CRCNS:hc-16

These genotypic differences were unlikely to be due to differences in locomotion because the changes in locomotion across the two halves of the session were not distinct between the WT and Fmr1-null mice

Results, Abnormal Spike-Field Coordination, paragraph 1; Results, Reduced Spatial Discharge Coordination, paragraph 3; Figure S5B

What the paper reports

• genotype: two-way genotype x epoch ANOVA, F(1, 27) = 0.94, p = 0.94
• epoch: F(1, 27) = 7.39, p = 0.01
• genotype x epoch interaction: F(1, 27) = 2.16, p = 0.15

34 stationary recordings. Recordings as units, genotype F(1,32) = 0.24, p = 0.63, interaction F = 0.18, p = 0.67, half F = 22.8, p < 0.0001 (WT 6.63 to 5.86, Fmr1-null 6.31 to 5.66 cm/s). Mice as units, genotype p = 0.44, interaction p = 0.32, half p = 0.0058; one recording per mouse and day (25), interaction p = 0.20. Within each environment alone the interaction is significant, in opposite directions; in the large box WT slows more (1.37 against 0.14 cm/s, p = 0.021, 15 mice), in the stationary arena Fmr1-null does (1.01 against 0.25 cm/s, p = 0.0076 with recordings, 0.024 with 10 mice). No set of recordings here gives the 29 units the paper's 27 error degrees of freedom imply (34, 25, 15, 19, 41 recordings or 15 mice). Which recordings Figure S5B uses is a question for the paper's authors (see findings). Linearity is not judged.

How the test differed from the paper’s method

• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• Speed from room-frame positions smoothed over 1 s, a standard choice.
• Assumption. The large box and stationary arena recordings are pooled, as the STAR Methods put all four environments under "Constant Conditions Open-Field" and the Figure S5B legend speaks of "the open field recordings". Within one kind of environment the interaction is significant, in opposite directions.
• The recordings are all stationary sessions of the data set, which the paper does not list.

A lower proportion of these were classified as interneurons in WT mice (28, 12.4%) than Fmr1-null mice (98, 29.3%; test of proportions z = 9.4, p = 10�17)

Results, Non-spatial and Spatial Single-Cell Discharge Features, paragraph 1

What the paper reports

• test of proportions, z = 9.4, p = 10⁻¹⁷ (n = 226 WT and 334 Fmr1-null cells)

Interneurons (rate above 2 AP/s) are 68 of 203 WT clusters (33.5%) and 106 of 238 Fmr1-null clusters (44.5%) in R1, and 35.4% and 48.5% in R2; with the classified cells as denominator 38.6% to 41.4% and 51.7% to 55.7%. Every percentage is far outside 12.4 +/- 1.24 and 29.3 +/- 2.93. The direction holds under every reading and is significant in the main ones (R1 z = -2.36, p = 0.018; R2 z = -3.33, p = 0.0009); with the refractory filter (clusters with at most 0.5% of ISIs below 2 ms) p = 0.092 in R1 and 0.023 in R2. The effect is present but the percentages are not the paper's, so partially reproduced. Only 70 of 774 clusters have more than 0.5% of ISIs below 2 ms and the filter changes the percentages by at most 4.9 points, so this filter does not remove the excess. It is a weak proxy for the isolation measures (interneuron rates are still 4.34 and 6.95 AP/s at the median in R1), so whether poorly isolated clusters explain the excess cannot be told from these data.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.
• Rate alone and rate with bursting give the same interneurons, as the rate decides it.
• A refractory filter (at most 0.5% of ISIs below 2 ms, a standard choice) is shown as a proxy for the isolation measures the data set lacks.

The fundamental extracellular waveform and non-spatial discharge characteristics of the classified cells were indistinguishable between the genotypes

Figure S3C to S3F; Results, Non-spatial and Spatial Single-Cell Discharge Features, paragraph 1

What the paper reports

• action potential width, genotype: two-way genotype x cell class ANOVA, F(1, 386) = 0.97, p = 0.32
• burst probability, genotype: two-way genotype x cell class ANOVA, F(1, 386) = 0.69, p = 0.41
• interspike interval, genotype: two-way genotype x cell class ANOVA, F(1, 386) = 2.78, p = 0.1
• overall firing rate, genotype: two-way genotype x cell class ANOVA, F(1, 386) = 0.32, p = 0.57

Two-way genotype x class ANOVA on 381 classified clusters (R1). No genotype effect for burst probability (p = 0.13) or the interspike interval (mean p = 0.40, median p = 0.63), but one for the overall rate (F = 8.28, p = 0.0042, partial eta squared 0.022), from faster Fmr1-null interneurons (11.5 against 7.3 AP/s). In R2 (547 clusters) burst probability also differs (p = 0.029) and the rate (p = 0.0005). With the refractory filter the rate effect is not significant (R1 p = 0.070, R2 p = 0.057), so that no measure differs in R1, but burst probability differs in R2 (p = 0.0037). Waveform width is not in the data set and is not tested.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.
• Waveform width cannot be measured; the claim is judged on three of its four measures.

the fundamental measures of spatial discharge quality did not differ between either the WT and the Fmr1-null place cells or interneurons (Table S1)

Table S1; Results, Non-spatial and Spatial Single-Cell Discharge Features, paragraph 1

What the paper reports

• place cell firing rate: unpaired t test, t(160) = 1.19, p = 0.23 (n = 73 WT and 89 Fmr1-null place cells)
• in-field rate: t(160) = 1.74, p = 0.08
• spatial coherence: t(160) = 1.22, p = 0.23
• information content: t(160) = 0.45, p = 0.65
• proportion active pixels: t(160) = 0.35, p = 0.73
• interneuron firing rate: t(67) = 0.18, p = 0.85 (n = 13 WT and 56 Fmr1-null interneurons)

Large-box place cells, 34 WT and 36 Fmr1-null. No place cell measure differs (rate p = 0.49, in-field rate p = 0.15 and 0.69, coherence p = 0.39, information p = 0.47, active pixels p = 0.44, by t test or Mann-Whitney as Shapiro-Wilk chooses). The interneuron rates differ (WT 5.40, Fmr1-null 10.34 AP/s, Mann-Whitney p = 0.0105). No mean is within the Table S1 SEMs (information 2.32 and 2.27 bits/spike against 1.57 and 1.5; active pixels 0.24 and 0.26 against 0.45 and 0.46). On the 32 x 32 grid (40 and 46 place cells) no place cell measure differs (p = 0.12 to 0.64) and information (1.74, 1.74) and active pixels (0.40, 0.41) come nearer Table S1 without reaching it; with the refractory filter the interneuron rates still differ (p = 0.036).

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• Only the large box is in the data set.

This greater spatial discharge reliability of individual Fmr1-null place cells is accompanied by significantly less spatial discharge covariance (Figure 4A) of the Fmr1-null cell pairs with overlapping place fields and at least 30 passes through both firing fields

Results, Reduced Spatial Discharge Coordination, paragraph 2; Figure 4A and 4B

What the paper reports

• t test, t(356) = 2.39, p = 0.02 (n = 109 WT and 249 Fmr1-null pairs)
• Mann-Whitney, U = 12,424, p = 0.2
• after random permutation of z values within pairs: t test, t(356) = 0.8, p = 0.4

R2. 122 WT and 210 Fmr1-null pairs with at least 30 joint passes; WT r = 0.061 +/- 0.020, Fmr1-null 0.021 +/- 0.010, t on Fisher z = 2.29, p = 0.023 (Mann-Whitney p = 0.26). WT is within tolerance, Fmr1-null is not (0.04 +/- 0.01). After one permutation per pair the difference is gone (p = 0.21; 100 permutations p = 0.22). R1 gives p = 0.033. Under this literal pass rule every pair has at least 39 joint passes, so the 30 pass minimum selects nothing. With the count rule (joint passes 3 to 129, median 28) 50 WT and 103 Fmr1-null pairs have at least 30; r = 0.165 against 0.007, t = 3.54, p = 0.0005, neither average within tolerance; after 100 permutations p = 0.51, after one permutation per pair p = 0.0004. That draw is the most extreme of the 100 permutations taken one at a time (|t| 3.61, median 0.68), of which 4 give p < 0.05, as a null should. The outcome is the same under both rules. With the refractory filter p = 0.023; on the 32 x 32 grid 183 WT and 290 Fmr1-null pairs, r = 0.074 against 0.023, p = 0.0009, the Fmr1-null average outside tolerance under either rule.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.
• The pass rule is read literally, as the claims file does (exp, a count, at least the cell's average rate in AP/s); it passes a median 76% of a cell's 5 s intervals. The count expected at the average rate over 5 s is shown as a variant; which is meant is asked of a reader (questions/crcns-hc-16.yaml).

In fact, the difference is maintained when at least 40 qualified passes were required

Results, Reduced Spatial Discharge Coordination, paragraph 2; Figure 4B

What the paper reports

• t test, t(227) = 4.2, p = 10⁻⁴ (n = 56 WT and 173 Fmr1-null pairs)
• Mann-Whitney, U = 3,791, p < 0.01

The fewest joint passes of any pair is 39, so the sample hardly changes, 122 WT and 209 Fmr1-null pairs; r = 0.061 against 0.020, t = 2.33, p = 0.021. The difference holds but is not larger, and both averages are outside tolerance (0.19 +/- 0.04, 0.06 +/- 0.01). With the count rule the 40 pass minimum selects (30 WT, 86 Fmr1-null pairs) and the difference is larger, as claimed, r = 0.279 against 0.039, t = 4.41, p < 0.0001, but neither average is within tolerance; the outcome is the same under both rules on 1.25 cm maps. With the refractory filter p = 0.021. On the code's 32 x 32 grid, under the literal rule both averages are outside tolerance (0.074, 0.023); under the count rule 65 WT and 121 Fmr1-null pairs give r = 0.182 against 0.056, both within tolerance, t p = 0.0004, which meets the criterion. That combines two choices the claims file does not make (the count rule, asked of a reader, and the grid), so the outcome stays on the claims file's rule and maps.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.
• The pass rule is read literally, as the claims file does (exp, a count, at least the cell's average rate in AP/s); it passes a median 76% of a cell's 5 s intervals. The count expected at the average rate over 5 s is shown as a variant; which is meant is asked of a reader (questions/crcns-hc-16.yaml).

Room-specific Ipos is greater in Fmr1 null compared to WT (Mann-Whitney U = 208, p < 0.05, d = 0.45) whereas arena-specific Ipos is only marginally greater in Fmr1 null

Figure 5B; Results, During Dissociation of Spatial Frames, paragraph 1

What the paper reports

• room frame: Mann-Whitney, U = 208, p < 0.05 (n = 20 WT and 31 Fmr1-null place cells)
• arena frame: Mann-Whitney, U = 213, p = 0.06

Room-frame Ipos is higher in Fmr1-null under both readings of which session classifies the cells, significantly in one, and that one rests on one mouse. Reading B (place cells classified in the stationary session of the same day, 10 WT, 8 Fmr1-null) gives room-frame Ipos 0.106 (Fmr1-null) against 0.066 (WT) bits, Mann-Whitney U = 12, p = 0.012, and arena-frame 0.097 against 0.063, p = 0.016, where the paper finds the arena difference only marginal; leaving out M37 (3 of 8 Fmr1-null cells) gives p = 0.13, leaving out any other mouse p < 0.05. Reading A (place cells classified in the rotating session, 9 WT, 7 Fmr1-null) gives the same direction without significance (0.103 against 0.084, p = 0.25). The cells come from 3 mice per genotype. Which session classifies the rotating-arena place cells is a question for the paper's authors (see findings). The measured values are reading B's; reading A gives 0.084 (WT) and 0.103 (Fmr1-null) bits.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• Both readings of which session classifies the place cells are run (A, the rotating session; B, the same-day stationary session); the paper does not say which classifies the cells of Figure 5.
• Pixels with fewer than 10 position samples are left out, as the PositionalInfo code counts visits per position sample.
• An Ipos interval is four position samples (133 ms at 30 Hz), located at its first sample, a standard choice.
• 20 WT and 31 Fmr1-null place cells in the paper; 10 and 8 here.

The session-averaged, frame-specific spatial firing was better organized in the room frame, and this preference was similar for the two genotypes (Figure S6)

Results, During Dissociation of Spatial Frames, paragraph 1; Figure S6B

What the paper reports

• map coherence, environment effect: two-way genotype x environment ANOVA, repeated measures on environment, F(3, 147) = 9.02, p = 1.6 × 10⁻⁵
• information content, environment effect: two-way genotype x environment ANOVA, repeated measures on environment, F(3, 147) = 4.17, p = 0.046

Reading B, 10 WT and 7 Fmr1-null place cells with all four maps. Coherence has an environment effect (F(3,45) = 6.24, p = 0.0012) and no interaction (p = 0.58); Tukey puts the rotating arena frame below stationary arena I (0.29, adjusted p = 0.011) but not below stationary arena II (0.12, p = 0.53; paired t p = 0.056). Log information has an environment effect (p = 0.0009), no interaction (p = 0.13). In reading A (9 and 6 cells) the coherence environment effect is p = 0.067, so under reading A the claim would not be reproduced. The outcome rests on reading B, which settles a question the record leaves open for fig5b-frame-specific-ipos.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• Assumption. The place cells are those classified in the same-day stationary session (reading B), as the Results classify cells in standard recordings (Results, Non-spatial and Spatial Single-Cell Discharge Features, paragraph 1); the paper does not say which session classifies the cells of Figure S6, and reading A gives an environment effect on coherence of p = 0.067. This settles for this claim a question the record leaves open elsewhere; fig5b-frame-specific-ipos is judged on both readings, and the question is put to the paper's authors (see findings).
• Tukey's HSD is applied across the four environments without the repeated structure; paired t tests are shown as a variant.
• M29 drops out, as its stationary arena 1 has another cut file, so its cells cannot be matched across the four maps.

when the putative principal cells were classified as place cells (WT n = 108 [47.8%], Fmr1 null n = 141 [42.2%]), the prevalence was lower in the Fmr1-null mice (test of proportions z = 2.04, p = 0.02)

Results, Non-spatial and Spatial Single-Cell Discharge Features, paragraph 1

What the paper reports

• test of proportions, z = 2.04, p = 0.02

Place cells are 48 of 203 WT (23.6%) and 51 of 238 Fmr1-null (21.4%) clusters in R1, and 55 (18.9%) and 70 (20.5%) in R2, against 108 (47.8%) and 141 (42.2%). As a share of principal cells, WT 44.4% and Fmr1-null 51.5% (R1), 37.7% and 53.0% (R2). No one-sided test over 26 readings, including single-cell measures on the code's 32 x 32 grid and the refractory filter, finds Fmr1-null lower (smallest one-sided p 0.27); with principal cells in R2 Fmr1-null is higher (two-sided p = 0.010; on the 32 x 32 grid p = 0.005).

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• Rate maps are unsmoothed on 1.25 cm pixels, with any occupancy, as the claims file sets. The code's 32 x 32 grid of the tracker frame, which needs no spatial scale, is a variant; 45 of 774 clusters change class on it.

Spatial firing quality differs between place cells and interneurons, but not between genotypes, as assessed by two-way genotype 3 cell class ANOVAs

Figure 2B to 2F; Results, Non-spatial and Spatial Single-Cell Discharge Features, paragraph 1

What the paper reports

• spatial coherence, genotype: two-way genotype x cell class ANOVA, F(1, 340) = 0.60, p = 0.44
• proportion of active pixels, genotype: F(1, 340) = 3.7, p = 0.06
• information content, genotype: F(1, 340) = 3.6, p = 0.06
• number of place fields, genotype: F(1, 340) = 0.13, p = 0.72
• in-field firing rate, genotype: F(1, 340) = 0.24, p = 0.62

Large box, 213 classified clusters. Genotype effects for coherence (F = 9.01, p = 0.0030), active pixels (F = 18.7, p < 0.0001), information (F = 7.27, p = 0.0076) and the number of fields above the mean rate (p = 0.011), with interactions for coherence and active pixels (p = 0.0077, 0.0098), mostly from interneurons. In-field rate (p = 0.11, 0.42) and the count of fields above 20% of the peak (p = 0.48) show none. The class effect is significant for every measure (p <= 0.0004). The genotype effects on coherence, active pixels and information remain on the code's 32 x 32 grid (p = 0.0003, 0.0016, 0.0073) and with the refractory filter (p = 0.0084, 0.0001, 0.013).

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• Only the large box is in the data set; the small-box and both-box readings cannot be run.

Not quoted: Numbers and tests from Table S1; no sentence of the paper states this difference.

What the paper reports

• unpaired t test, t(113) = 2.37, p = 0.02 (n = 58 WT and 57 Fmr1-null non-spatial pyramidal cells)
• Mann-Whitney, U = 1,219, p = 0.02

Large-box non-spatial pyramidal clusters discharge in 0.152 (WT, n = 29) and 0.222 (Fmr1-null, n = 35) of the visited pixels; Fmr1-null is higher but Mann-Whitney p = 0.10 (both fail Shapiro-Wilk), and neither value is within tolerance. R1 and R2 give p = 0.09 to 0.23. On the 32 x 32 grid both values are within tolerance (0.241 and 0.355) but Mann-Whitney p = 0.099; with the refractory filter p = 0.091.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.

Overdispersion was lower (spatial firing was more reliable) in Fmr1-null mice when the unit of analysis is a pass through a place field

Results, Reduced Spatial Discharge Coordination, paragraph 1

What the paper reports

• F test, F(6775, 9101) = 1.19, p = 10⁻¹⁷ (n = 6,776 WT and 9,102 Fmr1-null passes)
• Levene's test, F(1, 15874) = 22.21, p = 10⁻⁶

R2, consecutive 5 s intervals. 12,802 WT and 18,521 Fmr1-null passes, variance of z 3.69 and 3.55 (expected 5.81 and 4.89). The F test finds WT larger (F = 1.04, one-sided p = 0.008) but Levene's test does not (p = 0.998). In R1 the Fmr1-null variance is the larger (3.83 against 3.53). With a 1 s step, F one-sided p = 0.007 and Levene p = 0.61. With the count rule (exp at least the average rate x 5 s, 33% of intervals pass) R2 gives 5,442 and 8,250 passes, variances 5.28 and 4.66, both within tolerance, F = 1.13 (one-sided p < 0.0001) and Levene p = 0.0006, which meets the criterion (the WT pass count is outside 6,776 +/- 678); in R1 it gives Fmr1-null the larger variance (5.16 against 5.00). The outcome stays on the claims file's literal rule; the rule is asked of a reader. Under the literal rule neither the refractory filter (variances 3.71 and 3.55, Levene p = 0.86) nor the code's 32 x 32 grid (3.73 and 3.73, F one-sided p = 0.49) changes the result. Under the count rule the filter leaves both variances within tolerance (Levene p = 0.0004); on the grid variances (5.34, 4.89) and pass counts (6,712, 9,607) are within tolerance, F one-sided p = 0.0001, but Levene p = 0.052.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.
• Levene's test is centred on the mean, as the original test is.
• The pass rule is read literally, as the claims file does (exp, a count, at least the cell's average rate in AP/s); it passes a median 76% of a cell's 5 s intervals. The count expected at the average rate over 5 s is shown as a variant; which is meant is asked of a reader (questions/crcns-hc-16.yaml).

Overdispersion was also marginally lower in Fmr1-null mice when the unit of analysis is a place cell

Results, Reduced Spatial Discharge Coordination, paragraph 1

What the paper reports

• t test, one-tailed, t(176) = 1.87, p = 0.03 (n = 80 WT and 98 Fmr1-null place cells)
• Mann-Whitney, one-tailed, U = 3,386, p = 0.06

Every place cell has at least 30 passes, so the three minimums give one sample. R2, WT 3.68 +/- 0.24 (n = 55) and Fmr1-null 3.53 +/- 0.21 (n = 70), one-tailed t p = 0.32; R1, 3.56 against 3.72 (p = 0.67). Neither mean is within tolerance. With the count rule (at least 49 passes per cell) R2 gives 5.33 +/- 0.46 against 4.69 +/- 0.33 (one-tailed p = 0.12) and R1 5.02 against 5.08 (p = 0.54); on 1.25 cm maps no reading has both means within tolerance. With the refractory filter p = 0.29 (literal) and 0.11 (count); on the 32 x 32 grid p = 0.41 (literal), and with the count rule both means are within tolerance (5.57, 5.05) but p = 0.18.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.
• The pass rule is read literally, as the claims file does (exp, a count, at least the cell's average rate in AP/s); it passes a median 76% of a cell's 5 s intervals. The count expected at the average rate over 5 s is shown as a variant; which is meant is asked of a reader (questions/crcns-hc-16.yaml).

At the 25 ms (Figures 4C and 4D) and 40 ms timescales of gamma oscillations, as well as the 125 ms timescale of theta oscillations, Fmr1-null cell pair discharge correlations (n = 504 pairs) were significantly weaker than those of WT cell pairs

Results, Reduced Spatial Discharge Coordination, paragraph 3; Figure 4C and 4D

What the paper reports

• 25 ms: Mann-Whitney, U = 51,894, p = 2 × 10⁻⁴ (n = 247 WT and 504 Fmr1-null pairs)
• 40 ms: Mann-Whitney, U = 53,512, p = 0.002
• 125 ms: Mann-Whitney, U = 55,708, p = 0.02
• 1 s: Mann-Whitney, U = 60,801, p = 0.6
• 5 s: Mann-Whitney, U = 60,313, p = 0.5

Eight readings (K1, K2 x whole session, first 5 min x R1, R2); none gives p < 0.05 at 25, 40 or 125 ms (smallest p 0.24). K1, whole session, R2 (122 WT, 210 Fmr1-null pairs) gives mean tau at 25 ms 0.029 against 0.0067, Mann-Whitney p = 0.87, 0.67 and 0.60 at 25, 40 and 125 ms, and p = 0.049 at 1 s and 0.0012 at 5 s, the reverse of the claimed pattern. Without M16, which gives 52 of the 122 WT pairs, the pattern is the same (K1, session, R2, p = 0.98, 1.00, 0.27 at 25, 40, 125 ms and 0.025, 0.0059 at 1 s, 5 s); one of the eight readings without M16 has one short timescale below 0.05 (K1, first 5 min, R1, 125 ms, p = 0.045). With the refractory filter and on the 32 x 32 grid no reading has a short timescale below 0.05 (K1, session, R2, 1 s and 5 s p = 0.043 and 0.0009 with the filter, 0.022 and 0.0011 on the grid).

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.
• Bins are placed on the spike clock; positions per bin come from sample times estimated from the spikes, since the files' nominal sample times drift.

pairwise t tests could not reject the hypothesis that the WT vector of Kendall correlations from the two halves of a recording are the same whereas the corresponding Fmr1-null comparisons were all significantly different

Results, Reduced Spatial Discharge Coordination, paragraph 3; Figure 4D legend

What the paper reports

• WT halves, 25 ms: paired t test, t(246) = 0.9, p = 0.4
• Fmr1-null halves, 25 ms: paired t test, t(503) = 3.03, p = 0.002
• WT halves, 40 ms: paired t test, t(246) = 0.6, p = 0.5
• Fmr1-null halves, 40 ms: paired t test, t(503) = 3, p = 0.003
• WT halves, 125 ms: paired t test, t(246) = 1.02, p = 0.3
• Fmr1-null halves, 125 ms: paired t test, t(503) = 1.7, p = 0.09
• WT halves, 250 ms: paired t test, t(246) = 1.0, p = 0.3
• Fmr1-null halves, 250 ms: paired t test, t(503) = 2.4, p = 0.02
• WT halves, 1 s: paired t test, t(246) = 0.9, p = 0.4
• Fmr1-null halves, 1 s: paired t test, t(503) = 2.6, p = 0.001
• WT halves, 5 s: paired t test, t(246) = 0.3, p = 0.7
• Fmr1-null halves, 5 s: paired t test, t(503) = 2.0, p = 0.046

Readings H1 and H2 with K1 and K2, pairs per session (R2) and R1. No reading has five significant Fmr1-null paired tests; under H1 none of the six reaches p < 0.05, under H2 at most one. Under H2 with K1 the WT intervals differ at 25 and 40 ms (p = 0.035, 0.016). PCo is higher in WT at all six timescales in six readings (K1, H1, R2, 25 ms, 0.991 against 0.827) and at five in the other two (K2, H2, R1 and R2), where Fmr1-null is higher at 1 s (0.519 against 0.279, 0.499 against 0.266). The refractory filter and the 32 x 32 grid were not run on this claim.

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.

Similar to place cell ensembles, Fmr1-null cell pair discharge correlations of non-spatial pyramidal cells (n = 203 pairs) were significantly weaker than those of WT cell pairs (n = 252 pairs

Results, Reduced Spatial Discharge Coordination, paragraph 4

What the paper reports

• 25 ms: Mann-Whitney, U = 22,437, p = 0.03
• 40 ms: Mann-Whitney, U = 22,318, p = 0.02
• 125 ms: Mann-Whitney, U = 21,745, p = 0.007
• 1 s: Mann-Whitney, U = 21,977, p = 0.01
• 5 s: Mann-Whitney, U = 23,333, p = 0.12

380 WT and 146 Fmr1-null non-spatial pairs (R2). With K1 over the session the distributions differ at 25, 40 and 125 ms (p = 0.0084, 0.0024, 0.0054) but Fmr1-null mean tau is not lower (0.0099 against 0.0089 at 25 ms), and 1 s does not differ (p = 0.12). No reading gives Fmr1-null lower at p < 0.05 on more than one of the four timescales. With the refractory filter two readings have two (K2, session); on the 32 x 32 grid none has any, and Fmr1-null tau is the higher (K1, session, R2, 25 ms, 0.0110 against 0.0083).

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• The session that classifies a cell ('standard recordings') is open; R1 (each cell from its first stationary session) and R2 (each stationary session on its own) are both run, and a cluster name within a day's cut file is taken as one cell.

Spike train correlations did not differ at the gamma timescale ... whereas the Fmr1-null correlation distribution was weaker at theta and 1 s timescales

Results, Reduced Spatial Discharge Coordination, paragraph 5

What the paper reports

• 25 ms: Mann-Whitney, U = 11,248, p = 0.24 (n = 125 WT pairs (35 cells, 6 mice) and 195 Fmr1-null pairs (39 cells, 4 mice))
• 40 ms: Mann-Whitney, U = 11,026, p = 0.15
• 125 ms: Mann-Whitney, U = 10,583, p < 0.05
• 250 ms: Mann-Whitney, U = 10,397, p = 0.03
• 1 s: Mann-Whitney, U = 10,629, p < 0.05
• 5 s: Mann-Whitney, U = 10,917, p = 0.12

No recording is labelled sleep or rest, so only the reading of periods below 2 cm/s runs (21% of a session's samples, median). A bin enters when the speed is below 2 cm/s at every position sample in it. 25 and 40 ms do not differ (K1, R2, p = 0.78, 0.45) as claimed, but neither do 125 ms, 250 ms and 1 s (p = 0.41, 0.56, 0.50); 5 s keeps few bins (50 WT and 45 Fmr1-null pairs, p = 0.12). Keeping a bin when the speed at its centre is low gives the same pattern (p = 0.39, 0.92, 0.78 at 125 ms, 250 ms, 1 s). Pairs are 122 WT and 210 Fmr1-null (expected 125 and 195).

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• The spatial scale is assumed from the extent of the tracked positions and the enclosure sizes of the STAR Methods (large box 50 cm, arena 40 cm); the data set gives none.
• Sleep and quiet wakefulness recordings are not in the data set; the low-speed reading is the only one run.
• Speed from room-frame positions smoothed over 1 s, a standard choice.

The association between the ensemble and the single left-out cell was weaker in Fmr1-null compared to WT place cells (Figure 5C)

Figure 5C; Results, During Dissociation of Spatial Frames, paragraph 2

What the paper reports

• t test, t(48) = 2.4, p < 0.01 (n = 20 WT and 30 Fmr1-null place cells)

Under every reading Fmr1-null cells correlate with the ensemble at least as well as WT cells. |delta Ipos| with all other clusters as the ensemble gives r = 0.071 against 0.070 (reading A, p = 0.98) and 0.076 against 0.045 (reading B, p = 0.27); delta Ipos gives Fmr1-null higher in reading B, with all clusters (0.175 against 0.115, p = 0.045) and with the other place cells (0.092 against 0.045, p = 0.0042).

How the test differed from the paper’s method

• The data set has no isolation measures (IsoIBG, IsoINN) or waveforms, so every sorted cluster other than cluster 0 is used in place of the paper's well-isolated cells, and cells are classified by rate and bursting without waveform width.
• Both place cell readings (A, B) and both ensembles (all clusters, as the Results say, and place cells) are run.

Of the 1,115 single-unit waveform clusters that were identified, 499 were recorded from 12 WT mice and 616 were recorded from 9 Fmr1-null mice (Table S2) ... 226 WT and 334 Fmr1-null presumptive cells comprised the dataset

Results, Non-spatial and Spatial Single-Cell Discharge Features, paragraph 1; Table S2

The data set holds 8 WT mice (M1, M3, M11, M16, M28, M32, M35, M36) and 7 Fmr1-null mice (M9, M10, M18, M29, M33, M34, M37) of the paper's 12 (or 11) and 9, so it is a subset and its counts cannot test the paper's. The well-isolated cells cannot be selected either, as no isolation measure is given (metadata_insufficient). For the record, the cluster counts are within 10% of 499 and 616 only when every session is counted on its own with cluster 0 included (490 WT, 557 Fmr1-null), which with 15 of 21 mice is a coincidence; counted once per cut file without cluster 0 they are 203 and 294.

How the test differed from the paper’s method

• Clusters are counted from the file names; the paper's isolation criterion cannot be applied.