DANDI:001177

There were fewer dopaminergic cell bodies in the substantia nigra pars compacta (SNc) in 6-OHDA-treated versus control mice

Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 1; Figure S1D

What the paper reports

• Mann-Whitney U, p < 0.001 (n = 9 control and 21 parkinsonian mice)

TH+ cells over the four sections and both sides divided by the area counted. 29 of 32 mice could be given a group (9 control, 20 parkinsonian; paper 9 and 21); median density 316.0 against 186.6 cells per mm2, Mann-Whitney p = 0.0003. The three unlinked mice (C0.1, C2.1, D5.2) do not change the result in either group (p = 0.0001 and 0.0019).

How the test differed from the paper’s method

• The workbook has no group column. Groups are taken from the histology workbook of the same deposit, the README's PPX lists and the DANDI subjects' lesion, in that order.
• Two rows (D3 and N4 at -3.51 mm) have shifted count and area columns and are left out.

On the accelerating rotarod test, parkinsonian mice showed a lower latency to fall, but improved with pramipexole (PPX) (0.5 mg/kg)

Figure 1C

What the paper reports

• Ctrl against Park/PPX: Kruskal-Wallis, Dunn's, p < 0.001
• Park against PPX: Kruskal-Wallis, Dunn's, p < 0.001
• Park against saline: Kruskal-Wallis, Dunn's, p = 0.95 (n = Ctrl 31, Park/PPX 45, Park/Saline 27 mice)

Medians control 91.3 s (n = 40), parkinsonian before PPX 47.3 s (n = 45), on PPX 66.7 s, second cohort before saline 51.0 s and on saline 62.3 s (n = 27). Kruskal-Wallis over the five columns p = 6.7e-15; Dunn control against parkinsonian p = 1.1e-15, before against PPX p = 1.8e-05, before against saline p = 0.83. Variant (Supplementary methods) Mann-Whitney p = 5e-13, Wilcoxon p = 3e-08 (PPX) and 0.074 (saline). With Dunn's correction over all ten pairs of the five columns, p = 3.8e-15, 6.0e-05 and 1.

How the test differed from the paper’s method

• The deposited columns are labelled Park_cohort1_baseline and Park_cohort1_PPX, Park_cohort2_baseline and Park_cohort2_Saline; the pre-drug sessions are the baseline columns, as the claim's criterion asks the author to read from the file.
• Dunn's comparisons are corrected for the three comparisons the criterion names, with all ten pairs as a variant; the workbook has 40 controls where Table 1 gives 31 and does not say which 31 were used, so all 40 enter.

PPX-naive control and parkinsonian mice showed similar delay discounting

Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 2; Figure 1F and G

What the paper reports

• depletion main effect: two-way ANOVA, Bonferroni, F(1, 180) = 3.14e-003, p = 0.9554 (n = 16 control and 31 parkinsonian mice)

Reading (a), every deposited baseline session (16 control, 31 parkinsonian mice); group F(1, 178) = 0.007, p = 0.93 (paper F(1, 180) = 0.003, p = 0.96); every per-delay Bonferroni p = 1. Reading (b), the stable-run rule, keeps 12 mice; group p = 0.11, per-delay p >= 0.45. Both readings find no difference. Reading (b) is not a sample of the paper's cohort; mice have 1 to 5 baselines before the first drug session (median 3), only 3 of 16 control and 9 of 31 parkinsonian mice pass, and these discount little (control 95.4% at 0 s to 82.4% at 9 s, against 91.7% to 32.1% for every mouse), so the outcome rests on reading (a).

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• Two control mice of the 2020 cohort have no 9 s block, so error df is 178.

both control and parkinsonian mice showed classic discounting behaviour, choosing the large reward less often as the associated delay increased

Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 2; Figure 1F and G

What the paper reports

• delay main effect, baseline, control against parkinsonian: two-way ANOVA, Bonferroni, F(3, 180) = 38.53, p < 10⁻⁴ (n = 16 control and 31 parkinsonian mice)

Reading (a); delay F(3, 178) = 38.15, p = 4.3e-19 (paper F(3, 180) = 38.53); delayed/large choices fall from 91.7 to 32.1% (control) and 85.7 to 32.9% (parkinsonian) between 0 and 9 s. Reading (b), 12 mice; the means fall (95.4 to 82.4%, 95.7 to 58.8%) but delay p = 0.11. The criterion is met under reading (a). Reading (b) keeps the mice that discount least (see fig1fg-baseline-similar) and does not bear on the claim.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

PPX did not significantly change decision-making in healthy control mice

Figure 1F

What the paper reports

• PPX main effect: two-way repeated measures ANOVA, Bonferroni, F(1, 60) = 0.9564, p = 0.332 (n = 16 mice)

16 control mice. PPX effect p = 0.10 (first session), 0.16 (four sessions) with every baseline, 0.38 and 0.27 with stable baselines (3 mice); paper F(1, 60) = 0.96, p = 0.33. No per-delay Bonferroni comparison reaches p < 0.05 on the pooled error or by paired tests (smallest 0.21). Largest effect 5.5 points at 0 s.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

but reduced the likelihood of delayed/large choices at every delay in parkinsonian mice

Figure 1G; Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 2

What the paper reports

• PPX main effect: two-way repeated measures ANOVA, F(1, 120) = 95.48, p < 10⁻⁴ (n = 31 mice)
• 0 s: Bonferroni, p < 10⁻⁴
• 3 s: Bonferroni, p < 10⁻⁴
• 6 s (Table 1 gives p = 0.0001; the legend gives P < 0.0001): Bonferroni, p = 10⁻⁴
• 9 s (Table 1 gives p = 0.0087; the legend gives P < 0.0087): Bonferroni, p = 0.0087

Every baseline and the mean of four PPX sessions, 30 mice; PPX F(1, 116) = 91.46, p = 2.5e-16 (paper F(1, 120) = 95.48); lower at 0, 3, 6 and 9 s by 12.5, 14.7, 9.7 and 6.4 points, Bonferroni p < 0.0001, < 0.0001, 0.0002, 0.022 (paired p <= 0.0034). With the first PPX session, F = 28.56, p = 4.6e-07, two of four delays significant on the pooled error. Mouse E6 has no PPX session in the deposit, so 30 of 31 mice enter.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• PPX ON is read two ways, as the criterion lists; the outcome rests on the mean of the first four sessions, as Summary_Cognitive.m averages a mouse's sessions within a condition.

Mice showed low omission rates during baseline and PPX sessions

Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 3; Figure S2D and E

What the paper reports

• center omissions, parkinsonian, 6 s: two-way ANOVA, Bonferroni, p = 6 × 10⁻⁴ (n = 31 mice)
• center omissions, parkinsonian, PPX main effect: two-way ANOVA, F(1, 120) = 16.54, p < 10⁻⁴ (n = 31 mice)
• side omissions, parkinsonian, PPX main effect: two-way ANOVA, F(1, 120) = 2.06, p = 0.1536 (n = 31 mice)

Four PPX sessions. Parkinsonian center omissions differ only at 6 s (4.5 against 11.7%, Bonferroni p < 0.0001; 0 s p = 1, 3 s p = 0.15, 9 s p = 0.49); PPX main effect F(1, 116) = 24.82, p = 2.2e-06 (paper F(1, 120) = 16.54). Side omissions F(1, 116) = 0.96, p = 0.33 (paper 2.06, p = 0.15) and no per-delay difference (smallest p = 0.94); control comparisons all p >= 0.60. Mean rates in parkinsonian mice 5.0 and 7.8% (center), 1.6 and 1.9% (side) at baseline and on PPX.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• Omissions are counted in the free choice blocks as a percentage of the block's trials, the code's 'center omi' and 'side omi' events; the PPX sessions are the mean of four, as in fig1g-park-ppx-fewer-large.

in sessions following a 48 h PPX washout period, the A value recovered to baseline values, while differences in K and AUC persisted

Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 4; Figure S2I-K

What the paper reports

• A, baseline against PPX OFF (Figure S2I legend): Friedman, Dunn's (Supplementary methods); one-way ANOVA, Bonferroni (Table 1), p > 0.99 (n = 31 mice)
• K, baseline against PPX OFF (Figure S2J legend): Friedman, Dunn's (Supplementary methods); one-way ANOVA, Bonferroni (Table 1), p = 0.01 (n = 31 mice)
• AUC, baseline against PPX OFF (Figure S2K legend): Friedman, Dunn's (Supplementary methods); one-way ANOVA, Bonferroni (Table 1), p < 0.01 (n = 31 mice)
• A, PPX ON against PPX OFF (Table 1): one-way ANOVA, Bonferroni, p < 10⁻⁴ (n = 31 mice)
• K, PPX ON against PPX OFF (Table 1): one-way ANOVA, Bonferroni, p > 0.99 (n = 31 mice)
• AUC, PPX ON against PPX OFF (Table 1): one-way ANOVA, Bonferroni, p = 0.0019 (n = 31 mice)

Four PPX sessions and every washout session, 30 mice. Reading (a), Friedman with Dunn's; A after washout 0.879 against 0.891 at baseline (p = 0.73), K 0.419 against 0.258 (p = 0.020), AUC 3.38 against 3.95 (p = 0.078). Reading (b), repeated measures ANOVA with Bonferroni comparisons by paired t tests; A p = 1, K p = 0.063, AUC p = 0.039; ON against OFF A p = 3e-05, K p = 1, AUC p = 0.0011 (Table 1 gives < 0.0001, > 0.99, 0.0019). With the Bonferroni comparisons on the ANOVA's pooled error, all three parts hold (A p = 1, K p = 0.028, AUC p = 0.0072; ON against OFF A p = 5e-06, K p = 1, AUC p = 0.0051). The outcome rests on the pooled comparisons. Under the paired ones A recovers and both K and AUC stay on the PPX side of baseline but no test makes both significant (partially reproduced in rounds 1 and 2). Against the paper's six printed p-values (Figure S2I-K legend, Table 1), the pooled comparisons match its significance at 0.05 in all six and the paired ones in five, missing K (paper 0.01, paired 0.063, pooled 0.028); for AUC the paper gives p < 0.01, which pooled meets (0.0072) and paired does not (0.039); paired is closer only for ON against OFF AUC (paper 0.0019, paired 0.0011, pooled 0.0051). Changed to reproduced in round 3. With the first OFF or the first ON session, no test gives both K and AUC.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• PPX OFF is every washout session of the mouse; with the first ON and first OFF session none of the three differs from baseline (A p = 1, K p = 0.59, AUC p = 0.38). All four ON and OFF pairs are run.
• Reading (b) does not say how the Bonferroni comparisons are made. GraphPad Prism 10, which the paper used ('Statistical tests were performed using GraphPad Prism 10', Methods, Statistical analysis), makes them on the pooled error when sphericity is assumed and by paired comparisons when it is not, and its guide recommends not assuming sphericity for repeated measurements over time. The outcome takes the pooled-error comparisons as the assumption, since they match the paper's printed p-values better; with paired comparisons the claim is partially reproduced. Asked of the authors.

At a dose of 1 mg/kg, A77636 increased locomotor speed in parkinsonian but not healthy mice

Results, Dopamine D1R agonist A77636 causes impulsive decision-making in parkinsonian mice, paragraph 1; Figure 2B and C

What the paper reports

• control, before against A77636: Wilcoxon signed-rank, p = 0.16 (n = 10 mice)
• parkinsonian, before against A77636: Wilcoxon signed-rank, p = 0.02 (n = 16 mice)

26 files. The 10 control mice are not faster 30 to 100 min after A77636 (median 3.19 to 2.90 cm/s, Wilcoxon p = 0.065; paper p = 0.16); the other 16 are (3.14 to 3.69 cm/s, 13 of 16 faster, p = 0.016; paper p = 0.02). Six of these 16 (O6, R0, R1, R3.2, R5.3, R6) are in neither A77636 list of the README; the 10 mice on the README's parkinsonian list alone are also faster (3.25 to 3.69 cm/s, 9 of 10, p = 0.014).

How the test differed from the paper’s method

• The files hold time and velocity without an injection mark. Injection is assumed at 20 min after the start, where 25 of 26 files have a short run of untracked frames (20.0 to 20.3 min), consistent with the mouse being lifted for the injection and with a 20 min baseline (-20 to 0 min, the window stated for CNO); a -30 to 0 min baseline does not exist in these files, so the criterion's reading (b) cannot be run.
• Mice not in the README's A77636 control list are taken as parkinsonian, because the control list is complete in the files (10 of 10) and 16 mice remain, the paper's N.

A77636 did not significantly change decision-making in healthy control mice

Figure 2D

What the paper reports

• A77636 main effect: two-way repeated measures ANOVA, Bonferroni, F(1, 36) = 1.05, p = 0.3115 (n = 10 mice)

10 control mice; A77636 effect p = 0.80 and 0.61 (first, four sessions) with every baseline, 0.79 and 0.31 with stable baselines (8 mice); paper F(1, 36) = 1.05, p = 0.31. No per-delay comparison reaches p < 0.05 (smallest pooled p = 0.71).

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

A77636 shifted the discounting curve in parkinsonian mice towards immediate/small rewards

Figure 2E; Results, Dopamine D1R agonist A77636 causes impulsive decision-making in parkinsonian mice, paragraph 1

What the paper reports

• A77636 main effect: two-way repeated measures ANOVA, F(1, 60) = 45.68, p < 10⁻⁴ (n = 16 mice)
• 0 s: Bonferroni, p = 0.57
• 3 s: Bonferroni, p = 0.0047
• 6 s: Bonferroni, p = 3 × 10⁻⁴
• 9 s: Bonferroni, p = 2 × 10⁻⁴

16 parkinsonian mice, every baseline, four A77636 sessions; F(1, 60) = 45.06, p = 7.7e-09 (paper 45.68); lower at 3, 6 and 9 s by 15.3, 19.0 and 19.3 points (Bonferroni p = 0.0049, 0.0003, 0.0003; paper 0.0047, 0.0003, 0.0002), not at 0 s (p = 0.56; paper 0.57). With the first session, F = 16.85, p = 0.0001, with 6 and 9 s significant.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

A77636 had no significant effects on A in parkinsonian mice, but led to a significant increase in K ... A77636 also reduced the AUC

Figure 2F-H; Results, Dopamine D1R agonist A77636 causes impulsive decision-making in parkinsonian mice, paragraph 1

What the paper reports

• A: Friedman, Dunn's, p = 0.31 (n = 16 mice)
• K: Friedman, Dunn's, p = 0.0053 (n = 16 mice)
• AUC (Table 1 gives p = 0.0008, the legend P < 0.0008): Friedman, Dunn's, p = 8 × 10⁻⁴ (n = 16 mice)

16 mice, four A77636 sessions, every washout session, no failed fit. A unchanged (Friedman p = 0.31, Dunn p = 0.31; paper 0.31), K higher (median 0.115 to 0.291, Dunn p = 0.0053; paper 0.0053), AUC lower (5.82 to 3.97, Dunn p = 0.0008; paper 0.0008). With the first session only, K and AUC are not significant (Dunn p = 0.43, 0.10).

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• Dunn's comparisons against baseline, corrected for two (Supplementary methods, Behavior).

K values did not correlate with TH immunofluorescence, whether in DMS, DLS or VS

Figure 3B

What the paper reports

• DLS: Pearson correlation, r = 0.34, p = 0.07 (n = 28 mice)
• DMS: Pearson correlation, r = -0.02, p = 0.92 (n = 28 mice)
• VS: Pearson correlation, r = 0.18, p = 0.36 (n = 28 mice)

K on PPX against TH, n = 24; DMS r = 0.03 (p = 0.89), DLS r = 0.40 (p = 0.056), VS r = 0.32 (p = 0.13). With all controls as reference, p = 0.70, 0.060, 0.068. With the first PPX session, every p >= 0.32. None is significant.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

A77636-induced impulsivity did not correlate with dopaminergic lesions in any subregions of the striatum

Figure S4; Results, Changes in sensitivity to reward size relate to dopaminergic lesions in the striatum

What the paper reports

• smallest p of the six correlations (K against VS TH): Pearson correlation, r = -0.31, p = 0.23 (n = 16 mice (legend); Table 1 gives N = 28)

A and K on A77636 against TH in 16 mice; all six p >= 0.24, the smallest K against VS (r = -0.31, p = 0.24; paper r = -0.31, p = 0.23). With the first A77636 session the smallest is p = 0.071 (K against VS, r = -0.46). All 16 mice are in one staining cohort, so the two normalizations give the same r.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

Application of the DREADD agonist, clozapine-N-oxide (CNO), reduced oIPSC amplitude ... Application of CNO triggered a significant increase in firing rate

Results, Chemogenetic manipulations of striatal neurons mimic the effect of pramipexole in parkinsonian mice, paragraph 1; Figure S6

What the paper reports

• oIPSC amplitude: Wilcoxon signed-rank, p < 0.01 (n = 2 mice, 6 cells)
• dMSN firing to 300 pA: Wilcoxon signed-rank, p = 0.01 (n = 3 mice, 9 cells)
• input-output curve, CNO effect: two-way ANOVA, Bonferroni, F(1, 32) = 17.64, p = 2 × 10⁻⁴ (n = 3 mice, 9 cells)

oIPSC amplitude 647.5 to 55.0 pA (median, 6 cells, Wilcoxon p = 0.031); firing at 300 pA 8 to 16 Hz (9 cells, p = 0.016); CNO effect on the input-output curve (100 to 400 pA) F(1, 32) = 17.64, p = 0.0002, the paper's values, with each cell and current as the repeated unit. An ordinary two-way ANOVA gives p = 0.082.

How the test differed from the paper’s method

• The input-output ANOVA uses the layout of the delay discounting ANOVA, which the paper's F(1, 32) for 9 cells and 4 currents implies.

robustly shifted choices towards immediate/small rewards over delayed/large rewards in hM4D i -expressing or hM3D q -expressing but not mCherry control mice

Figure 5G

What the paper reports

• CNO main effect: two-way repeated measures ANOVA, Bonferroni, F(1, 56) = 0.97, p = 0.3297 (n = 15 mice)

14 mCherry mice with delay discounting files (README lists 15; R5.5 has none). CNO effect F(1, 52) = 0.94, p = 0.34 (paper F(1, 56) = 0.97, p = 0.33); stable baselines (7 mice) p = 0.34; every per-delay p = 1.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• Post-CNO sessions are averaged per mouse over the first four CNO sessions ('Post CNO administration sessions were averaged per animal across four CNO treatment sessions', Supplementary methods, Chemogenetic manipulation).

In healthy control mice, delay discounting remained consistent across sessions

Figure 6B; Results, Impulsivity develops with repeated pramipexole dosing, paragraph 1

What the paper reports

• first against fourth PPX session: two-way repeated measures ANOVA, Bonferroni, p > 0.99 (n = 16 mice)

Control mice, 12 of 16 with a fourth PPX session; baseline against first p = 0.46, against fourth p = 0.75, first against fourth p = 1 (paper p > 0.99), Bonferroni on condition means; per-delay first against fourth all p = 1.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• The first-against-fourth test compares condition means pooled over delays, with Bonferroni over the three pairs ('the 1st and 4th PPX injections were compared across different delays using two-way repeated measures ANOVA, followed by Bonferroni multiple comparisons', Supplementary methods, Behavior); per-delay comparisons are shown as a variant.

PPX injection induced a modest increase in impulsivity, which became more marked by the fourth session

Figure 6C; Results, Impulsivity develops with repeated pramipexole dosing, paragraph 1

What the paper reports

• first against fourth session: two-way repeated measures ANOVA, Bonferroni, p < 0.0081 (n = 31 mice)
• baseline against first and against fourth: two-way repeated measures ANOVA, Bonferroni, p < 10⁻⁴ (n = 31 mice)

27 of 31 parkinsonian mice with four PPX sessions; delayed/large choices 50.7% at baseline, 41.6% first and 36.2% fourth session; baseline against first p = 2.7e-05, against fourth p = 2.7e-11, first against fourth p = 0.024 (paper p < 0.0081). No single delay differs between first and fourth (smallest p = 0.19).

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

the initial exposure to A77636 treatment resulted in increased impulsivity, which persisted across four consecutive treatments

Figure 6D and E; Results, Impulsivity develops with repeated pramipexole dosing, paragraph 1

What the paper reports

• parkinsonian, first against fourth: two-way repeated measures ANOVA, Bonferroni, p = 0.59 (n = 16 mice)
• parkinsonian, baseline against first and against fourth: two-way repeated measures ANOVA, Bonferroni, p < 10⁻⁴ (n = 16 mice)
• control, first against fourth: two-way repeated measures ANOVA, Bonferroni, p = 0.49 (n = 10 mice)

Parkinsonian (16 mice); baseline against first p = 0.0004, against fourth p < 0.0001, first against fourth p = 1 (paper 0.59). Control (10 mice); p = 1, 0.31, 0.48 (paper first against fourth 0.49).

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

we found approximately 50% loss of TH signal in the rostral dorsolateral and dorsomedial striatum (DMS), with less marked depletion in the ventral striatum (VS

Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 1; Figure 1B

What the paper reports

• DMS against DLS: Kruskal-Wallis, Dunn's, p = 0.09 (n = 44 animals (Table 1, N (animals); the legend gives n = 44 without unit))
• VS against DMS: Kruskal-Wallis, Dunn's, p = 3 × 10⁻⁴ (n = 44 animals (Table 1, N (animals); the legend gives n = 44 without unit))
• VS against DLS: Kruskal-Wallis, Dunn's, p < 10⁻⁴ (n = 44 animals (Table 1, N (animals); the legend gives n = 44 without unit))

Corrected total fluorescence per ROI, normalized to the control mice of the same staining cohort, in 45 parkinsonian mice (Table 1 gives N = 44). Mean losses are DLS 64.3%, DMS 56.8% and VS 35.3% (medians 66.3, 58.4, 39.7%), so DLS and DMS are outside 50 +/- 5%. Kruskal-Wallis p = 3.4e-08; Dunn DLS against DMS p = 0.38, DLS against VS p < 0.0001, DMS against VS p = 0.0001, so the pattern holds. Normalized to all controls, DLS 60.5% and DMS 52.4% (DMS inside the tolerance), and DMS against VS p = 0.052. The measured value is the mean of the DLS and DMS losses. Cohorts 3 and 4 have one control mouse each; without them (40 mice) the losses are 65.7% and 58.1% and the pattern holds (DMS against VS p = 0.0002). No mouse has sections in two cohorts, so which of the 45 mice the paper's 44 leaves out cannot be told.

How the test differed from the paper’s method

• Normalization reference not given beyond 'normalized TH fluorescence levels' (Supplementary methods, Histology); the control mice of the same staining cohort are used, as the deposit README says each cohort includes healthy and parkinsonian samples 'to control for inter-batch staining variability', and all controls pooled is run as a variant. Both readings give a partial outcome.
• VS is the mean of the VLS and VMS ROIs of both sides ('VLS and VMS were then combined into ventral striatum (VS)', Supplementary methods); background is the mean of the four background ROIs of each section (same passage).

Parkinsonian mice showed slightly slower response latencies and learning rates during the first phase ... but eventually achieved similar performance in the second phase

Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 2; Figure S1H-L

What the paper reports

• success rate, blocked learning, sessions 1 to 6: Mann-Whitney U, Bonferroni, session 1 p < 0.01, 2 p = 0.06, 3 p < 0.05, 4 p < 0.01, 5 p = 0.10, 6 p = 0.09, p < 0.01 (n = 34 control and 81 parkinsonian mice)
• latency to nosepoke, blocked learning, sessions 1 to 6: Mann-Whitney U, Bonferroni, session 1 p > 0.99, 2 p > 0.05, 3 p < 0.001, 4 p < 0.01, 5 p < 0.01, 6 p = 0.10, p < 0.001 (n = 34 control and 81 parkinsonian mice)
• success rate, random learning, sessions 1 to 6: Mann-Whitney U, Bonferroni, p = 0.56 (n = 34 control and 81 parkinsonian mice)

From the workbook's Figure S1H-I and S1K-L sheets (32 control and 81 parkinsonian mice; paper 34 and 81). In blocked learning, parkinsonian success lower in sessions 1, 3 and 4 (Bonferroni p = 0.007, 0.0055, 0.0035), latency longer in sessions 3 to 6 (p = 0.0005 to 0.012). In random learning, success is similar in sessions 2 to 6 but different in session 1 (65.9 against 86.9%, p = 0.003), and initiation latency similar in all sessions (p >= 0.13). The criterion's last part, no random session differing, fails on session 1.

How the test differed from the paper’s method

• The S1K-L sheet has a second 'Session1_Park' header at row 118, between session 1 and session 2 parkinsonian blocks, with 30 rows; it is read as session 2 control. Without it, random session 2 has no control data.

PPX had variable effects on A and K in individual parkinsonian mice, but overall led to a decrease in A and increase in K ... PPX treatment led to a significant reduction in AUC in parkinsonian mice

Figure 1I-K; Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 4

What the paper reports

• A: Friedman, Dunn's, p < 10⁻⁴ (n = 31 mice)
• K: Friedman, Dunn's, p = 0.03 (n = 31 mice)
• AUC: Friedman, Dunn's, p < 10⁻⁴ (n = 31 mice)

Four PPX sessions, 30 mice, no failed fit. A falls (median 0.891 to 0.785, Dunn p = 1.9e-06), AUC falls (3.95 to 2.91, p = 2.7e-08), K rises (0.258 to 0.321) without significance (Dunn p = 0.19; Friedman over the three conditions p = 0.032, which is the paper's printed p = 0.03). With the first PPX session, A and AUC fall (p = 0.012, 5.3e-05) and K does not change (p = 1). The same procedure reproduces the paper's A77636 values exactly (Dunn K p = 0.0053, AUC p = 0.0008), so the method matches the paper's; the paper's K result for PPX may come from the Friedman test (asked of the authors). The Dunn p for K depends on which OFF sessions enter the ranks; over the four pairs of ON (four, first) and OFF (every, first) readings it is 0.19, 0.23, 0.49 and 1, never significant. The repeated measures ANOVA of Table 1 gives K an omnibus p = 0.028 and baseline against PPX p = 0.15 (paired) or 0.19 (pooled error), so the printed p = 0.03 also matches that omnibus test.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• Dunn's comparisons are those against baseline, corrected for two comparisons ('between baseline and PPX ON or OFF conditions', Supplementary methods, Behavior).
• PPX OFF is every washout session of the mouse (the claims file gives no reading; Summary_Cognitive.m averages a mouse's sessions within a condition), with the first OFF session as a variant; all four ON and OFF pairs are run.

In parkinsonian mice treated with PPX, the latency to choose small/immediate rewards was shorter than at baseline, and modulation in response latencies according to outcome was absent

Results, Dopamine D2/3R agonist pramipexole causes impulsive decision-making in parkinsonian mice, paragraph 3; Figure S2G and H

What the paper reports

• PPX effect on latency to small reward, parkinsonian: two-way ANOVA, Bonferroni, F(1, 114) = 4.505, p = 0.02 (n = 31 mice)
• reward size effect on latency, parkinsonian on PPX: two-way ANOVA, Bonferroni, F(1, 117) = 0.1514, p = 0.6979 (n = 31 mice)
• reward size effect on latency, parkinsonian at baseline: two-way ANOVA, Bonferroni, F(1, 114) = 7.016, p = 0.0092 (n = 31 mice)

Parkinsonian mice. Small-reward latency is shorter on PPX at 0 and 3 s (1.28 to 1.09 s, 1.38 to 1.02 s) but the PPX effect is F(1, 110) = 2.87, p = 0.093 with four sessions and p = 0.15 with the first (paper F(1, 114) = 4.505, p = 0.02). The reward size effect is significant at baseline (F(1, 114) = 5.33, p = 0.023; paper p = 0.0092) and not on PPX (p = 0.71 with four sessions, 0.19 with the first; paper p = 0.70). Two of the criterion's three parts hold.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• Latency is from the 'cen tri' event (side cue light) to the first side nosepoke of the trial, classed by the reward delivered, per the Supplementary methods' definition; per mouse and condition, the mean over the block's trials and the condition's sessions.

Similarly to our observations with PPX, the differences in K and AUC persisted even after a 48 h A77636 washout period ... Like PPX, A77636 led to a loss of outcome-dependent modulation of choice latency

Results, Dopamine D1R agonist A77636 causes impulsive decision-making in parkinsonian mice, paragraph 1; Figure S3

What the paper reports

• K, A77636 against A77636 OFF: Friedman, Dunn's, p = 0.96 (n = 16 mice)
• AUC, A77636 against A77636 OFF: Friedman, Dunn's, p = 0.43 (n = 16 mice)
• reward size effect on latency, parkinsonian at baseline: two-way ANOVA, F (1, 52) = 11.70, p = 0.0012 (n = 16 mice)
• reward size effect on latency, parkinsonian on A77636: two-way ANOVA, F (1, 55) = 3.463, p = 0.07 (n = 16 mice (legend); Table 1 gives N = 10)

K and AUC after washout (0.209, 4.93) do not differ from A77636 (Dunn p = 1 and 0.65, three comparisons; paper 0.96, 0.43) and stay on the A77636 side of baseline (0.115, 5.82; Dunn baseline against OFF p = 0.043 each). No per-delay center or side omission comparison is significant (smallest p = 0.65). The reward size effect on latency is present at baseline (F(1, 52) = 14.70, p = 0.0003; paper F(1, 52) = 11.70, p = 0.0012) and also on A77636 (F(1, 56) = 4.84, p = 0.032; paper p = 0.07), where the criterion requires p >= 0.05. The K and AUC results hold under all four ON and OFF pairs (ON against OFF Dunn p >= 0.65).

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

injection of CNO increased locomotion in both hM4D i and hM3D q animals, but not in mCherry control animals

Figure 5C-F; Results, Chemogenetic manipulations of striatal neurons mimic the effect of pramipexole in parkinsonian mice, paragraph 2

hM3Dq mice move faster 20 to 30 min after CNO (median 2.20 to 4.35 cm/s, 11 of 11, Wilcoxon p = 0.001) and hM4Di mice too (3.83 to 6.13 cm/s, 9 of 10, p = 0.004); mCherry mice move slower (2.96 to 2.23 cm/s, 1 of 12 faster, p = 0.016), where the criterion needs p >= 0.05. H0 and H0.4 have identical velocity traces (one recording deposited twice); without H0.4, hM4Di gives p = 0.008 (8 of 9). Twelve files (7 hM4Di and 5 mCherry, every 2023 H and J file except H1 and J3, whose gaps start at 20.96 and 19.57 min) have no tracking gap starting 19.5 to 21 min, and their per-minute velocity from 15 to 25 min does not mark an injection. In the 21 files of listed mice with a gap at 19.5 to 21 min, mCherry mice are slower in 6 of 7 but not significantly (p = 0.16 with injection at 20 min, 0.22 with injection at the gap), and only 3 hM4Di files remain (p = 0.5; a signed-rank test on 3 cannot go below 0.25), so that subset cannot test the claim. The outcome is the one under injection at 20 min in every file. Four files (R0, F0, F3, R1.2) are of mice in no DREADD list and are left out.

How the test differed from the paper’s method

• Injection is assumed at 20 min after the start of each file, as in fig2bc-a77636-locomotion, which the Supplementary methods' '-20 to 0 min' baseline before CNO supports; twelve files (2023 H and J mice other than H1 and J3) have no tracking gap starting 19.5 to 21 min, so for them the time rests on the assumption alone. The files with a gap, with injection at 20 min or at the gap, are shown as variants.
• H0.4's file duplicates H0's; both are kept in the main test, as the README lists both mice, and the result without H0.4 is shown.

Chemogenetic inhibition of iMSNs or excitation of dMSNs robustly shifted choices towards immediate/small rewards over delayed/large rewards

Figure 5H and I

What the paper reports

• hM4Di, CNO main effect: two-way repeated measures ANOVA, F(1, 44) = 22.22, p < 10⁻⁴ (n = 12 mice)
• hM4Di per delay, 0 / 3 / 6 / 9 s (3 s given as p < 0.01): Bonferroni, 0 s p > 0.99, 3 s p < 0.01, 6 s p = 0.049, 9 s p = 0.027, p = 0.049
• hM3Dq, CNO main effect: two-way repeated measures ANOVA, F(1, 44) = 42.30, p < 10⁻⁴ (n = 12 mice)
• hM3Dq per delay: Bonferroni, 0 s p = 0.29, 3 s p = 0.0002, 6 s p = 0.0008, 9 s p = 0.04, p = 2 × 10⁻⁴

hM3Dq (12 mice); CNO F(1, 44) = 40.94, p = 8.8e-08 (paper 42.30), lower at 3, 6 and 9 s (p = 0.0006, 0.0013, 0.026). hM4Di (12 mice); F(1, 44) = 12.77, p = 0.0009 (paper 22.22), lower at 3 s only (p = 0.022; 6 s p = 0.38, 9 s p = 0.62). The criterion's partial case holds (one group).

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.

in parkinsonian mice treated with PPX, A values were positively correlated with residual TH immunofluorescence in the DMS, but not in the DLS or VS

Figure 3A; Results, Changes in sensitivity to reward size relate to dopaminergic lesions in the striatum

What the paper reports

• DMS: Pearson correlation, r = 0.51, p = 0.006 (n = 28 mice)
• DLS: Pearson correlation, r = 0.30, p = 0.11 (n = 28 mice)
• VS: Pearson correlation, r = 0.35, p = 0.07 (n = 28 mice)

A on PPX (four sessions) against normalized DMS TH in the 24 parkinsonian mice with both measures (paper 28); r = 0.32, p = 0.13, outside 0.51 +/- 0.051. DLS r = 0.23 (p = 0.29) and VS r = 0.28 (p = 0.18). Normalized to all controls, DMS r = 0.14 (p = 0.52). With the first PPX session, r = 0.08 (p = 0.71) per cohort and 0.07 (p = 0.75) with all controls, so no reading reproduces. Of the 30 PPX mice, six have no histology (C3.1, E3.3, E4, H4.1, N1, N5.1); E6 has histology but no PPX session; two histology mice (E3, E3.2) are in no README list. E3.3 and E3 or E3.2 may be one mouse under two names, but nothing in the deposit links them, so the join uses exact identifiers.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.
• TH as in fig1b-th-loss (per staining cohort), A as in fig1ijk-ppx-a-k-auc. Mice are matched by their identifiers in the two workbooks; 24 of the 30 PPX mice have histology.

chemogenetic inhibition of indirect pathway or excitation of direct pathway significantly increased the K value and decreased AUC ... CNO did not impact the A value in the hM4D i -expressing group, but significantly reduced the A value in the hM3D q -expressing group

Figure 5J-L; Results, Chemogenetic manipulations of striatal neurons mimic the effect of pramipexole in parkinsonian mice, paragraph 2

What the paper reports

• mCherry A / K / AUC: Wilcoxon signed-rank (Table 1), A p = 0.89, K p = 0.45, AUC p = 0.56, p = 0.45 (n = 15 mice)
• hM4Di A: Wilcoxon signed-rank (Table 1), p = 0.11 (n = 12 mice)
• hM4Di K: Wilcoxon signed-rank (Table 1), p = 0.02 (n = 12 mice)
• hM4Di AUC: Wilcoxon signed-rank (Table 1), p < 0.001 (n = 12 mice)
• hM3Dq A: Wilcoxon signed-rank (Table 1), p = 0.01 (n = 12 mice)
• hM3Dq K: Wilcoxon signed-rank (Table 1), p < 0.01 (n = 12 mice)
• hM3Dq AUC: Wilcoxon signed-rank (Table 1), p < 0.001 (n = 12 mice)

Reading (a), Wilcoxon (Table 1). hM3Dq; A 0.923 to 0.880 (p = 0.012), K 0.133 to 0.198 (p = 0.0015), AUC 5.10 to 4.31 (p = 0.0005). hM4Di; A unchanged (p = 0.20), AUC 6.08 to 5.78 (p = 0.043), K 0.123 to 0.136 (7 of 12 higher, p = 0.20; paper p = 0.02). mCherry; no change (p >= 0.36). Reading (b), Friedman, reduces to a sign test with two conditions and finds fewer effects (hM4Di AUC p = 0.15). Eight of the nine parts hold under reading (a); the K rise in hM4Di does not. The criterion is reproduced when all hold and partially reproduced when the K and AUC results hold but the A pattern does not; here the A pattern holds and a K result fails, so by its wording the claim is not reproduced, narrowly. Changed from partially reproduced after verification round 1.

How the test differed from the paper’s method

• Choices, omissions and latencies are recounted from the raw operant logs, as the paper's per-mouse spreadsheets are not deposited, with the event and block names of the paper's code (UCSF-Nelson-Lab/ICD-Cognitive-Analysis at 5528e222, delay_discounting_task.py, processCsv_1; free choice blocks B3, B6, B9, B12).
• The two-way 'repeated measures' ANOVA treats each mouse and delay pair as the repeated unit with delay between units, the layout implied by the degrees of freedom the paper prints (F(1, 60) for 16 control mice, F(1, 120) for 31 parkinsonian mice; Figure 1 legend and Table 1); per-delay Bonferroni comparisons use the pooled within-unit error, with paired t tests as a variant.
• The six control mice of the 2020 cohort ran delays of 0, 6, 12 and 18 s (their block labels and the measured reward delays agree), and 30 of their sessions have no fourth block; their blocks are counted as the other mice's, as the paper's code does.