5 ER and 10 +- 4 EL trials.

  average number of trials per imaged plane ... CR ... 34 ± 13, CL ... 36 ± 14; ER ... 12 ± 5, EL ... 10 ± 4

  Results, Mapping Activity Using Two-Photon Calcium Imaging, paragraph 3

  Mean trials per plane CR 34.06 +- 13.43, CL 35.51 +- 13.94, ER 12.18 +- 4.71, EL 10.40 +- 4.42, all within the paper's spread. Every plane has at least 3 trials of each type.

• ✓ReproducedResults, Mapping Activity Using Two-Photon Calcium Imaging, paragraphs 2 and 3

  15,431 neurons were imaged across 6 mice, 8,837 in ALM and 6,594 in MM.

  Our sample (15,431 neurons across 6 mice) is therefore biased toward the active population

  Results, Mapping Activity Using Two-Photon Calcium Imaging, paragraphs 2 and 3

  15,431 ROIs over the 105 distinct planes, 8,837 in planes labelled left_alm and 6,594 in planes labelled left_vm1 (reading A1), exactly the paper's counts. Reading A2 cannot be run, as no coordinates are in the files.

  How the test differed from the paper’s method

  • Area from the files' experiment_description (left_alm, left_vm1), as ImagingPlane.location reads ALM for every plane.
• ✓ReproducedResults, Analysis of Behavior-Related Activity, paragraph 2; Figure S5 (not read)

  The selectivity of most neurons can be attributed to one of the three coefficients, especially early in the trial, so most neurons mainly encode one behavioural variable at a time.

  The selectivity of most neurons can be attributed to one of the three coefficients, especially during the early part of the trial (Figure S5)

  Results, Analysis of Behavior-Related Activity, paragraph 2; Figure S5 (not read)

  Among ROIs trial-type selective in a bin, the fraction purely selective exceeds one half in all 59 bins (T1) and 63 bins (T2) from trial start to the go cue, 0.845 to 0.928 (T1). After the go cue its mean is 0.793 (T1), 0.784 (T2). With permuted trial types (T1) the fraction is 0.845 to 0.911 and above one half in all 59 bins, so the criterion is met at chance as well; the outcome rests on the criterion, which sets no chance level, and does not show that selectivity is more often single-factor than chance gives.

• ✓ReproducedResults, Analysis of Behavior-Related Activity, paragraph 3; Figure 4A

  Different neurons coded for lick direction in different epochs, some during the sample and delay epochs before licking, some during the response epoch, and some after the mouse stopped consuming water.

  Some neurons were selective during the sample and delay epochs, long before onset of licking ... Other neurons became selective during the response epoch ... Some neurons even represented licking direction after the animal stopped consuming water

  Results, Analysis of Behavior-Related Activity, paragraph 3; Figure 4A

  With T1 bins, 4,125 ROIs are purely lick-direction selective in the delay epoch, 3,611 are first so after the go cue and 4,426 more than 2 s after it (T2 3,825, 3,514 and 4,155). With permuted trial types the counts are 3,273, 2,526 and 3,196, so each group exceeds chance but most of each count is what chance bins give, and the criterion, which asks only that each group exists, is met by the permuted labels too. The time after the last lick is not tested, as lick times are not in the files (the behavioural events of every plane are sample, delay and go).

  How the test differed from the paper’s method

  • The median last lick, which the criterion reports as the reading of "stopped consuming water", cannot be computed; lick times are not in the files.
• ✓ReproducedResults, Analysis of Behavior-Related Activity, paragraph 4; Figure 4B

  Many object-location-selective neurons were selective during the sample epoch, and a subset coded object location mainly in the delay epoch after the pole was removed, some up to the end of the delay.

  Many of these neurons were selective during the sample epoch when mice explored the object ... However, a subset of neurons coded object location primarily during the delay epoch, after the pole was removed ... Some of these neurons carried information about object location up to the end of the delay epoch

  Results, Analysis of Behavior-Related Activity, paragraph 4; Figure 4B

  With T1 bins, 2,148 ROIs are purely object-location selective in the sample epoch, 2,487 have most of their object-location bins in the delay and 1,149 are selective in the last 0.5 s before the go cue (T2 2,067, 2,483, 1,074). With permuted trial types the counts are 1,694, 1,719 and 707; each group exceeds chance, and the criterion, which asks only that each group exists, is met by the permuted labels too.

• ✓ReproducedResults, Analysis of Behavior-Related Activity, paragraph 6; Figure 5

  Most neurons with mixed selectivity were seen during the response epoch; a subset responded in only one trial type and a few switched the variable they coded during the trial.

  Most of these neurons with ‘‘mixed’’ selectivity were observed during the response epoch

  Results, Analysis of Behavior-Related Activity, paragraph 6; Figure 5

  76.5% of 23,363 mixed-selective ROI-bins fall after the go cue (T1; T2 77.6%), against 50% of bins and 52.3% with permuted labels. 4,682 ROIs (T1) are purely object-location selective before the go cue and purely lick-direction or outcome selective after it.

• ✓ReproducedResults, Spatial Map of Behavior-Related Activity, paragraph 1; Figure 6B to 6D, 6H

  During the sample and early delay epochs, object location was represented mainly in MM.

  During the sample and the early part of the delay epoch, representation of object location was prominent, mainly in MM

  Results, Spatial Map of Behavior-Related Activity, paragraph 1; Figure 6B to 6D, 6H

  The fraction of purely object-location-selective ROIs is higher in MM than in ALM in 27 of 31 bins from trial start to 2.2 s and 34 of 38 to 2.7 s (T1; T2 29 of 33 and 37 of 41). Mean fractions 0.037 in MM and 0.012 in ALM to 2.2 s (T1). In every mouse with both areas MM is higher.

• ✓ReproducedResults, Spatial Map of Behavior-Related Activity, paragraph 2; Figure 6H, 6I, S6 and S8 (not read)

  Throughout the trial preparatory activity, as the fraction of lick-direction-selective neurons or the average coding coefficient, was most pronounced in ALM, while object location selectivity was mainly limited to MM.

  Throughout the trial, preparatory activity, quantified either as fraction of lick direction-selective neuron (Figures 6H and 6I) or as average coding coefficient (Figures S6 and S8), was most pronounced in ALM, whereas object location related selectivity was mainly limited to MM

  Results, Spatial Map of Behavior-Related Activity, paragraph 2; Figure 6H, 6I, S6 and S8 (not read)

  In the delay epoch ALM is above MM on both lick-direction measures in 24 of 42 bins (T1) and 30 of 45 (T2); on the fraction alone in all delay bins, on mean |beta| alone in 24 and 30. From trial start to the go cue MM is above ALM on both object-location measures in 52 of 59 (T1) and 57 of 63 (T2) bins.

• ✓ReproducedResults, Projection-Specific Coding, paragraph 2; Figure 8D

  In ALM, lick direction selectivity (|beta|) was stronger and earlier in PT than in IT neurons, with an onset difference of 1.4 +- 0.6 s (p < 0.01, bootstrap).

  The magnitude of lick direction selectivity (jbj) appeared stronger and earlier in PT than in IT neurons (delta onset time; time when selectivity rose to >6 SD above baseline = 1.4 ± 0.6 s; p < 0.01, bootstrap

  Results, Projection-Specific Coding, paragraph 2; Figure 8D

  What the paper reports

  • onset of mean |beta|, ALM PT against ALM IT: bootstrap, p < 0.01 (n = 719 PT and 2,631 IT neurons from 3 mice (all areas))

  Reading I2 (all ALM IT). Mean |beta| is larger in PT than in IT (maximum 0.229 against 0.084, T1; 0.231 against 0.085, T2). The answer on the onset rule (the paper does not say) added rule S2, the first bin from which the curve stays above the threshold up to the go cue, beside S1, the first single bin, and fixed the threshold at the baseline mean plus 6 SD, as run before. Under S2 with T2 bins and the bootstrap over cells (H2), the IT minus PT onset is 1.801 s (O1, bootstrap SD 0.415 s, p = 0.003) and 1.401 s (O2, SD 0.434 s, p = 0.006), within 1.4 +- 0.6 s at p < 0.01, so the criterion (one combination) is met; the measured value is the first such combination in the criterion's order (O1, S2, H2). The result rests on those combinations. Every S2 combination is within tolerance (1.29 to 1.80 s), but with T1 bins p is 0.021 (O2) and 0.170 (O1) under H2, and with the hierarchical bootstrap (H1) p is 0.065 to 0.319, with up to 307 of 1,000 resamples having no S2 onset (counted as PT not earlier). Rule S1 gives 0.215 s (T1, O1), 0.200 s (T2) and -1.075 s (T1, O2), with p from 0.31 to 0.52, the earlier outcome (not_reproduced) unchanged under S1. Per mouse (T1, O1, S1) PT is earlier in an063 and an064 and later in an066. Reading I1, deep ALM IT only, could not be run without depth, so the outcome rests on I2. The passing combinations use T2 bins of 66.7 ms, interpolated from frames of 71.7 ms (13.95 Hz) or 70.4 ms (14.20 Hz), so neighbouring T2 bins are correlated; T2 also raises the counts of rule R2 in fraction-selective-alm-mm (ALM 45% against 32.6% with T1).

  How the test differed from the paper’s method

  • Reading I1 not run (no plane depth).
• ✓ReproducedResults, Projection-Specific Coding, sentence beginning 'However, this difference in lick direction selectivity onset'; Discussion, paragraph beginning 'We analyzed neuronal selectivity'uses data from outside the dataset

  The difference in onset of lick direction selectivity between PT and IT neurons was not detected in an electrophysiology dataset recorded under similar conditions (45 PT and 27 IT cells, Li et al., 2015).

  However, this difference in lick direction selectivity onset was not detected in an electrophysiology dataset recorded under similar conditions (Li et al., 2015)

  Results, Projection-Specific Coding, sentence beginning 'However, this difference in lick direction selectivity onset'; Discussion, paragraph beginning 'We analyzed neuronal selectivity'

  Blocked (methods_underspecified) before the answer. The answer (the paper does not say) named the readings, run here on dandiset 000010 version 0.220126.1905. Spike counts in 400 ms windows in 100 ms steps from the sample event, on the 9,828 trials of the four types without photostimulation and without an early lick in the 50 sessions (14 mice) with a tagged unit, each unit counted only in the trials within its observation intervals; 4,152 of 16,571 trials had photostimulation (laser pulses or photostim events) and are left out. One IT unit (mouse 219253) has no ER trial left and cannot be fitted, so 45 PT and 26 IT units are analysed. No combination finds PT earlier at p < 0.01. The IT minus PT onset of mean |beta| is 0.7 s with rule S1 (PT 0.1 s, IT 0.8 s; bootstrap SD 0.66 to 0.75 s, p 0.30 to 0.34) and 0.2 s with rule S2 (PT 1.2 s, IT 1.4 s; SD 0.63 to 0.64 s, p 0.15 to 0.20), under both bootstraps. Baselines O1 and O2 are the same two windows (centres -0.3 and -0.2 s), as spikes are stored only from 0.52 s before the sample event, so the 8 combinations give 4 distinct results, each with a baseline SD from two overlapping windows; the IT baseline SD is 0.0004 (threshold 0.0228 against a mean of 0.0205), and the PT onset under S1 (0.1 s) falls in the sample epoch, so the test carries little information. The difference is positive under every reading but not significant.

  How the test differed from the paper’s method

  • One of the 27 IT units is left out, as it has no error-right trial without photostimulation.
  • Baselines O1 and O2 coincide, as no spikes are stored before 0.52 s ahead of the sample event.
• ✓ReproducedDiscussion, paragraph beginning 'We analyzed neuronal selectivity'; dandiset 000010 version 0.220126.1905checks the link to the paper’s data

  The tagged units of dandiset 000010 (Li et al. 2015) are the 45 PT and 27 IT cells of the electrophysiology dataset the paper compares its imaging with.

  (PT/IT cells; electrophysiology, 45/27; imaging, 719/2,631)

  Discussion, paragraph beginning 'We analyzed neuronal selectivity'; dandiset 000010 version 0.220126.1905

  Dandiset 000010 version 0.220126.1905 labels 45 units PT (40 in sim1-cre and 5 in rbp4-cre mice) and 27 IT (24 in tlx-cre mice, 3 in rbp4-cre), the paper's 45/27, among 1,441 units in 99 files. Matched by label and count only; no identifier links them to the paper's analysis.

• ✓ReproducedDiscussion, paragraph beginning 'We analyzed neuronal selectivity'; Figures 6, 8 and S10 (not read)

  Even in ALM, less than 20% of PT or IT cells displayed lick direction selectivity at any given time, and the onset timing was variable.

  even in ALM, less than 20% of PT or IT cells displayed lick direction selectivity at any given time (Figures 6, 8, and S10), and the onset timing was variable

  Discussion, paragraph beginning 'We analyzed neuronal selectivity'; Figures 6, 8 and S10 (not read)

  The largest per-bin fraction of lick-direction-selective ALM PT ROIs is 12.7% (reading a, purely selective) or 15.6% (reading b, beta significant), and of ALM IT ROIs 3.3% or 6.0% (T1); with T2 at most 16.1%. Per mouse at most 19.8% (an063, PT, reading b). All below 22%.