eLife, Version of Record. This deposit contains the numerical values behind every quantitative panel of the paper, one file per panel, in plain CSV.

What is here

source_data_by_panel/ holds 56 files. Each one corresponds to a panel of a main figure or of a figure supplement and is named with the eLife panel designation, for example Figure_2-figure_supplement_3a_-_source_data.csv. Panels that show three fibre components (Figure 4d, 4e, 4f and 5f left) have all three in a single file, as do the three colocalisation panels of Figure 1-figure supplement 3.

Every file is plain tabular CSV: one header row of column names, then the data, with no comment lines and nothing above the header. The two sections that follow define every term and every column used anywhere in the deposit. The per-file sections at the end give only what is specific to each file: its size, what it measures, its sample size, the statistic as published, the same statistic recomputed from the table itself, and the analysis file it came from.

Terms and abbreviations

These are used in column names, in categorical values and in the per-file sections below.

Column dictionary

Every column used anywhere in the deposit, with its unit. A column carries the same meaning in every file in which it appears, so this table is the single place where the variables are defined; the per-file sections below do not repeat them.

Conventions

• Layout. Panels with a single factor are wide: one row per experimental unit, one column per condition, so the pairing is explicit. Panels with two factors are long: subject, group, condition, value, which handles unequal group sizes.
• Identifiers are positional. subject and cell number the rows of the source analysis file. Those files do not store animal or cell identifiers, so these numbers cannot be matched to individual animals. The one exception is Figure_2-figure_supplement_3f_-_source_data.csv, whose source spreadsheet does record mouse numbers; they are given there as mouse_id.
• Units are those of the source file and have not been converted. Freezing is stored as a proportion in some panels and as a percentage in others; the figures plot percentages throughout. Each file section states which.
• Signed-rank statistics. For the Wilcoxon matched-pairs signed rank test, W is the signed sum of ranks (W+ minus W-) as GraphPad Prism reports it, so its sign follows the order in which the two conditions were subtracted and carries no further meaning. Pairs with a zero difference are dropped before ranking.
• Nothing has been removed. Where the published analysis excluded a value, the row is kept and an included_in_analysis column marks it, with the reason given in the file section (Figure 4c and Figure 5f right).
• Missing values are empty cells. They occur in two files only, and the file section says which rows and why.
• Every value table was checked twice: the plot was reproduced from it, and the published test statistic was recomputed from it. The recomputed value is given for each file below. Where it differs from the published one, the reason is stated.

Completeness

Every quantitative panel of the paper is represented. Panels that are photomicrographs, histology, schematics, example traces or fibre placement maps carry no value table and are not represented here.

Files

Figure_1b_-_source_data.csv

Figure 1b. Freezing during fear retrieval in wild-type mice: baseline, CS- and CS+.

• Rows and columns: 12 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 12 mice.
• Units: freezing as a proportion (0-1), as stored; the figure plots it as a percentage.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,22) = 160.8606
• Source: "WT_FCA.svd" (StatView), variables "BL_F", "CS-_F", "CS+_F".

Figure_1e_-_source_data.csv

Figure 1e. Latency of the NC response on the FC hot-plate test, CS- vs CS+.

• Rows and columns: 12 rows; subject, latency_CSminus_s, latency_CSplus_s
• Experimental unit: mouse. n = 12 mice.
• Units: latency of the nociceptive response, in seconds.
• Published statistic: Wilcoxon matched-pairs signed rank test, P = 0.0005.
• Recomputed from this table: W = 78, P = 0.000488
• Source: "WT_FCA.svd" (StatView), variables "HP_CS-time", "HP_CS+time".

Figure_1f_right_-_source_data.csv

Figure 1f right. Average freezing during the 10 s after CS onset and during the 10 s before the NC response, CS- vs CS+.

• Rows and columns: 12 rows; subject, freezing_CSminus_at_CS_onset, freezing_CSplus_at_CS_onset, freezing_CSminus_before_NC, freezing_CSplus_before_NC
• Experimental unit: mouse. n = 12 mice.
• Units: freezing as a proportion (0-1) of the 10 s window; the figure plots it as a percentage. The left-hand part of panel 1f is a per-mouse raster of freezing probability in 0.5 s bins; it has no separate value table and is drawn from the trial-by-trial freezing time series.
• Published statistic: Wilcoxon matched-pairs signed rank test; P at CS onset = 0.0015, before the NC response ns (P = 0.8330).
• Recomputed from this table: at CS onset W = 74, P = 0.0015 (exact); before the NC response W = 6, P = 0.8311 against the published 0.8330. W is the signed sum of ranks (W+ minus W-), as reported by GraphPad Prism. In the second comparison one pair has a zero difference and is dropped, leaving 11 pairs, and several absolute differences are tied; the small residual gap on P comes from the tie handling.
• Source: "Figure 1f.prism" (GraphPad Prism), data tables 471A94A2... (at CS onset) and EEAC8FE0... (before the NC response).

Figure_1g_-_source_data.csv

Figure 1g. Freezing during fear extinction: 24 CS+ presented as six blocks across two sessions.

• Rows and columns: 10 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus_block1, freezing_CSplus_block2, freezing_CSplus_block3, freezing_CSplus_block4, freezing_CSplus_block5, freezing_CSplus_block6
• Experimental unit: mouse. n = 10 mice (the extinction subgroup of the 21 rows in the source file).
• Units: freezing as a proportion (0-1), as stored; the figure plots it as a percentage.
• Published statistic: Friedman test followed by Dunn.
• Recomputed from this table: Q = 48.1333
• Source: "WT_Extinction-FCA.svd" (StatView), variables "BL_T1", "CS-_T1", "CS+1_T1", "CS+2_T1", "CS+3_T1", "CS+1_T2", "CS+2_T2", "CS+3_T2".

Figure_1h_-_source_data.csv

Figure 1h. NC response latency after fear extinction, CS- vs CS+.

• Rows and columns: 10 rows; subject, latency_CSminus_s, latency_CSplus_s
• Experimental unit: mouse. n = 10 mice.
• Units: latency of the nociceptive response, in seconds.
• Published statistic: paired t-test, ns.
• Recomputed from this table: |t9| = 0.8143 (t9 = -0.8143 for CS- minus CS+; mean latency 202.37 s on CS- against 205.75 s on CS+). ns.
• Source: "WT_Extinction-FCA.svd" (StatView), variables "HP T CS-", "HP T CS+".

Figure_1i_-_source_data.csv

Figure 1i. NC response latency for the different hot-plate tests: no tone, unconditioned tone, CS- and CS+.

• Rows and columns: 46 rows; subject, test, latency_s
• Experimental unit: mouse. No tone n = 10, Tone n = 12, CS- and CS+ n = 12 (the same 12 mice for both CSs).
• Units: latency of the nociceptive response, in seconds.
• Layout: one row per measurement. Subject numbers identify mice: the CS- and CS+ rows carry the same numbers because they are the same twelve mice. The no-tone and tone groups are two other cohorts, numbered separately. The published test treats the four groups as independent and does not use this pairing.
• Published statistic: Kruskal-Wallis test followed by Dunn, P < 0.0001.
• Recomputed from this table: P = 7.0e-05, P < 0.0001 as published.
• Source: "WT_basal pain response.svd" ("HP t trial1"), "WT_tone-specific pain response.svd" ("T CS-") and "WT_FCA.svd" ("HP_CS-time", "HP_CS+time"), all StatView.

The source files hold two trials for the no-tone and the tone test; which of the two was used for this four-group panel is not recorded, and the first trial of each is deposited here.

Figure_1-figure_supplement_1a_-_source_data.csv

Figure 1-figure supplement 1a. Freezing during habituation and conditioning: baseline, CS- and CS+.

• Rows and columns: 12 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 12 mice.
• Units: freezing as a proportion (0-1), as stored; the figure plots it as a percentage.
• Published statistic: Friedman test followed by Dunn, Q = 16.55, P = 0.0003.
• Recomputed from this table: Q = 15.1667 without tie correction, 16.5455 with it (the published value)
• Source: "FCA_Habituation and conditioning.svd" (StatView), variables "BL", "CSM", "CSP".

Figure_1-figure_supplement_1b_-_source_data.csv

Figure 1-figure supplement 1b. Temperature of the NC response, CS- vs CS+.

• Rows and columns: 12 rows; subject, temperature_CSminus_C, temperature_CSplus_C
• Experimental unit: mouse. n = 12 mice.
• Units: infrared skin temperature at the nociceptive response, in degrees Celsius.
• Published statistic: Wilcoxon matched-pairs signed rank test, W = 78, P = 0.0005.
• Recomputed from this table: W = 78, P = 0.000488
• Source: "WT_FCA.svd" (StatView), variables "IR CS-", "IR CS+".

Figure_1-figure_supplement_1d_-_source_data.csv

Figure 1-figure supplement 1d. Skin temperature of the back across the trial, CS- vs CS+.

• Rows and columns: 260 rows; subject, CS, timepoint, temperature_C
• Experimental unit: mouse. n = 13 mice; 12 with a complete series (two values missing).
• Units: skin temperature in degrees Celsius.
• Layout: one row per mouse, timepoint and CS. Empty cells are the missing values. The source file stores the two CS series as duplicate column names: the plain names and the ".2" names. It carries category variables named "Category for CS-" and "Category for CS+" in that order, which is how the two series are labelled here. The reported interaction is unchanged if the two labels are exchanged.
• Published statistic: mixed-effects model followed by Bonferroni; the published CS x Time interaction is F(9,238), ns.
• Recomputed from this table: F(9,99) = 0.7453 on the 12 complete subjects, treating CS and Time as within-subject factors. The published F(9,238) comes from the restricted-maximum-likelihood mixed-effects model of GraphPad Prism, which keeps the incomplete subject and is not reproducible outside Prism; both are ns.
• Source: "WT_VasoC_BACK.svd" (StatView), the ten timepoint columns and their ".2" duplicates.

Figure_1-figure_supplement_1e_-_source_data.csv

Figure 1-figure supplement 1e. Skin temperature of the tail across the trial, CS- vs CS+.

• Rows and columns: 260 rows; subject, CS, timepoint, temperature_C
• Experimental unit: mouse. n = 13 mice; 12 with a complete series (two values missing).
• Units: skin temperature in degrees Celsius.
• Layout: one row per mouse, timepoint and CS. Empty cells are the missing values. The source file stores the two CS series as duplicate column names: the plain names and the ".2" names. It carries category variables named "Category for CS-" and "Category for CS+" in that order, which is how the two series are labelled here. The reported interaction is unchanged if the two labels are exchanged.
• Published statistic: mixed-effects model followed by Bonferroni; the published CS x Time interaction is F(9,238), ns.
• Recomputed from this table: F(9,99) = 1.3171 on the 12 complete subjects, treating CS and Time as within-subject factors. The published F(9,238) comes from the restricted-maximum-likelihood mixed-effects model of GraphPad Prism, which keeps the incomplete subject and is not reproducible outside Prism; both are ns.
• Source: "WT_VasoC_TAIL.svd" (StatView), the ten timepoint columns and their ".2" duplicates.

Figure_1-figure_supplement_2b_-_source_data.csv

Figure 1-figure supplement 2b. NC response latency for the no-tone test, the unconditioned tone test and the CFCA paradigm.

• Rows and columns: 68 rows; subject, test, value
• Experimental unit: mouse. No tone n = 10 (two trials), Tone n = 12 (two trials), CFCA n = 12 (CS- and CS+).
• Units: latency of the nociceptive response, in seconds.
• Layout: one row per measurement. Subject numbers identify mice. A number appearing in two groups is the same animal: the two no-tone trials are the same ten mice, the two tone trials the same twelve, and the CS- and CS+ rows the same twelve. The three cohorts are distinct animals. The published test treats the six groups as independent and does not use this pairing.
• Published statistic: Kruskal-Wallis test followed by Dunn, H = 25.45, P = 0.0001.
• Recomputed from this table: H = 25.4369
• Source: "WT_basal pain response.svd", "WT_tone-specific pain response.svd" and "WT_FCA.svd" (StatView).

Figure_1-figure_supplement_2c_-_source_data.csv

Figure 1-figure supplement 2c. Temperature of the NC response for the no-tone test, the unconditioned tone test and the CFCA paradigm.

• Rows and columns: 68 rows; subject, test, value
• Experimental unit: mouse. No tone n = 10 (two trials), Tone n = 12 (two trials), CFCA n = 12 (CS- and CS+).
• Units: infrared skin temperature at the nociceptive response, in degrees Celsius.
• Layout: one row per measurement. Subject numbers identify mice. A number appearing in two groups is the same animal: the two no-tone trials are the same ten mice, the two tone trials the same twelve, and the CS- and CS+ rows the same twelve. The three cohorts are distinct animals. The published test treats the six groups as independent and does not use this pairing.
• Published statistic: Kruskal-Wallis test followed by Dunn, H = 29.30, P < 0.0001.
• Recomputed from this table: H = 29.2977
• Source: "WT_basal pain response.svd", "WT_tone-specific pain response.svd" and "WT_FCA.svd" (StatView).

Figure_1-figure_supplement_2d_-_source_data.csv

Figure 1-figure supplement 2d. Temperature of the NC response after fear extinction, CS- vs CS+.

• Rows and columns: 10 rows; subject, temperature_CSminus_C, temperature_CSplus_C
• Experimental unit: mouse. n = 10 mice.
• Units: infrared skin temperature at the nociceptive response, in degrees Celsius.
• Published statistic: paired t-test, P = 0.3276, ns.
• Recomputed from this table: t9 = -1.0351
• Source: "WT_Extinction-FCA.svd" (StatView), variables "IR CS-", "IR CS+".

Figure_1-figure_supplement_2f_-_source_data.csv

Figure 1-figure supplement 2f. Freezing during fear retrieval, round 1 of the stability protocol.

• Rows and columns: 10 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 10 mice.
• Units: freezing as a proportion (0-1), as stored; the figure plots it as a percentage.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,18) = 558.2503
• Source: "WT_FCA_2xHP.svd" (StatView), variables "BL_T1", "CS-_T1", "CS+_T1".

Figure_1-figure_supplement_2g_-_source_data.csv

Figure 1-figure supplement 2g. Freezing during fear retrieval, round 2 of the stability protocol.

• Rows and columns: 10 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 10 mice.
• Units: freezing as a proportion (0-1), as stored; the figure plots it as a percentage.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,18) = 55.9970
• Source: "WT_FCA_2xHP.svd" (StatView), variables "BL_T2", "CS-_T2", "CS+_T2".

Figure_1-figure_supplement_2h_left_-_source_data.csv

Figure 1-figure supplement 2h left. NC response latency, FC hot-plate test 1.

• Rows and columns: 10 rows; subject, latency_CSminus_s, latency_CSplus_s
• Experimental unit: mouse. n = 10 mice.
• Units: latency of the nociceptive response, in seconds.
• Published statistic: paired t-test, P = 0.0029.
• Recomputed from this table: t9 = -4.0540
• Source: "WT_FCA_2xHP.svd" (StatView), variables "HP1 T CS-", "HP1 T CS+".

Figure_1-figure_supplement_2h_right_-_source_data.csv

Figure 1-figure supplement 2h right. Temperature of the NC response, FC hot-plate test 1.

• Rows and columns: 10 rows; subject, temperature_CSminus_C, temperature_CSplus_C
• Experimental unit: mouse. n = 10 mice.
• Units: infrared skin temperature at the nociceptive response, in degrees Celsius.
• Published statistic: paired t-test, P = 0.0006.
• Recomputed from this table: t9 = -5.1842
• Source: "WT_FCA_2xHP.svd" (StatView), variables "IR1 CS-", "IR1 CS+".

Figure_1-figure_supplement_2i_left_-_source_data.csv

Figure 1-figure supplement 2i left. NC response latency, FC hot-plate test 2.

• Rows and columns: 10 rows; subject, latency_CSminus_s, latency_CSplus_s
• Experimental unit: mouse. n = 10 mice.
• Units: latency of the nociceptive response, in seconds.
• Published statistic: paired t-test, P = 0.0362.
• Recomputed from this table: t9 = -2.4588
• Source: "WT_FCA_2xHP.svd" (StatView), variables "HP2 T CS-", "HP2 T CS+".

Figure_1-figure_supplement_2i_right_-_source_data.csv

Figure 1-figure supplement 2i right. Temperature of the NC response, FC hot-plate test 2.

• Rows and columns: 10 rows; subject, temperature_CSminus_C, temperature_CSplus_C
• Experimental unit: mouse. n = 10 mice.
• Units: infrared skin temperature at the nociceptive response, in degrees Celsius.
• Published statistic: paired t-test, P = 0.0245.
• Recomputed from this table: t9 = -2.6967
• Source: "WT_FCA_2xHP.svd" (StatView), variables "IR2 CS-", "IR2 CS+".

Figure_1-figure_supplement_3_c_e_g_-_source_data.csv

Figure 1-figure supplement 3, panels c, e and g. Colocalisation of Sst mRNA with Cre, Slc32a1 and Slc17a6 mRNA in the dlPAG and vlPAG, by single-molecule fluorescent in situ hybridization.

• Rows and columns: 12 rows; panel, region, bar, numerator, denominator, n_colocalized, n_total, percent
• Experimental unit: cell. Panel c: 2 mice, 2 sections each (dlPAG 6 images, vlPAG 7 images). Panels e and g: 3 mice; dlPAG 6 images from 1 section per mouse; vlPAG 9 images for Slc32a1 and 10 for Slc17a6, from 2 sections per mouse.
• Units: counts of cells, and the percentage they give.
• Published statistic: none. Each bar is a single proportion; no error bars, no test.
• Source: counts obtained by the co-author and printed on the figure and in its legend.

How the percentages were obtained, as stated by the co-author who performed the quantification: the cells of each category are summed across all images of a condition and the percentage is taken on those pooled totals. Percentages are deliberately NOT averaged over images, because the images are taken at high magnification to make the labelling unambiguous, and the number of cells per field varies enough that averaging per image would bias the result.

The percent column below is therefore the ratio of the two count columns, which is both the quantity plotted and the whole dataset for these panels: each bar carries its own count pair and there is nothing behind them to deposit.

Figure_2c_-_source_data.csv

Figure 2c. Freezing during fear retrieval, GFP and ChR2 groups pooled (photoactivation cohort).

• Rows and columns: 16 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 16 mice.
• Units: freezing expressed as a proportion (0-1), as stored in the analysis file; the figure plots the same values as percentages.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,30) = 396.7868
• Source: "SST_FCA_ChR2.svd" (StatView), variables "BL_HP Freezing", "CS-_HP Freezing", "CS+_HP Freezing".

Figure_2d_-_source_data.csv

Figure 2d. NC response latency during CS- and CS+, photoactivation of SST+ vlPAG cells.

• Rows and columns: 32 rows; subject, opsin, condition, latency_s
• Experimental unit: mouse. GFP n = 7, ChR2 n = 9.
• Units: latency of the nociceptive response, in seconds.
• Layout: one row per mouse and condition. 'subject' numbers each mouse uniquely across both groups.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(1,14) = 8.5136
• Source: "SST_FCA_ChR2.svd" (StatView), variables "HP CS- time", "HP CS+ time"; group membership from the nominal variable "Virus".

Figure_2e_-_source_data.csv

Figure 2e. Freezing during fear retrieval, GFP and ArchT groups pooled (photoinhibition cohort).

• Rows and columns: 23 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 23 mice.
• Units: freezing expressed as a proportion (0-1), as stored in the analysis file; the figure plots the same values as percentages.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,44) = 280.4195
• Source: "SST_FCA_ArchT.svd" (StatView), variables "BL_T1", "CS-_T1", "CS+_T1".

Figure_2f_-_source_data.csv

Figure 2f. NC response latency during CS- and CS+, photoinhibition of SST+ vlPAG cells.

• Rows and columns: 46 rows; subject, opsin, condition, latency_s
• Experimental unit: mouse. GFP n = 12, ArchT n = 11.
• Units: latency of the nociceptive response, in seconds.
• Layout: one row per mouse and condition. 'subject' numbers each mouse uniquely across both groups.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(1,21) = 10.6374
• Source: "SST_FCA_ArchT.svd" (StatView), variables "HP CS- T", "HP CS+ T"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_2a_-_source_data.csv

Figure 2-figure supplement 2a. Discrimination index during fear retrieval, GFP vs ChR2.

• Rows and columns: 16 rows; subject, opsin, discrimination_index
• Experimental unit: mouse. GFP n = 7, ChR2 n = 9.
• Units: discrimination index, dimensionless.
• Layout: One row per mouse.
• Published statistic: unpaired t-test, P = 0.6052, ns.
• Recomputed from this table: P = 0.6052
• Source: "SST_FCA_ChR2.svd" (StatView), variables "DI_HP"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_2b_-_source_data.csv

Figure 2-figure supplement 2b. Discrimination index during fear retrieval, GFP vs ArchT.

• Rows and columns: 23 rows; subject, opsin, discrimination_index
• Experimental unit: mouse. GFP n = 12, ArchT n = 11.
• Units: discrimination index, dimensionless.
• Layout: One row per mouse.
• Published statistic: unpaired t-test, P = 0.4575, ns.
• Recomputed from this table: P = 0.4575
• Source: "SST_FCA_ArchT.svd" (StatView), variables "DI_T1"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_2c_-_source_data.csv

Figure 2-figure supplement 2c. Conditioning index during fear retrieval, GFP vs ChR2.

• Rows and columns: 16 rows; subject, opsin, conditioning_index
• Experimental unit: mouse. GFP n = 7, ChR2 n = 9.
• Units: conditioning index, dimensionless.
• Layout: One row per mouse.
• Published statistic: unpaired t-test, P = 0.2333, ns.
• Recomputed from this table: P = 0.2331
• Source: "SST_FCA_ChR2.svd" (StatView), variables "CI_HP"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_2d_-_source_data.csv

Figure 2-figure supplement 2d. Conditioning index during fear retrieval, GFP vs ArchT.

• Rows and columns: 23 rows; subject, opsin, conditioning_index
• Experimental unit: mouse. GFP n = 12, ArchT n = 11.
• Units: conditioning index, dimensionless.
• Layout: One row per mouse.
• Published statistic: unpaired t-test, P = 0.1320, ns.
• Recomputed from this table: P = 0.1321
• Source: "SST_FCA_ArchT.svd" (StatView), variables "CI_T1"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_3a_-_source_data.csv

Figure 2-figure supplement 3a. Temperature of the NC response during CS- and CS+, photoactivation.

• Rows and columns: 32 rows; subject, opsin, condition, temperature_C
• Experimental unit: mouse. GFP n = 7, ChR2 n = 9.
• Units: infrared skin temperature at the nociceptive response, in degrees Celsius.
• Layout: one row per mouse and condition. 'subject' numbers each mouse uniquely across both groups.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(1,14) = 11.3997
• Source: "SST_FCA_ChR2.svd" (StatView), variables "IR CS-", "IR CS+"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_3b_-_source_data.csv

Figure 2-figure supplement 3b. Temperature of the NC response during CS- and CS+, photoinhibition.

• Rows and columns: 46 rows; subject, opsin, condition, temperature_C
• Experimental unit: mouse. GFP n = 12, ArchT n = 11.
• Units: infrared skin temperature at the nociceptive response, in degrees Celsius.
• Layout: one row per mouse and condition. 'subject' numbers each mouse uniquely across both groups.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(1,21) = 10.8410
• Source: "SST_FCA_ArchT.svd" (StatView), variables "IR CS-", "IR CS+"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_3c_-_source_data.csv

Figure 2-figure supplement 3c. Open-field locomotion before, during and after photoactivation of SST+ vlPAG cells.

• Rows and columns: 48 rows; subject, opsin, condition, distance_travelled
• Experimental unit: mouse. GFP n = 7, ChR2 n = 9.
• Units: distance travelled in the open field, in the units stored in the source file.
• Layout: one row per mouse and condition. 'subject' numbers each mouse uniquely across both groups.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,28) = 0.6855
• Source: "SST_FCA_OF (1).svd" (StatView), variables "PreStimulation", "Stimulation", "PostStimulation"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_3d_-_source_data.csv

Figure 2-figure supplement 3d. Open-field locomotion before, during and after photoinhibition of SST+ vlPAG cells.

• Rows and columns: 69 rows; subject, opsin, condition, distance_travelled
• Experimental unit: mouse. GFP n = 12, ArchT n = 11.
• Units: distance travelled in the open field, in the units stored in the source file.
• Layout: one row per mouse and condition. 'subject' numbers each mouse uniquely across both groups.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,42) = 0.9243
• Source: "SST_FCA_OF.svd" (StatView), variables "Pre-Stim", "Stim1-3", "Post-Stim"; group membership from the nominal variable "Virus".

Figure_2-figure_supplement_3f_-_source_data.csv

Figure 2-figure supplement 3f (formerly Supplementary Figure 5f). Real-time place preference: time spent in the stimulated compartment, photoactivation of SST+ vlPAG cells. GFP n = 6, ChR2 n = 8.

• Rows and columns: 14 rows; mouse_id, opsin, percent_time_in_stimulated_compartment
• Experimental unit: mouse. 14 mice, 6 GFP and 8 ChR2.
• Units: percentage of the session time spent in the stimulated compartment.
• Layout: One row per mouse.
• Published statistic: Mann-Whitney test (two-tailed), U = 13, P = 0.1812, ns.
• Recomputed from this table: U = 13, same group sizes.
• Source: "Winke et al. Source Data.xlsx", sheet "ED Fig 5", panel f.

NOTE on mouse IDs: in the 2022 source workbook "1513" appears twice, once in the ChR2 group and once in the GFP group. A later version of the same workbook (2023) lists the RTPP roster as ChR2 = 1493, 1512, 1514, 1516, 1518, 1520, 1524, 1526 and GFP = 1494, 1513, 1515, 1517, 1519, 1521, 1525. The ChR2 entry recorded here as 1513 is therefore mouse 1516; 1513 belongs to the GFP group. Only the identifier is affected: the measured values, the group sizes (6 / 8) and U = 13 are those used in the publication and are unchanged.

Figure_2-figure_supplement_4b_-_source_data.csv

Figure 2-figure supplement 4b. Freezing during fear conditioning with the laser as the unconditioned stimulus.

• Rows and columns: 39 rows; subject, opsin, condition, freezing_percent
• Experimental unit: mouse. GFP n = 6, ChR2 n = 7.
• Units: freezing in percent.
• Layout: one row per mouse and condition.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,22) = 0.7582
• Source: "ED6.pzfx" (GraphPad Prism), data table "ED6b".

Figure_2-figure_supplement_4c_-_source_data.csv

Figure 2-figure supplement 4c. Freezing during fear retrieval with the laser as the unconditioned stimulus.

• Rows and columns: 39 rows; subject, opsin, condition, freezing_percent
• Experimental unit: mouse. GFP n = 6, ChR2 n = 7.
• Units: freezing in percent.
• Layout: one row per mouse and condition.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,22) = 1.9262
• Source: "ED6.pzfx" (GraphPad Prism), data table "ED6c".

Figure_2-figure_supplement_4d_-_source_data.csv

Figure 2-figure supplement 4d. Freezing during fear conditioning with footshock plus laser as the unconditioned stimulus.

• Rows and columns: 39 rows; subject, opsin, condition, freezing_percent
• Experimental unit: mouse. GFP n = 6, ChR2 n = 7.
• Units: freezing in percent.
• Layout: one row per mouse and condition.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,22) = 0.1444
• Source: "ED6.pzfx" (GraphPad Prism), data table "ED6d".

Figure_2-figure_supplement_4e_-_source_data.csv

Figure 2-figure supplement 4e. Freezing during fear retrieval with footshock plus laser as the unconditioned stimulus.

• Rows and columns: 39 rows; subject, opsin, condition, freezing_percent
• Experimental unit: mouse. GFP n = 6, ChR2 n = 7.
• Units: freezing in percent.
• Layout: one row per mouse and condition.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,22) = 0.1095
• Source: "ED6.pzfx" (GraphPad Prism), data table "ED6e".

Figure_2-figure_supplement_5c_-_source_data.csv

Figure 2-figure supplement 5c. Freezing during fear retrieval in Sst-IRES-Cre heterozygotes.

• Rows and columns: 13 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 13 mice.
• Units: freezing in percent.
• Layout: One row per mouse.
• Published statistic: Friedman test followed by Dunn.
• Recomputed from this table: Q = 19.8462
• Source: "ED7v2.prism" (GraphPad Prism), data table B2100381...

Figure_2-figure_supplement_5d_-_source_data.csv

Figure 2-figure supplement 5d. NC response latency during CS- and CS+ in Sst-IRES-Cre heterozygotes.

• Rows and columns: 26 rows; subject, opsin, condition, latency_s
• Experimental unit: mouse. GFP n = 6, ChR2 n = 7.
• Units: latency of the nociceptive response, in seconds.
• Layout: one row per mouse and CS.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(1,11) = 12.3903
• Source: "ED7v2.prism" (GraphPad Prism), data table 70208965...

Figure_2-figure_supplement_5g_-_source_data.csv

Figure 2-figure supplement 5g. Freezing during fear retrieval in Vip-IRES-Cre mice.

• Rows and columns: 10 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 10 mice.
• Units: freezing in percent.
• Layout: One row per mouse.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,18) = 147.1254
• Source: "ED7v2.prism" (GraphPad Prism), data table 9AFCAAA1...

Figure_2-figure_supplement_5h_-_source_data.csv

Figure 2-figure supplement 5h. NC response latency during CS- and CS+ in Vip-IRES-Cre mice.

• Rows and columns: 20 rows; subject, opsin, condition, latency_s
• Experimental unit: mouse. GFP n = 5, ArchT n = 5.
• Units: latency of the nociceptive response, in seconds.
• Layout: one row per mouse and CS.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(1,8) = 0.3984
• Source: "ED7v2.prism" (GraphPad Prism), data table AB1311C5...

Figure_3c_-_source_data.csv

Figure 3c. Freezing during fear retrieval and extinction, photoinhibition of SST+ vlPAG cells.

• Rows and columns: 70 rows; subject, opsin, condition, freezing_proportion
• Experimental unit: mouse. GFP n = 6, ArchT n = 8.
• Units: freezing expressed as a proportion (0-1), as stored in the analysis file; the figure plots the same values as percentages.
• Layout: one row per mouse and condition. 'subject' numbers each mouse uniquely across both groups.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(4,48) = 0.1954
• Source: "SST_FR_ArchT.svd" (StatView), variables "BL", "CS-", "CS+1-4", "CS+5-8", "CS+9-12"; group membership from the nominal variable "Virus".

Figure_3d_-_source_data.csv

Figure 3d. Freezing during fear retrieval and extinction, photoactivation of SST+ vlPAG cells.

• Rows and columns: 85 rows; subject, opsin, condition, freezing_proportion
• Experimental unit: mouse. GFP n = 9, ChR2 n = 8.
• Units: freezing expressed as a proportion (0-1), as stored in the analysis file; the figure plots the same values as percentages.
• Layout: one row per mouse and condition. 'subject' numbers each mouse uniquely across both groups.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(4,60) = 7.9578
• Source: "SST_FR_ChR2.svd" (StatView), variables "BL", "CS-", "CS+_1-4", "CS+_5-8", "CS+_9-12"; group membership from the nominal variable "Virus".

Figure_4b_-_source_data.csv

Figure 4b. Spinal field potentials before, during and after photoactivation of SST+ vlPAG cells.

• Rows and columns: 8 rows; subject, AUC_OFF_1, AUC_ON, AUC_OFF_2
• Experimental unit: mouse. n = 8 mice.
• Units: area under the curve of the evoked spinal field potential, in the units stored in the source file (V.s), given to six significant figures.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,14) = 17.9426
• Source: "pascal data - Nanci.svd" (StatView), variables "Off1", "ON", "Off2".

Figure_4c_-_source_data.csv

Figure 4c. Area under the curve of the evoked spinal field potential before, during and after photoinhibition of SST+ vlPAG cells.

• Rows and columns: 6 rows; subject, AUC_control, AUC_photoinhibition, AUC_recovery, included_in_analysis
• Experimental unit: mouse. 6 mice recorded, 5 analysed (subject 4 excluded as an outlier).
• Units: area under the curve, in the units stored in the source file (V.s), given to six significant figures. All 6 recorded subjects are kept; included_in_analysis marks the one excluded. Subject 4 is the subject flagged "outlier" in the co-author's own spreadsheet ("detail panel 4c.xlsx"), which is retained in the working files for provenance but is not the deposited record: it differs from this table on two subjects and does not reproduce the published F. The StatView values below do.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni, F(2,8) = 5.826, P = 0.0275.
• Recomputed from this table: F(2,8) = 5.8383, P = 0.0273 on the 5 analysed subjects.
• Source: "pascal data - Nanci.svd" (StatView), the "ArchT AUC" block: the third set of "Off1", "ON", "Off2" columns in the file.

Figure_4d_-_source_data.csv

Figure 4d. WDR neuron spike counts in response to noxious stimulation, OFF vs ON photoactivation of SST+ vlPAG cells.

• Rows and columns: 12 rows; cell, A_beta_OFF_spikes, A_beta_ON_spikes, A_delta_OFF_spikes, A_delta_ON_spikes, C_OFF_spikes, C_ON_spikes
• Experimental unit: cell. 12 cells from 8 mice.
• Units: spike count per stimulation, averaged over the trials of that condition.
• Layout: One row per cell; OFF and ON are the paired values for each fibre component.
• Published statistic: paired t-test per fibre component.
• Recomputed from this table: A-beta t11 = -0.6319, P = 0.5403 against the published 0.5430 (probable digit transposition; ns either way); A-delta t11 = -3.9587, P = 0.0022 (exact); C t11 = -3.7387, P = 0.0033 (exact)
• Source: "detail panel 4d.xlsx" (co-author spreadsheet), columns B/C for A-beta, H/I for A-delta, N/O for C.

Figure_4e_-_source_data.csv

Figure 4e. WDR neuron spike counts in response to noxious stimulation, OFF vs ON photoinhibition of SST+ vlPAG cells.

• Rows and columns: 14 rows; cell, A_beta_OFF_spikes, A_beta_ON_spikes, A_delta_OFF_spikes, A_delta_ON_spikes, C_OFF_spikes, C_ON_spikes
• Experimental unit: cell. 14 cells from 5 mice.
• Units: spike count per stimulation, averaged over the trials of that condition.
• Layout: One row per cell; OFF and ON are the paired values for each fibre component.
• Published statistic: paired t-test per fibre component.
• Recomputed from this table: A-beta t13 = 1.5778, P = 0.1386 (exact); A-delta t13 = 2.5530, P = 0.0241 (exact); C t13 = 6.2736, P < 0.0001 (exact)
• Source: "detail panel 4e.xlsx" (co-author spreadsheet), columns B/C for A-beta, F/G for A-delta, J/K for C.

Figure_4f_-_source_data.csv

Figure 4f. WDR neuron spike counts in response to subthreshold stimulation, OFF vs ON photoactivation of SST+ vlPAG cells.

• Rows and columns: 10 rows; cell, A_beta_OFF_spikes, A_beta_ON_spikes, A_delta_OFF_spikes, A_delta_ON_spikes, C_OFF_spikes, C_ON_spikes
• Experimental unit: cell. 10 cells from 8 mice.
• Units: spike count per stimulation, averaged over the trials of that condition.
• Layout: One row per cell; OFF and ON are the paired values for each fibre component.
• Published statistic: Wilcoxon matched-pairs signed rank test per fibre component.
• Recomputed from this table: A-beta W = 7, P = 0.3750 (exact); A-delta W = 28, P = 0.0156 (exact); C W = 55, P = 0.0020 (exact). W is the signed sum of ranks (W+ minus W-), as reported by GraphPad Prism. On A-beta six of the ten pairs have a zero difference and are dropped, so that test rests on four pairs.
• Source: "detail panel 4f.xlsx" (co-author spreadsheet), columns B/C for A-beta, F/G for A-delta, I/J for C.

Figure_4-figure_supplement_1c_-_source_data.csv

Figure 4-figure supplement 1c (formerly Supplementary Figure 8c). Mean firing rate of purely tactile spinal cord neurons in response to gentle brushing of the skin, before vs during photoactivation of ChR2-expressing SST+ vlPAG cells.

• Rows and columns: 5 rows; cell, firing_rate_before_Hz, firing_rate_during_photoactivation_Hz
• Experimental unit: cell. 5 cells.
• Units: mean firing rate in spikes per second (Hz).
• Layout: One row per cell; the two columns are the paired values before and during photoactivation.
• Published statistic: paired t-test (two-tailed), t4 = 0.4247, P = 0.6929, ns.
• Recomputed from this table: t4 = 0.4247, P = 0.6929 (exact).
• Source: "Electrophysiologie moelle.xlsx", sheet "Brush Chr2", in "Winke et al LAST/second submission/LAST DATA PASCAL/".

Figure_5e_-_source_data.csv

Figure 5e. WDR C-fibre response, OFF vs ON photoactivation of SST+ vlPAG projections to the spinal cord.

• Rows and columns: 11 rows; cell, OFF_spikes, ON_spikes
• Experimental unit: cell. 11 cells from 7 mice.
• Units: spike count per stimulation, averaged over the trials of that condition.
• Layout: One row per cell; OFF and ON are the paired values.
• Published statistic: paired t-test (two-tailed), t10 = 0.94222, P = 0.3683, ns.
• Recomputed from this table: t10 = -0.94221, P = 0.3683 (exact); the sign differs only because the published value is reported for the reverse subtraction order.
• Source: "Winke et al. Source Data.xlsx", sheet "Figure 5", panel e. The same values are independently present in the co-author spreadsheet "detail panel 5e (Pascal).xlsx".

Figure_5f_left_-_source_data.csv

Figure 5f left. WDR neuron spike counts in response to noxious stimulation, OFF vs ON photoactivation of the SST+ vlPAG-to-RVM projection.

• Rows and columns: 10 rows; cell, A_beta_OFF_spikes, A_beta_ON_spikes, A_delta_OFF_spikes, A_delta_ON_spikes, C_OFF_spikes, C_ON_spikes
• Experimental unit: cell. 10 cells from 7 mice.
• Units: spike count per stimulation, averaged over the trials of that condition.
• Layout: One row per cell; OFF and ON are the paired values for each fibre component.
• Published statistic: paired t-test for A-beta and A-delta, Wilcoxon matched-pairs signed rank test for C.
• Recomputed from this table: A-beta t9 = -2.0203, P = 0.0741 (exact); A-delta t9 = -2.4877, P = 0.0345 against the published 0.0346 (rounding); C W = 55, P = 0.0020 (exact)
• Source: "detail panel 5f left.xlsx" (co-author spreadsheet), columns B/C for A-beta, G/H for A-delta, L/M for C.

Figure_5f_right_-_source_data.csv

Figure 5f, right panel. WDR C-fibre response to subthreshold electrical stimulation, OFF vs ON photoactivation of the SST+ vlPAG-RVM pathway.

• Rows and columns: 10 rows; cell, OFF_spikes, ON_spikes, included_in_analysis, exclusion_note
• Experimental unit: cell. 10 cells recorded from 7 mice; 8 cells in the published analysis.
• Units: spike count per stimulation, averaged over the trials of that condition.
• Layout: One row per cell; OFF and ON are the paired values. 'included_in_analysis' marks the 8 cells that entered the published test; 'exclusion_note' carries the reason for the other two.
• Published statistic: Wilcoxon matched-pairs signed rank test (two-tailed), W = -36, P = 0.0078, on the 8 included cells.
• Recomputed from this table: |W| = 36, P = 0.0078 (exact) on the 8 cells flagged included_in_analysis = yes. W is the signed sum of ranks (W+ minus W-), as reported by GraphPad Prism; all eight included cells increase from OFF to ON, so the sign follows the subtraction order and the published value is negative.
• Source: "detail panel 5f right.xlsx" (co-author spreadsheet). In that file the two excluded cells carry a yellow fill on their OFF and ON cells, which matches the two cells identified here.

Cells 2 and 9 were not included. The Methods state that outliers were removed using the Tukey method. Applied across cells to the ON values, the Tukey fence is [-1.750 ; 8.250] and flags cells 2 and 9, and no others; the same two cells are flagged on the ON-OFF differences and under either quartile convention. The exclusion is therefore consistent with the stated criterion and can be reproduced from this file.

Both excluded cells are high responders (ON = 12.5 and 9.33, against a maximum of 4.5 among the eight retained), so their removal makes the reported effect smaller, not larger. Including all ten cells would give W = -55, P = 0.0020.

All ten recorded cells are retained in this file; none has been deleted.

Figure_5-figure_supplement_1b_-_source_data.csv

Figure 5-figure supplement 1b. WDR neuron spike counts in response to noxious stimulation, before vs during photoactivation, in GFP control mice (vlPAG-to-RVM pathway).

• Rows and columns: 7 rows; cell, A_beta_control_spikes, A_beta_photoactivation_spikes, A_delta_control_spikes, A_delta_photoactivation_spikes, C_control_spikes, C_photoactivation_spikes
• Experimental unit: cell. 7 cells from 3 mice.
• Units: spike count per stimulation, averaged over the trials of that condition.
• Layout: One row per cell; control and photoactivation are the paired values for each fibre component. This is the GFP control for the experiment shown in Figure 5f left.
• Published statistic: paired t-test per fibre component; t6 = 0.9564, 0.3520 and 0.8755, all ns.
• Recomputed from this table: t6 = 0.9564, P = 0.3758 (exact); t6 = 0.3520, P = 0.7369 (exact); t6 = -0.8755, P = 0.4150; |t| as published (exact)
• Source: "RVM_GFP_detail_fibres.xlsx", supplied by the co-author who performed the recordings. Columns C/D for A-beta, G/H for A-delta, J/K for C.

Figure_6b_-_source_data.csv

Figure 6b. Freezing across baseline, CS- and three CS+ blocks, in GFP and ChR2 mice (vlPAG-to-RVM pathway).

• Rows and columns: 65 rows; subject, opsin, condition, freezing_percent
• Experimental unit: mouse. GFP n = 7, ChR2 n = 6 (13 mice, 5 conditions each).
• Units: freezing in percent. Long format: one row per mouse and condition, 'subject' numbering each mouse uniquely across both groups. The source table declares 7 replicate columns per group; the seventh ChR2 column is empty, hence 6 ChR2 mice.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: interaction F(4,44) = 0.6292, P = 0.6443.
• Source: Figure6.pzfx (GraphPad Prism), data table "Fig6 b".

Figure_6c_-_source_data.csv

Figure 6c. Freezing during retrieval, GFP and ChR2 groups pooled (vlPAG-to-RVM pathway).

• Rows and columns: 14 rows; subject, freezing_BL, freezing_CSminus, freezing_CSplus
• Experimental unit: mouse. n = 14 mice.
• Units: freezing in percent. 'subject' is the row position in the source table, not an animal identifier.
• Published statistic: one-way repeated-measures ANOVA followed by Bonferroni.
• Recomputed from this table: F(2,26) = 274.1762.
• Source: Figure6.pzfx (GraphPad Prism), data table "Fig 6c".

Figure_6d_-_source_data.csv

Figure 6d. NC response latency during CS- and CS+, photoactivation of the SST+ vlPAG-to-RVM projection.

• Rows and columns: 28 rows; subject, opsin, condition, latency_s
• Experimental unit: mouse. GFP n = 7, ChR2 n = 7.
• Units: latency of the nociceptive response, in seconds.
• Layout: one row per mouse and CS. The analysis sheet stored in the source file was set to match on Opsin rather than on CS, which treats each mouse as having received both GFP and ChR2; that setting produced the F = 10.91, P = 0.0063 of the Reviewed Preprint. With matching declared on CS, as the design requires, the same table gives F(1,12) = 19.87, P = 0.0008. The interaction sum of squares is identical under both (331.3408); only the error term changes. The corrected values were re-run in Prism by the author who performed the original analysis and are the ones reported in the Version of Record.
• Published statistic: two-way repeated-measures ANOVA followed by Bonferroni. The corrected values are F(1,12) = 19.87, P = 0.0008 (see note).
• Recomputed from this table: F(1,12) = 19.8748, P = 0.0008.
• Source: "Figure6.prism" (GraphPad Prism), data table 8B6EA88C... ("Fig6d").