Claims 41
From the paper Peri-head distance coding in the mouse brainstem · claims file · checked 2026-10-21: ready
- figs1f-angle-awake-vs-anesthetized
Wall-induced change in whisker midpoint angle does not differ significantly between awake and anesthetized mice for the C1, C2 and Greek whiskers (paired t-tests across 4 mice, Bonferroni alpha 0.017); the mean awake minus anesthetized difference is 2.28 degrees.
Awake mice showed similar bending angle (mean difference: 2.28°) and curvature (mean difference: 1.66 × 10 −5 mm −1 ) changes, with no significant differences from the anesthetized condition across whiskers
Results, A wall-passing paradigm, paragraph 2; Figure S1F legendHow it will be judged
A claim of absence, judged on the paper's test (Methods, Data Analysis: Wall-Induced Whisker Retraction; FigS1_Stats.m). Per pass, the midpoint angle (Savitzky-Golay 21 frames, order 3, mean of the upper and lower envelopes) is averaged over 3 to 11 s minus its mean over -4.5 to 0 s; per mouse, whisker and condition, averaged over distances and passes; for each whisker a two-tailed paired t-test across the 4 mice, significant at 0.05/3. Reproduced when no whisker reaches p < 0.0167; not reproduced when any does. The mean difference is reported for each whisker; the paper does not say whether 2.28 degrees is pooled over whiskers or for one, so the readings are the mean over the three whiskers and each whisker alone. The code (FigS1_Code.m) takes the midpoint as a 101-frame Savitzky-Golay smoothing of the twice 21-frame-smoothed angle, with the envelope method commented out; this is the variant, shown, and the Methods' envelope midpoint is the main.
- figs1j-curvature-awake-vs-anesthetized
Wall-induced change in whisker curvature does not differ significantly between awake and anesthetized mice for any of the three whiskers.
(H-J) Same as D-F, but for curvature. Significance reflects mouse-level paired comparisons (awake vs. anesthe- tized) with Bonferroni correction (α = 0.017): C1, p = 0.56 (n.s.); C2, p = 0.053 (n.s.), Greek: p = 0.43 (n.s.)
Results, A wall-passing paradigm, paragraph 2; Figure S1H-J legendHow it will be judged
A claim of absence, as figs1f-angle-awake-vs-anesthetized, for curvature (signed planar curvature from a local second-order fit to a 4-mm arclength segment, minus its baseline mean over -4.5 to 0 s, averaged over 3 to 11 s; Methods). Reproduced when no whisker reaches p < 0.0167. The Methods do not say where along the whisker the 4-mm segment lies; this is open, and the author reports the segment used.
- figs1d-bending-larger-closer
Under anesthesia the passing wall bends whiskers in the retraction direction, with larger angle changes and larger curvature changes at closer wall distances, so wall-induced bending grows roughly monotonically as the wall approaches.
Under anesthesia, the moving wall pushed whiskers in the retraction direction, producing larger absolute bending angles ( Supplemental Figure S1D , right) and greater change in whisker curvature ( Supplemental Figure S1H , right) at closer wall distances
Results, A wall-passing paradigm, paragraph 2; Figure S1D, HHow it will be judged
The paper reports no test. Per distance, the across-mouse mean of the absolute change in midpoint angle and of the absolute change in curvature in the anesthetized condition (measures as in figs1f-angle-awake-vs-anesthetized). Reproduced when, for each of the two measures and each of the three whiskers, the value at 7 mm exceeds the value at 23 mm and the Spearman correlation between distance and value over the nine distances is negative. These two summaries are the readings of "larger ... at closer wall distances"; the outcome is the one both give. The cited panels (S1D and S1H, right) show one example C1 whisker in one mouse; the across-mouse curves per whisker are in S1F and S1J, which this criterion uses.
- figs2-tungsten-responsiveness
Of 46 tungsten-recorded PrV units from 14 mice, 41 (89.1%) were wall-activated, 3 (6.5%) wall-suppressed, and 43 (93.5%) wall-responsive in all, one unit being both.
Of the 46 units, 41/46 (89.1%) were wall-activated and 3/46 (6.5%) were wall-suppressed; one unit showed both activation and suppression at different distances, yielding a total of 43/46 (93.5%) wall-responsive
Results, PrV contains both proximity and map codes, paragraphs 1-2How it will be judged
Classification as in the opening comment, applied to the tungsten sessions with their own distances (4 to 20 mm in 1-mm steps, so the binning into nine 2-mm bins does not apply; the code leaves the number of distances to each session). Responsiveness is judged with the Methods' uncorrected p < 0.05 per distance; the code's BH-FDR across distances is the variant, shown. Reproduced when all four counts are within tolerance under the Methods' test.
- figs2-tungsten-tuning
Of the 43 wall-responsive tungsten units, 39 (90.7%) were distance-tuned; 24 (55.8%) were broadly monotonic, firing more at closer distances, and 6 (14.0%) peaked at intermediate or far distances.
Wall-responsive units with significant differences in mean firing rates for at least one distance pair were classified as "distance-tuned" (i.e., distance-selective; 90.7%, 39/43) ... The majority of wall-responsive units (55.8%, 24/43) showed a broadly monotonic, inversely proportional relationship with wall distance ... a distinct subset (14.0%, 6/43) showed peak responses at intermediate or far wall distances
Results, PrV contains both proximity and map codes, paragraph 2; Figure S2D-FHow it will be judged
Distance tuning as in the opening comment (36 pairs become all pairs of the session's distances). The paper does not say how the 24 monotonic and 6 peaked units were defined for the tungsten data. Readings: (a) the silicon-probe rule, proximity (preferred at the closest distance) and map (preferred farther, with a significant BH-FDR one-tailed test against the closest distance), with activated or suppressed units both eligible as in the code's tungsten branch; (b) preferred distance at the closest distance against elsewhere, without the map test. Reproduced when all three counts are within tolerance under one reading.
- si-unit-yield
Silicon probe recordings in wild-type mice yielded 1,141 single units from 36 mice passing isolation and firing criteria, out of 1,358 sorted units.
For silicon probe data, in wild-type mice, 1141 of 1358 units (36 mice), in TeLC experiments, 587 of 665 units (11 mice)
Results, PrV contains both proximity and map codes, paragraph 3; Methods, Isolation of Single Unit ActivityHow it will be judged
Units with SNR of at least 5 and ISI false-alarm rate at most 2% (0.8 ms), not low-firing as defined in the opening comment, in wild-type full-array sessions (including the pre-cut part of whisker-cutting sessions). The number of mice is a count and must match exactly. If the deposit carries only curated units, the count of deposited units is compared and the difference reported.
- fig2-wall-responsive-counts
Of the 1,141 units, 768 (67.3%) were wall-responsive; of these 532 (69.3%) were activated at one or more distances, 333 (43.4%) suppressed, and 97 both.
Of the 1,141 units, 67.3% (768/1141) were significantly modulated by the passing-wall stimulus and classified as "wall-responsive". Among these wall-responsive units, 532/768 (69.3%) showed significant activation at one or more distances, 333/768 (43.4%) showed significant suppression, and 97 units met both criteria
Results, PrV contains both proximity and map codes, paragraph 3How it will be judged
Classification as in the opening comment (Methods, uncorrected p < 0.05 per distance; variant, the code's BH-FDR across the nine distances, shown). Reproduced when all four counts are within tolerance under the Methods' test.
- fig2-distance-tuned-counts
Over the full wall-contact period, 656 of 768 wall-responsive units (85.4%) were distance tuned; of these, 472 (72.0%) were activated at one or more distances (the activated population), 213 (32.5%) suppressed (the suppressed population), and 29 both.
Across the full wall-contact period (8 s), we identified 656 out of the 768 (85.4%) wall-responsive units that were significantly tuned to wall distance. Within this distance-tuned population, 472/656 (72.0%) showed significant activation relative to baseline at one or more distances ... whereas 213/656 (32.5%) showed significant suppression at one or more distances ... A small subset of units (29/656; 4.4%) exhibited both
Results, PrV contains both proximity and map codes, paragraph 4How it will be judged
Distance tuning as in the opening comment. Reproduced when all four counts are within tolerance under the Methods' responsiveness test (the code's BH-FDR variant shown). The 472 activated units are the population of Figures 2D-E, 5D, 6G-I and S4.
- fig2f-tuning-categories
Among the 768 wall-responsive units, 266 (34.6%) were proximity units, 118 (15.4%) map units, 88 (11.5%) ambiguous, and 119 (15.5%) suppressed-only; the remainder (177, 23.0%) were untuned, giving a proximity to map ratio of 2.3 to 1.
We identified 266 proximity units (34.6%) and 118 map units (15.4%) out of the 768 wall-responsive units ( Figure 2F ) ... An additional 88 units (11.5%) were distance-selective but did not meet criteria for either map or proximity and were labeled "ambiguous" ... Suppressed-only units ... accounted for 119 out of 768 wall-responsive units (15.5%; Figure 2F )
Results, PrV contains both proximity and map codes, paragraph 7; Figure 2F; Figure 5E legendHow it will be judged
Categories as in the opening comment. The paper's counts are not consistent with each other (472 + 213 - 29 = 656 tuned implies 184 suppressed-only and 112 untuned; the five Figure 2F counts sum to 768 with 119 and 177). Readings for the last two: (a) the Figure 2F counts as printed; (b) the counts implied by fig2-distance-tuned-counts (184 and 112, same 10% tolerance; the Methods also count 112 silicon-probe units as wall-responsive but not distance-tuned, Characterizing Wall Distance Tuning). Reproduced when the proximity, map and ambiguous counts are within tolerance and the suppressed-only and untuned counts are within tolerance under one reading; partially reproduced when only the proximity, map and ambiguous counts are.
- figs4a-baseline-activated-lower
Before wall movement, units of the activated population fire less than untuned units (5.46 against 6.93 sp/s).
At baseline prior to wall movement onset, units in the activated population showed lower firing rates than untuned units (5.46 ± 0.670 vs. 6.93 ± 0.947 sp/s; two-tailed Kolmogorov-Smirnov (K-S) test: p = 2.47 × 10 − 7 )
Results, PrV contains both proximity and map codes, paragraph 4; Figure S4AHow it will be judged
Per unit, the mean baseline rate (-4.5 to 0 s) over all passes; two-sample K-S test (two-tailed, kstest2) between the activated population and untuned units (Methods). The tolerances are the paper's printed spread (bootstrap CI half-width in the Methods, SEM in the legend). Reproduced when p < 0.05 with the activated mean lower and both means within tolerance; partially reproduced when the test holds but a mean is out of tolerance.
- figs4a-baseline-suppressed-similar
Units of the suppressed population have baseline firing comparable to untuned units (7.31 against 6.93 sp/s, not significant).
whereas units in the suppressed population had baseline firing rates comparable to untuned units (7.31 ± 1.124 vs. 6.93 ± 0.947 sp/s; two-tailed K-S test: p = 0.63
Results, PrV contains both proximity and map codes, paragraph 4; Figure S4AHow it will be judged
A claim of absence, judged on the paper's test, a two-tailed two-sample K-S test of per-unit mean baseline rates, suppressed population (213) against untuned units, at p < 0.05. Reproduced when p >= 0.05; not reproduced when p < 0.05. The difference of means is reported either way.
- figs4b-activated-increase-suppressed-decrease
During wall passes the activated population fires more than at baseline (16.44 against 6.36 sp/s at the preferred distance) and the suppressed population fires less (4.47 against 6.19 sp/s).
During wall-passing period, the activated population increased firing relative to baseline (16.44 ± 1.640 vs. 6.36 ± 0.704 sp/s; one-tailed K-S test: p = 9.26 × 10 −28 ; Supplemental Figure S4B , left), while the suppressed population exhibited a significant decrease in firing (4.47 ± 1.000 vs. 6.19 ± 1.076 sp/s; one-tailed K-S test: p = 3.06 × 10 −4
Results, PrV contains both proximity and map codes, paragraph 5; Figure S4BHow it will be judged
Per unit, the mean wall rate (3 to 11 s) at the preferred distance (activated) or at the distance of minimum rate (suppressed; Figure S4B legend), against the unit's baseline rate; one-tailed two-sample K-S tests in the stated direction (Methods). The baseline paired with it is not specified, and the paper's 6.36 sp/s differs from the 5.46 sp/s of the same population's all-pass baseline (figs4a-baseline-activated-lower). Readings: (a) baseline of the passes at that same distance; (b) baseline over all passes. Reproduced when both tests give p < 0.05 in the stated direction and the four means are within tolerance under one reading; partially reproduced when both tests hold but no reading puts the means within tolerance.
- figs4c-untuned-no-change
Untuned units do not change their firing between baseline and wall passes (6.93 against 6.32 sp/s).
Untuned units showed no significant change between baseline and wall-passing periods (6.93 ± 0.947 vs. 6.32 ± 0.953 sp/s; two-tailed K-S test: p = 0.53
Results, PrV contains both proximity and map codes, paragraph 5; Figure S4CHow it will be judged
A claim of absence, judged on the paper's test, a two-tailed two-sample K-S test between per-unit mean baseline rates and per-unit wall rates averaged across distances (Figure S4C legend) for untuned units, at p < 0.05. Reproduced when p >= 0.05; not reproduced when p < 0.05. The difference of means is reported either way.
- fig2de-preferred-distances-span
In both the activated and the suppressed populations a large proportion of units prefer the closest wall position (7 mm), while others are best tuned to intermediate or far distances.
In both populations, while a large proportion of units preferred the closest wall position (7 mm), others were best tuned to intermediate or far distances ( Figure 2E , Supplemental Figure S4G ), revealing coexistence of proximity and map-like coding schemes in PrV
Results, PrV contains both proximity and map codes, paragraph 6; Figure 2D-E; Figure S4D, GHow it will be judged
Preferred distance is the maximum of the mean tuning curve (activated population, 472) and the minimum (trough, suppressed population, 213; Figure S4G legend and the code's pref_dist_sup). Reproduced when, in each population, 7 mm is the most frequent preferred distance and at least one unit prefers each of the eight distances from 9 to 23 mm. The paper reports no test; the most frequent bin and coverage of the range are the reading of "a large proportion ... others".
- figs4ef-crossvalidated-stability
Preferred distances of the activated population are stable across two random halves of the wall passes, whereas the suppressed population is only partly reproducible, with only proximity-preferring responses stable.
The resulting distance-tuning heatmaps of the activated population revealed little changes in the preferred distance between cross-validated partitions ( Supplemental Figure S4E ). In contrast, the suppressed population showed partial reproducibility, where only proximity-preferring responses were stable, while map-like profiles lacked consistent structure
Results, PrV contains both proximity and map codes, paragraph 6; Figure S4E-FHow it will be judged
Per unit, passes at each distance are split at random into two halves (the code draws floor(reps/2) repetition indices without replacement, the same split for all distances of a unit); preferred distance (maximum for activated, minimum for suppressed units) is taken in each half. The paper gives no measure of stability. Readings: (a) the fraction of units whose preferred distance is the same in both halves; (b) the Spearman correlation of preferred distance across halves. Reproduced when, under both readings, the activated population is more stable than the suppressed population, and, within the suppressed population, units whose first-half trough is at 7 mm keep it more often than units with a trough at 9 mm or more. A fixed seed is a standard choice; the split is repeated 100 times and the median reported.
- figs4j-ventral-location
Wall-responsive units, and proximity and map units among them, lie preferentially in ventral PrV, toward the probe tip.
We identified 266 proximity units (34.6%) and 118 map units (15.4%) out of the 768 wall-responsive units ( Figure 2F ), which were preferentially located in ventral PrV ( Supplemental Figure S4J )
Results, PrV contains both proximity and map codes, paragraphs 3 and 7; Figure S4I-JHow it will be judged
Depth is the shank coordinate of the unit's peak channel (largest peak-to-trough amplitude of the mean waveform) from the probe tip, Poly2 probes only, in 100-um bins (Methods, Unit Depth Along Silicon Probe; FigS4I_probemap.m), n = 1050 well-isolated and 703 wall-responsive units (Figure S4I-J legends). The paper reports no test. Readings: (a) the median depth of wall-responsive units is smaller than that of unresponsive units, and the median depth of proximity and of map units is smaller than that of unresponsive units; (b) the fraction of units that are responsive is higher in the 0 to 400 um half of the shank than in the 400 to 800 um half, and so is the fraction that are proximity or map units. Reproduced when both readings agree.
- figs4l-temporal-windows-map
Applied to temporal windows, the classification finds map units most often in the sustained period (122 of 768, 15.9%), less in the onset period (83, 10.8%) and fewest in the offset period (34, 4.4%).
Map units were most prevalent during the sustained period among wall-responsive units (15.9%, 122/768), but were also present during the onset epoch (10.8%, 83/768). In contrast, offset responses yielded fewer map units (4.4%, 34/768)
Results, PrV contains both proximity and map codes, paragraph 8; Figure S4LHow it will be judged
The classification of the opening comment with the wall rate taken in the onset (2.5 to 4.5 s), sustained (4.5 to 9.5 s) or offset (9.5 to 11.5 s) window. The paper does not say whether the 768 wall-responsive units are reclassified for responsiveness in each window or kept from the full-period classification. Readings: (a) responsiveness kept from the full period, tuning and map tests rerun in each window; (b) all tests rerun in each window. Reproduced when the three counts are within tolerance, and ordered sustained > onset > offset, under one reading; partially reproduced when the order holds but a count is out of tolerance.
- fig3a-decoding-above-controls
Population decoders trained on map, proximity, suppressed-only and combined map plus proximity units decode wall distance during the contact period better than shuffled and untuned controls.
As expected, decoding accuracy during the wall-contact period was higher in the map, proximity, suppressed-only, and combined "map + proximity" groups compared to shuffled and untuned controls ( Figure 3A )
Results, The map code in PrV allows more accurate readout, paragraph 1; Figure 3AHow it will be judged
Decoding as in the opening comment. Reproduced when the overall accuracy (mean over the nine distances, 3 to 11 s) of each of the four groups exceeds that of both the shuffled and the untuned groups, and a two-sided Wilcoxon rank-sum test (MATLAB ranksum default, as in the code; the Table S2 value 1.83e-4 is the two-sided normal-approximation p for complete separation of 10 against 10) on the 10 subsample values has p < 0.05 for each such pair, under the Methods' 10-fold cross-validation and under the code's 15 splits.
- fig3-overall-decoding-accuracy
Averaged across the nine distances, the map group decodes best (71.5%), ahead of the combined map plus proximity group (58.6%) and the proximity group (48.9%); the suppressed-only group reaches 32.6% and untuned units 15.7%, against a chance level of 11.1%.
the map group achieved the highest overall decoding accuracy averaged across all nine distances (71.5%), outperforming the proximity (48.9%) and combined "map + proximity" (58.6%) groups. The suppressed-only group had an average overall decoding accuracy of 32.6%. Untuned units performed slightly above chance (15.7%)
Results, The map code in PrV allows more accurate readout, paragraph 1; Table S2How it will be judged
Decoding as in the opening comment; values from the Table S2 "Average" row. The text says "During the sustained wall-passing period" while the Methods, the code (summary_period 3000 to 11000 ms) and the Table give the 3 to 11 s contact period. Readings: window (i) 3 to 11 s (main), (ii) 4.5 to 9.5 s; cross-validation (x) 10-fold (Methods), (y) 15 splits (code). Reproduced when the order map > combined > proximity > suppressed-only > untuned > shuffled holds and all six values are within tolerance under one reading; partially reproduced when the order holds with the map group first but values are out of tolerance.
- fig3c-decoding-by-distance
At the two closest distances (7 and 9 mm) the combined map plus proximity and proximity groups decode most accurately, whereas the map group is most accurate at 11 to 23 mm.
At the closest distances (7 and 9 mm), the combined "map + proximity" and proximity populations achieved the highest accuracy. In contrast, the map population was most accurate at decoding more distal wall positions (11–23 mm)
Results, The map code in PrV allows more accurate readout, paragraph 2; Figure 3B-C; Table S2How it will be judged
Per-distance accuracy as in the opening comment, with the readings of fig3-overall-decoding-accuracy. Tolerances are 10% of the Table S2 values; the Table's 95% CIs over 10 subsamples (0.1 to 3 points) are left aside because a rerun draws new subsamples. Reproduced when, under one reading, (a) at 7 and at 9 mm the combined and proximity groups both exceed the map group, (b) at each of 11 to 23 mm the map group exceeds every other group, with the per-distance rank-sum tests (BH-FDR over nine distances) of map against proximity and map against combined giving q < 0.05 at 11 to 23 mm, and (c) the five values are within tolerance. Partially reproduced when (a) and (b) hold at all but one distance.
- fig3c-decoding-error
Decoding error grows with distance; proximity, map and combined groups reach submillimetre error at 7 and 9 mm, and only the map group is submillimetre at 13 and 15 mm.
While classification error generally increased with distance, proximity, map, and the combined "map + proximity" groups showed submillimeter precision when decoding the two closest distances (7 and 9 mm; Figure 3C , right; Table S3 ). Only the map population achieved submillimeter error at the intermediate distances of 13 and 15 mm
Results, The map code in PrV allows more accurate readout, paragraph 2; Figure 3C right; Table S3How it will be judged
Error per true distance is the confusion-matrix-weighted mean of the absolute difference between predicted and true distance, averaged over time bins in the contact period (Methods; code err_by_distance). Readings as in fig3-overall-decoding-accuracy. Reproduced when, under one reading, proximity, map and combined errors are below 1 mm at 7 and 9 mm, the map error is below 1 mm at 13 and 15 mm while proximity and combined errors are not, and the four values are within tolerance.
- figs5c-single-neuron-decoders
Single-neuron decoders show the same pattern with lower accuracy than population decoders, proximity units being more accurate than map units at 7 mm and map units more accurate at 11 to 17 mm.
These results are consistent with the performance of single-neuron decoders, although their overall decoding accuracy is lower than that of population decoders
Results, The map code in PrV allows more accurate readout, paragraph 2; Figure S5B-C; Table S4How it will be judged
Per unit, spike counts in 200-ms bins over 3 to 11 s; a template (mean over training passes) per distance; nearest template in Euclidean distance; leave-two-out cross-validation (two passes per distance held out, exhaustively); accuracy the fraction of correct test passes (Methods). Tolerances are the Table S4 95% CIs across units. Per-distance proximity against map with Wilcoxon rank-sum tests, BH-FDR over nine distances. Reproduced when proximity exceeds map at 7 mm and map exceeds proximity at 11, 13, 15 and 17 mm, each with q < 0.05, every single-neuron group's mean accuracy is below the corresponding population value of fig3-overall-decoding-accuracy, and the four values are within tolerance; partially reproduced when the directions hold but some tests or values do not.
- figs5de-svm-robust
The ranking of subpopulations is robust to the decoder; with a linear SVM the map group still has the highest average accuracy (72.1%) and proximity and combined groups are best at the closest distance.
Importantly, these findings were robust to the choice of decoder (see STAR Methods , Supplemental Figure S5D , E , and Tables S2 , S3 for SVM classifier)
Results, The map code in PrV allows more accurate readout, paragraph 2; Figure S5D-E; Tables S2-S3How it will be judged
As fig3-overall-decoding-accuracy and fig3c-decoding-by-distance with the Neural Decoding Toolbox libsvm_CL, linear kernel, C = 1, one-versus-one (Methods). Reproduced when, under one reading, the map group has the highest overall accuracy, the proximity and combined groups both exceed the map group at 7 mm, the map group is highest at 13 to 23 mm, and the four values are within tolerance.
- fig4-cut-baseline-unchanged
Cutting all whiskers but one does not change baseline firing across the recorded population (7.09 against 7.10 sp/s; 195 units from 18 mice).
Across the recorded population (n = 195 units from 18 mice), acute whisker cutting did not significantly alter baseline firing rates, measured before wall-movement onset for each wall pass (full array: 7.09 ± 1.034 vs. single whisker: 7.10 ± 1.038 sp/s; Wilcoxon signed-rank test: p = 0.70)
Results, PrV map tuning can persist, paragraph 2How it will be judged
A claim of absence, judged on the paper's test, a two-sided Wilcoxon signed-rank test of per-unit mean baseline rates (-4.5 to 0 s; the code averages the per-distance means) between full-array and single-whisker conditions, at p < 0.05 (Methods; Fig4BC_WhiskerCutting_pairwise_FR_PrefDist.m, signrank on fr_base). Units are those passing the unit criteria with stable waveforms in both conditions (Methods). Reproduced when p >= 0.05; not reproduced when p < 0.05. The difference of means is reported either way.
- fig4b-cut-wall-firing-reduced
In the 86 wall-activated, distance-tuned units, firing during wall passes falls after cutting to a single whisker (11.58 against 8.18 sp/s), in many cases to below 1 sp/s.
firing rates of wall-activated, distance-tuned units (activated population; n = 86) during wall-passing periods were significantly reduced following cutting down to a single whisker (full array: 11.58 ± 1.338 vs. single whisker: 8.18 ± 2.602 sp/s, p = 4.43 × 10 −8 ; Figure 4B )
Results, PrV map tuning can persist, paragraph 2; Figure 4BHow it will be judged
Units activated and distance-tuned in the full-array condition (classification of the opening comment). The per-unit wall rate in the code is the median over distances of the per-distance mean rate (fr_wall = median of fr_mean); the Methods say mean firing rate during wall passes. Readings: (a) median over distances (code); (b) mean over all passes (Methods). Two-sided Wilcoxon signed-rank test. The full-array spread is printed as 1.338 in the text and 2.986 in the Figure 4B legend; the larger is used as the tolerance. Reproduced when p < 0.05 with lower single-whisker rates and the three values within tolerance under one reading; partially reproduced when the test holds but a value does not.
- fig4b-cut-per-unit-changes
Of the 86 activated units, 66 change firing significantly after cutting, 52 decreasing and 14 increasing.
Among the 86 activated units, 66 showed significant changes in firing rates during wall-passing periods between cutting conditions (two-sided Wilcoxon signed-rank test, p < 0.05). Of these 66 units, 52 exhibited a significant decrease in firing rates ... whereas 14 units displayed increased firing rates
Results, PrV map tuning can persist, paragraph 2; Figure 4BHow it will be judged
Per unit, two-sided Wilcoxon signed-rank test of per-pass wall rates (3 to 11 s) between conditions, passes paired in order and truncated to the shorter condition (code), direction from the sign of the median difference (code). The paper gives three corrections. Readings: (a) uncorrected p < 0.05 (Results); (b) BH-FDR across the 86 units (Methods, Data Analysis: Acute Whisker-Cutting Manipulation; code mafdr); (c) Bonferroni across units (Methods, Characterizing Wall Distance Tuning; code alpha / (2n)). The code also differs on the pass set: (i) all passes pooled over distances; (ii) the code's exact-test block over repetitions. Reproduced when the three counts are within tolerance under one reading.
- fig4c-map-units-after-cut
Of 21 map units in the full-array condition, 10 remain distance-tuned after cutting to a single whisker; of these, 2 shift their preferred distance to the closest distance (becoming proximity units), 6 shift their preferred distance while keeping map tuning, and 2 keep their preferred distance with a change in gain.
Of the 21 units categorized as map cells in the full-array condition, 10 remained significantly tuned to wall distance after cutting down to a single whisker ... Two units shifted their preferred distance to the closest distance ( Figure 4D , E ), effectively converting from map cells to proximity cells. In contrast, six units shifted their preferred distance while retaining map tuning (e.g., Figure 4F , G ), and two additional units preserved their preferred distance but changed in response gain
Results, PrV map tuning can persist, paragraph 3; Figure 4C-IHow it will be judged
Classification of the opening comment run separately on full-array and single-whisker passes. "Remained tuned" is read as in the code, activated and distance-tuned in the single-whisker condition and not silenced (activated_tuned_mask); the variant drops the activation requirement. Map after cutting means the map category in the single-whisker condition (with the map test); a variant counts any preferred distance above 7 mm. Reproduced when all six counts are within tolerance under one reading; partially reproduced when the first two are and the split of the 10 is off by at most one unit per class. Gain change is reported, not judged, as the paper gives no measure.
- fig4jk-proximity-units-after-cut
Of 45 proximity units in the full-array condition only 15 remain distance-selective after cutting; 9 stay proximity units with lower gain, 4 stay proximity units with higher gain, and 2 become map units.
Of the 45 units categorized as proximity cells before whisker cutting, only 15 remained distance-selective after cutting down to a single whisker. Of these 15 units, nine remained proximity cells with reduced response gain ... four remained proximity cells with increased response gain ... and two shifted to exhibit map tuning
Results, PrV map tuning can persist, paragraph 4; Figure 4J-KHow it will be judged
As fig4c-map-units-after-cut. Gain is not defined in the paper. Readings for lower or higher gain: (a) the unit's peak mean wall rate in the single-whisker condition below or above that in the full-array condition; (b) its mean wall rate over distances below or above. Reproduced when all five counts are within tolerance under one reading.
- figs5fg-cut-decoding
Whisker cutting lowers population decoding accuracy in both map (31.0 to 21.5%) and proximity (24.1 to 20.2%) populations of 21 units, with the largest decreases at intermediate and far distances in the map population and less effect at the closest distances.
Whisker cutting reduced overall decoding accuracy in both populations (map: 31.0 ± 0.10% vs. 21.5 ± 0.08%; proximity: 24.1 ± 0.82% vs. 20.2 ± 1.22%; mean ± 95% CI for full-array vs. single-whisker conditions), with the largest decreases occurring at intermediate and far distances in the map population
Results, PrV map tuning can persist, paragraph 5; Figure S5F-GHow it will be judged
MCC decoders as in the opening comment, trained and tested within each condition on units recorded in both, grouped by full-array category; map n = 21, proximity 21 of 45 drawn 10 times (Methods). Tolerances are 10%: the printed CIs (0.08 to 1.2 points) describe the spread over subsamples, not a rerun's error. Readings for cross-validation as in fig3-overall-decoding-accuracy. Reproduced when, under one reading, accuracy falls in both populations, the map population's largest per-distance drop is at 11 mm or farther and its mean drop per distance over 11 to 23 mm exceeds its mean drop over 7 and 9 mm, and the four values are within tolerance; partially reproduced when the falls and the distance pattern hold but values do not.
- figs6-opto-local-inhibition-gain
Optogenetic silencing of local PrV inhibitory neurons (GtACR in vGat-Cre mice) strongly inhibits 18 of 112 units (putative inhibitory neurons); among 71 distance-tuned units, 17 are inhibited and 16 disinhibited by light, and disinhibited units keep their preferred distance, so local inhibition affects gain without changing distance preference.
Among the 112 well-isolated units, 18 were putative inhibitory neurons expressing GtACR, as they were strongly inhibited during the "Laser ON" period (modulation index ≤ −0.5 ... Of distance-tuned units (71/112), 17 were opto-inhibited ... Sixteen units showed a reduction in inhibition under the "Laser ON" condition (modulation index > 0 ... Nonetheless, these opto-disinhibited neurons, regardless of whether they were proximity or map cells, retained their preferred distance
Results, Inhibition contributes to the PrV map code, paragraph 2; Figure S6B-GHow it will be judged
MI = (FR on - FR off) / (FR on + FR off) of mean wall-period rates; a unit is modulated when its bootstrap CI on MI excludes zero; strongly inhibited MI <= -0.5, inhibited -0.5 < MI < 0, disinhibited MI > 0 (Methods). The paper does not give the CI level or what is resampled; readings, 95% percentile CI from 10,000 resamples (the paper's bootstrap elsewhere) of (a) passes within each laser condition, (b) passes within each condition and distance. Tuning is classified at the five distances (7 to 23 mm in 4-mm steps); the paper does not say on which passes, so the readings are (i) laser-off passes and (ii) all passes. Reproduced when the five counts are within tolerance under one reading and the disinhibited tuned units keep their preferred distance with laser on as off; the paper says they retained it but gives no number, so the readings are all of them and more than half of them, and the share is reported. The 17 inhibited units are counted among tuned units with MI below 0 (strongly or not); whether the 18 strongly inhibited are inside the 17 is not said, so both counts are compared as printed.
- figs6i-telc-kinematics-unchanged
Chronic TeLC expression in SpVi does not change whisker kinematics during wall passes (whisker-array midpoint angle; mean difference 0.84 degrees, Wald t = 1.02, p = 0.31).
Chronic TeLC expression in SpVi did not significantly alter whisker kinematics during wall passes, as measured by whisker array midpoint angle (mean difference across distance = 0.84°; two-tailed Wald's t -test: t = 1.02, p = 0.31
Results, Inhibition contributes to the PrV map code, paragraph 4; Figure S6IHow it will be judged
A claim of absence, judged on the paper's test (Methods, Whisker Kinematics, Wild-Type vs SpVi-TeLC). Per pass, the change in midpoint angle averaged across tracked whiskers per frame; fitlme by maximum likelihood, theta_mid ~ TeLC + distance + (1|subject) + (1|session), two-sided Wald test of TeLC at p < 0.05. Reproduced when p >= 0.05, and no per-distance model theta_mid ~ TeLC + (1|subject) + (1|session) gives p < 0.05 (Figure S6I legend); not reproduced otherwise. The effect size is reported either way. The legend's group means carry its printed spread as tolerance.
- telc-unit-yield
In SpVi-TeLC mice, 587 of 665 sorted PrV units from 11 mice passed the unit criteria, of which 395 were wall-responsive and 252 wall-activated and distance-tuned.
in TeLC experiments, 587 of 665 units (11 mice)
Methods, Isolation of Single Unit Activity; Figure 5C, E legendsHow it will be judged
Unit criteria and classification as in the opening comment (the same code, Fig5CE_TeLC_heatmap_stacked_bar.m). Reproduced when all four values are within tolerance under the Methods' responsiveness test (the code's BH-FDR variant shown).
- fig5d-telc-preferred-closer
Wall-activated, distance-tuned units in SpVi-TeLC mice prefer closer distances than those in wild-type mice.
Comparison of normalized tuning curves of all wall-activated, distance-tuned units revealed a reduction in the proportion of map units in SpVi-TeLC mice ( Figure 5C ) relative to wild-type mice ( Figure 2D ), shifting the population to prefer closer distances ( Figure 5D )
Results, Inhibition contributes to the PrV map code, paragraph 4; Figure 5C-DHow it will be judged
Preferred distance (maximum of the mean tuning curve, binned to the nine centres) of activated, distance-tuned units; one-tailed two-sample K-S test that TeLC preferred distances are smaller (Fig5D_TeLC_Pref_Dist_CDF.m, kstest2 with Tail smaller, wild type first). Population readings: (a) activated and distance-tuned (Figure 5D legend, "same units in C"); (b) activated without the tuning requirement (the code's mask1 and mask2). Reproduced when p < 0.05 under reading (a); the result under (b) is shown as a variant.
- fig5e-telc-category-shift
Among wall-responsive units, SpVi-TeLC mice have nearly threefold fewer map units than wild-type mice (5.8 against 15.4%), more proximity units (47.9 against 34.6%) and modestly fewer suppressed-only units (10.9 against 15.5%).
In particular, the proportion of map units was nearly threefold lower in SpVi-TeLC mice than in wild-type mice (5.8% vs 15.4%; Z -statistic: 4.721, p = 1.2×10 −6 ), whereas the proportion of proximity units increased (47.9% vs 34.6%; Z -statistic: −4.33, p = 7.5×10 −6 ) ... The proportion of suppressed units decreased modestly (10.9% vs 15.5%; Z -statistic: 2.15, p = 0.02)
Results, Inhibition contributes to the PrV map code, paragraph 4; Figure 5EHow it will be judged
Shares of the five categories among wall-responsive units in each group; pooled two-proportion Z = (p_WT - p_TeLC) / SE, one-tailed in the paper's direction (map and suppressed-only lower, proximity higher in TeLC; Methods; Fig5CE code). The Methods' denominator is "the total number of recorded units" while the legend and the code use wall-responsive units; readings (a) wall-responsive units (768 and 395, main), (b) all units passing criteria (1141 and 587). Reproduced when the three tests give p < 0.05 in the stated directions and the three TeLC shares are within tolerance under reading (a); partially reproduced when the map and proximity tests hold but the suppressed-only test or a share does not.
- fig5e-telc-untuned-ambiguous-unchanged
The shares of untuned (25.3 against 23.0%) and ambiguous (10.1 against 11.5%) units do not differ between SpVi-TeLC and wild-type mice.
whereas the proportions of untuned (25.3% vs 23.0%; Z -statistic: −0.86, p = 0.39) and ambiguous units (10.1% vs 11.5%; Z -statistic: 0.62, p = 0.53) remained unchanged
Results, Inhibition contributes to the PrV map code, paragraph 4; Figure 5E legendHow it will be judged
A claim of absence, judged on the paper's test, two-tailed pooled two-proportion Z-tests (Figure 5E legend; code) on shares of wall-responsive units, at p < 0.05. Reproduced when neither test gives p < 0.05; not reproduced when either does. The difference in share is reported for each.
- figs6j-telc-untuned-higher-firing
Untuned units in SpVi-TeLC mice fire more than in wild-type mice at baseline (10.01 against 6.93 sp/s) and during wall passes (9.64 against 6.32 sp/s).
During both baseline and wall-passing periods, untuned units in SpVi-TeLC mice exhibited significantly higher firing rates than those in the wild-type population, consistent with loss of long-range inhibitory inputs from SpVi (baseline: 10.01 ± 2.431 vs 6.93 ± 0.947 sp/s, p = 3.98 × 10 −5 ; wall passes: 9.64 ± 2.772 vs. 6.32 ± 0.953 sp/s, p =9.64 × 10 −4 ; one-tailed K-S tests
Results, Inhibition contributes to the PrV map code, paragraph 5; Figure S6JHow it will be judged
Per-unit mean baseline rate and wall rate averaged across distances; one-tailed two-sample K-S tests (TeLC higher), significant at 0.0125 (Figure S6K legend, four comparisons). The baseline p-value is printed as 3.98 x 10^-5 in the text and 3.98 x 10^-3 in the legend; both are below 0.0125. Untuned population readings: (a) wall-responsive but not distance-tuned (177 wild type, the population whose 6.93 and 6.32 sp/s match Figure S4); (b) the Table S5 "untuned" sets (550 wild type, 294 TeLC), every unit outside the activated-tuned and suppressed-only groups. Reproduced when both tests give p < 0.0125 and both TeLC means are within tolerance under one reading.
- figs6k-telc-tuned-baseline-similar
Wall-activated, distance-tuned units in SpVi-TeLC mice have baseline firing comparable to wild type (5.60 against 5.46 sp/s).
In contrast, distance-tuned units in SpVi-TeLC mice showed baseline activity comparable to that of wild-type mice in the absence of wall contact (5.60 ± 0.845 vs. 5.46 ± 0.670 sp/s; one-tailed K-S test: p = 0.14)
Results, Inhibition contributes to the PrV map code, paragraph 5; Figure S6K leftHow it will be judged
A claim of absence, judged on the paper's test, a one-tailed two-sample K-S test (TeLC higher) of per-unit mean baseline rates of activated, distance-tuned units, at the legend's Bonferroni level 0.0125. Reproduced when p >= 0.0125; not reproduced when p < 0.0125. The result at 0.05 and the difference of means are reported as well.
- figs6k-telc-tuned-wall-higher
During wall contact, wall-activated, distance-tuned units in SpVi-TeLC mice fire more than in wild type (9.96 against 9.02 sp/s, averaged across distances).
but exhibited significantly greater wall-evoked firing (9.96 ± 1.209 vs. 9.02 ± 1.005 sp/s; one-tailed K-S test: p = 9.34 × 10 −3
Results, Inhibition contributes to the PrV map code, paragraph 5; Figure S6K rightHow it will be judged
Per-unit wall rate (3 to 11 s) averaged across the nine distances; one-tailed two-sample K-S test (TeLC higher). Reproduced when p < 0.0125 and both means are within tolerance; partially reproduced when 0.0125 <= p < 0.05 or the test holds with a mean out of tolerance.
- figs6l-telc-increase-at-close-distances
The TeLC increase in firing of activated, distance-tuned units is distance dependent, largest at 7 mm (4.74 sp/s) and significant only at 7 and 9 mm, whereas untuned units show a roughly uniform increase of about 2 sp/s, significant at every distance.
This enhancement was distance dependent, with the largest increase at the closest wall distance ( Supplemental Figure S6L , bottom; Table S5 ). Thus, SpVi inhibition appears to be strongest at close distances for distance-tuned units
Results, Inhibition contributes to the PrV map code, paragraph 5; Figure S6L; Table S5How it will be judged
Per distance, per-unit mean wall rate; TeLC minus wild-type difference of means with a bootstrap 95% CI, and a two-tailed Wilcoxon rank-sum test at the Bonferroni level 0.00556 (Table S5; Figure S6L legend). Tolerances are the larger half of each Table S5 CI around its mean. Populations as Table S5 (tuned 472 and 252; untuned 550 and 294). Reproduced when, for tuned units, the largest difference is at 7 mm, the tests are significant at 7 and 9 mm and at no distance from 11 to 23 mm, and, for untuned units, every distance is significant, with the four values within tolerance; partially reproduced when the tuned-unit difference is largest at 7 mm and significant there but the pattern elsewhere differs.
- fig6-model-telc-loses-map
A feedforward model in which PrV neurons sum sparse sigmoidal TG excitation and SpVi-relayed inhibition yields both proximity-like and map-like tuning across 1,225 model neurons; halving long-range inhibition loses map units and expands proximity units.
When long-range inhibitory input to each PrV neuron was reduced by 50% to simulate TeLC-mediated silencing of SpVi inhibition, the model showed a loss of map units and a corresponding expansion of proximity units ( Figure 6D – F , right)
Results, Inhibition contributes to the PrV map code, paragraphs 6-7; Figure 6C-FHow it will be judged
A simulation, not a test on recorded data. Run PrV_simulation/simulate_prv.py at the commit with its defaults (M1 = 20, seed 42, p_untuned 0.4, long_inh_gain 10, local_inh_gain 1.7), without and with --telc (telc_inh_scale 0.5); classification as in the Methods (untuned when max - min < 0.1 over nine bins, suppressed when max below zero, proximity when the peak is at bin 0, map when it is later). Reproduced when the map share of distance-tuned model neurons falls and the proximity share rises with --telc, and the two tuned counts are within tolerance. The Methods' equations and the code's defaults are both recorded; where they differ (the code adds local excitation and inhibition terms the Methods do not write), the code is run, as only it gives the figure's numbers.
- fig6ghi-tuning-shape-telc
Compared with wild type, activated, distance-tuned units in SpVi-TeLC mice show little change in tuning width, a modest increase in tuning amplitude and a pronounced shift of monotonicity toward proximity-like (near-preferring) tuning, a pattern the model with reduced inhibition reproduces.
This simulated manipulation also recapitulated the experimental population-level changes in tuning curve shape ( Figure 6G – I ; see STAR Methods ): tuning width changed relatively little, tuning amplitude showed a modest rightward shift, and monotonicity exhibited a pronounced leftward shift toward more proximity-like responses
Results, Inhibition contributes to the PrV map code, paragraph 7; Figure 6G-IHow it will be judged
Experimental part, on 472 wild-type and 252 SpVi-TeLC activated, distance-tuned units (Methods, Distance Tuning Curve Shape; Fig6GHI_left_TC_shape_analysis.m). Width is the full width at half maximum above the 4.5-s baseline, in mm, by linear interpolation, truncated at the sampled edges, undefined units dropped; amplitude is peak minus baseline in sp/s; monotonicity is (far-edge minus near-edge response) / peak after subtracting the curve's minimum. The paper reports no test. Readings for "shift": (a) difference of medians; (b) difference of means. Reproduced when, under both readings, TeLC monotonicity is lower than wild type and TeLC amplitude is higher, and the width shift, expressed as a fraction of the pooled interquartile range, is smaller than the monotonicity shift so expressed. The model part (Fig6GHI_right_model_TC_shape_analysis.m, simulation outputs) is reported, not judged.
Pipeline
- Paper
- done (1 of 1)
- Claims
- done
- Claims check
- done (ready)
- Attempt
- todo
- Questions
- none
- Verified
- none