Datasets / DANDI:000946

DANDI:000946

✓ Verified by an independent agent

Neural Pathways Modulation in the Anesthetized Rat Elicited by Trials of Transcranial Focused Ultrasound Stimulation

electrophysiological Norway rat 65.2 GB 1,920 files
  • 2 not reproduced
  • 31 blocked

Tested on dataset version 0.250829.1249

View on DANDI Paper JSON How to read outcomes

Claims

From the paper Transcranial Focused Ultrasound Modulates Feedforward and Feedback Cortico-Thalamo-Cortical Pathways by Selectively Activating Excitatory Neurons

Reproducing Gao et al. (2025) from dandiset 000946 3

Notebook reproduce_gao_2025.ipynb

Not reproduced2

  • Not reproducedFigure 4 legendverifier disagreed at firstquestion to the reader

    In the 3000 Hz PRF (60% DC) sessions 312 RSUs and 54 FSUs were recorded from S1 and 233 RSUs and 17 FSUs from POm; in the 30 Hz PRF (0.6% DC) sessions 148 RSUs and 61 FSUs from S1 and 189 RSUs and 43 FSUs from POm.

    There are 312 RSUs and 54 FSUs recorded from S1, 233 RSUs and 17 FSUs recorded from POm for PRF 3,000 Hz. There are 148 RSUs and 61 FSUs recorded from S1 and189 RSUs and 43 FSUs recorded from POm for PRF 30 Hz.

    Figure 4 legend

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification); no file holds waveforms, and celltype_label is 0 in all 1516 files that have it, so the RSU/FSU split cannot be made. The k-means of the claims file header puts every unit in one of its two clusters, so RSU + FSU is the number of units with more than 2000 spikes under every typing reading, and an area total more than 10% from the paper's RSU + FSU rules out both the reproduced outcome (every count within 10%) and the partial one (area totals within 10%), whatever the split. Units with more than 2000 spikes in distinct recordings with tFUS at S1 and PD 200 us. POm has 352 units at 3000 Hz (339 without the BH499 POM_S1BF recording) against 250, and 315 at 30 Hz against 232, ratios 1.36 to 1.41, so the POm totals are outside 10% under both POm selections. S1 has, at 3000 Hz, 572 over all S1 recordings, 383 in the rats with POm recordings and 234 in the sessions with a POm file (509, 320 and 234 without the suffixed recordings NEAR, NEW and 2; 591 with the S1_R and S1BF files) against 366; at 30 Hz 530, 390 and 213 (624 with S1_R) against 209. One S1 selection matches at 3000 Hz and another at 30 Hz, but none matches both, and POm matches in neither. The POm units come from all 10 POm rats (6 to 65 each at 3000 Hz). The totals depend on which rats and which unit rule are used, though nothing marks a rat for exclusion and the criterion fixes the rule at more than 2000 spikes. Leaving out two of the 10 POm rats brings all three POm totals within 10% for 9 of the 45 pairs (BH511 and BH513 give 227, 241 and 1052). With more than 8000 spikes the POm totals are 242, 230 and 1031, all within 10%, while no S1 selection fits all three at 1000, 2000, 4000 or 8000 spikes; for AC, more than 4000 spikes gives 445 units, inside 453 +/- 45.3. So the paper's unit rule or spike counting may differ from more than 2000 spikes over the whole file. The RSU/FSU split was not tested.

    How the test differed from the paper’s method

    • The RSU/FSU split cannot be made (no waveforms or cell types), so the claim is judged on the area totals, which are the same under every typing reading of the claims file header.
    • Which sessions count beyond the sessions of the S1 and POm rats is not fixed; every selection the identifiers allow is shown, and the outcome is the same under all of them because the POm totals are outside 10% under each.
    • Recordings deposited twice (identical spike times and trial starts) are counted once.
  • Not reproducedFigure 6 legendverifier disagreed at first

    The PRF series (PD 200 us) comprises 1050 RSUs and 247 FSUs from S1 and 913 RSUs and 106 FSUs from POm.

    The comparison of neuronal responses for sorted 1,050 RSUs (A) and 247 FSUs (B) recorded from S1 was conducted using Kruskal–Wallis ... The comparison of neuronal responses for sorted 913 RSUs (C) and 106 FSUs (D) recorded from POm was conducted using Kruskal–Wallis.

    Figure 6 legend

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification); no file holds waveforms, and celltype_label is 0 in all 1516 files that have it, so the RSU/FSU split cannot be made. The k-means of the claims file header puts every unit in one of its two clusters, so RSU + FSU is the number of units with more than 2000 spikes under every typing reading, and an area total more than 10% from the paper's RSU + FSU rules out both the reproduced outcome (every count within 10%) and the partial one (area totals within 10%), whatever the split. Over the five PRFs at PD 200 us (units with more than 2000 spikes, distinct recordings, tFUS at S1), POm has 1470 units (1447 without the POM_S1BF recordings) against 913 + 106 = 1019, ratio 1.44 (1.42), outside 10%. S1 has 2414 over all S1 recordings, 1726 in the rats with POm recordings and 986 in the sessions with a POm file (2298, 1648 and 986 without the suffixed recordings; 2534 with the S1_R and S1BF files) against 1297, none within 10%. As for fig4-unit-counts, leaving out two POm rats (9 of 45 pairs) or counting only units with more than 8000 spikes (POm 1031) brings the POm total within 10%, and no S1 selection fits at any of the four spike rules shown, so the paper's unit rule may differ from the criterion's. The RSU/FSU split was not tested.

    How the test differed from the paper’s method

    • The RSU/FSU split cannot be made (no waveforms or cell types), so the claim is judged on the area totals, which are the same under every typing reading of the claims file header.
    • Which sessions count beyond the sessions of the S1 and POm rats is not fixed; every selection the identifiers allow is shown, and none brings either area total within 10%.
    • Recordings deposited twice (identical spike times and trial starts) are counted once.

Blocked1

  • BlockedResults, Auditory confound on CTC pathway modulation using tFUS, paragraph 1; Figure 11A and 11B· Methods underspecifiedverifier disagreed at firstquestion to the reader

    AC neurons in the S1-AC rats show time-locked responses to 1500 and 3000 Hz PRF and postsonication responses to 30, 1500, 3000 and 4500 Hz, both PRF dependent, least at 300 Hz.

    During sonication, the AC neurons showed time-locked responses to 1,500 Hz (n = 107, p = 9.9 × 10−5) and 3,000 Hz PRF (n = 104, p = 1.98 × 10−7). Postsonication, neuronal responses were activated by 30 Hz (n = 83, p = 1.77 × 10−20), 1,500 Hz, (n = 107, p = 1.36 × 10−5), 3,000 Hz (n = 104, p = 4.47 × 10−13), and 4,500 Hz (n = 100, p = 1.66 × 10−40).

    Results, Auditory confound on CTC pathway modulation using tFUS, paragraph 1; Figure 11A and 11B

    What the paper reports

    • across PRFs, during sonication: Kruskal-Wallis, p = 3.2 × 10⁻⁴ (n = 453 AC neurons)
    • across PRFs, postsonication: Kruskal-Wallis, p = 2.7 × 10⁻⁴³ (n = 453 AC neurons)
    • 1500 Hz, during: post hoc one-tail two-sample Wilcoxon against baseline, p = 9.9 × 10⁻⁵, Bonferroni-corrected (n = 107 AC neurons)
    • 3000 Hz, during: post hoc one-tail two-sample Wilcoxon against baseline, p = 2 × 10⁻⁷, Bonferroni-corrected (n = 104 AC neurons)
    • 30 Hz, postsonication: post hoc one-tail two-sample Wilcoxon against baseline, p = 1.8 × 10⁻²⁰, Bonferroni-corrected (n = 83 AC neurons)
    • 1500 Hz, postsonication: post hoc one-tail two-sample Wilcoxon against baseline, p = 1.4 × 10⁻⁵, Bonferroni-corrected (n = 107 AC neurons)
    • 3000 Hz, postsonication: post hoc one-tail two-sample Wilcoxon against baseline, p = 4.5 × 10⁻¹³, Bonferroni-corrected (n = 104 AC neurons)
    • 4500 Hz, postsonication: post hoc one-tail two-sample Wilcoxon against baseline, p = 1.7 × 10⁻⁴⁰, Bonferroni-corrected (n = 100 AC neurons)

    AC units with more than 2000 spikes in the 17 distinct AC recordings with tFUS at S1 and PD 200 us, from 6 rats; 112, 89, 138, 167 and 127 units at 30, 300, 1500, 3000 and 4500 Hz, 633 in all (expected 453 +/- 45.3, so the count does not match). The seventh AC rat, BH532, has no PD 200 us recording, and the AC_1 and AC_2 files of BH538 are at 1500 Hz and PD 40 us. The four readings of the claims file header give different outcomes, and the header blocks a claim when they do. N1 with B1 is partially reproduced: Kruskal-Wallis p = 7.83e-06 (0 to 67 ms) and 5.16e-49 (201 to 268 ms); during sonication 1500 Hz (140.3 +/- 8.8%, corrected p = 2.46e-08), 3000 Hz (107.1 +/- 1.2%, 9.72e-11) and also 4500 Hz (108.0 +/- 1.5%, 7.05e-05) are above baseline and 30 Hz is not (105.9%, 0.0189); postsonication exactly 30, 1500, 3000 and 4500 Hz are above baseline (176.4, 147.9, 310.5 and 210.8%) and 300 Hz is not (94.3%); 300 Hz has the lowest mean in both windows. N1 with B2 is not reproduced: 300 Hz is also above baseline during sonication (99.7%, corrected p = 6.29e-03). N2 with B1 and with B2 are not reproduced (Kruskal-Wallis p = 0.0213 during sonication). No reading is better supported: N2 gives 49.8 to 61.7% at trial starts shifted by 1 s, where N1 gives 94.3 to 99.0%; under N1 the B1 sample is 100% for every unit, so the two-sample test becomes a test against a constant; B2 uses a window the paper never names. No reading reproduces the claim. With one rat left out, N1 with B1 gives reproduced (without BH537), partially reproduced (without BH530, BH535 or BH538) or not reproduced (without BH539 or BH547, where 30 Hz joins the during set), and N1 with B2 gives reproduced without BH530, BH537 or BH538; 300 Hz has the lowest mean in both windows in all twelve runs. The statement's PRF dependence (both Kruskal-Wallis tests) holds under N1. The question of which normalization and baseline sample the paper used is for the paper's authors (finding methods_unavailable); whether the header's readings N2 and B1 are readings of the paper is asked of a reader (questions/000946.yaml).

    How the test differed from the paper’s method

    • The normalization (N1 or N2) and the baseline sample of the post hoc test (B1 or B2) are not fixed by the paper (Statistical analysis, "normalized to the pre-stim period (67 ms pre-stim)"); all four readings were run and they give different outcomes.
    • The Bonferroni factor is the five PRF levels of the panel (claims file header, reading C1); reading C2 applies only to a panel with a single comparison.
    • The dataset holds AC recordings at PD 200 us from 6 rats (7 AC rats in all) where the paper reports 12 S1-AC rats, and it gives 633 units with more than 2000 spikes where the paper reports 453; the paper's selection of AC units could not be matched.
    • Each PRF comes from 3 or 4 of the 6 rats, and the rats differ by PRF, so the Kruskal-Wallis test across PRFs mixes PRF with rat (for example BH538 averages 654.1% at 3000 Hz, 201 to 268 ms, against 310.5% pooled); the paper also pools units across rats.
    • Recordings deposited twice (identical spike times and trial starts) are counted once.
    • Trial start_time is taken as sonication onset; the clock check shows the firing changes after it and not 1 s later, which is consistent with this but does not prove there is no fixed delay.

Not tested in a notebook 30

30 data missing from the dataset

Blocked30

  • BlockedResults, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4A· Data missing from the dataset

    S1 RSUs are activated during sonication (0 to 67 ms) with 3000 Hz PRF, 60% DC tFUS aimed at S1.

    RSUs from S1 (n = 312, p = 3.13 × 10−7) and POm (n = 233, p = 0.0003) were both activated during 3,000 Hz PRF sonication

    Results, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4A

    What the paper reports

    • Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 3.1 × 10⁻⁷, Bonferroni-corrected (n = 312 S1 RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. As an untyped variant that decides nothing (notebook, Untyped responses), all units with more than 2000 spikes, reading N1 and B1: all S1 units at 3000 Hz (n = 572) have a mean normalized rate of 114% in 0 to 67 ms, above baseline at corrected p < 0.01.

  • BlockedResults, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4B· Data missing from the dataset

    POm RSUs are activated during sonication with 3000 Hz PRF tFUS aimed at S1, in synchrony with S1 RSUs.

    RSUs from S1 (n = 312, p = 3.13 × 10−7) and POm (n = 233, p = 0.0003) were both activated during 3,000 Hz PRF sonication

    Results, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4B

    What the paper reports

    • Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 3 × 10⁻⁴, Bonferroni-corrected (n = 233 POm RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. As an untyped variant that decides nothing (notebook, Untyped responses), all units with more than 2000 spikes, reading N1 and B1: all POm units at 3000 Hz (n = 352) have 105% in 0 to 67 ms, above baseline at corrected p < 0.01.

  • BlockedResults, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4A and 4B· Data missing from the dataset

    After 3000 Hz PRF sonication S1 RSUs, POm RSUs and S1 FSUs show delayed activation from 134 ms.

    Postsonication activations were presented for S1 RSUs (p = 5 × 10−18), POm RSUs (p = 0.0007), and S1 FSUs (n = 54, p = 4.03 × 10−8) from 134 ms.

    Results, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4A and 4B

    What the paper reports

    • S1 RSUs: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 5 × 10⁻¹⁸, Bonferroni-corrected (n = 312 S1 RSUs)
    • POm RSUs: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 7 × 10⁻⁴, Bonferroni-corrected (n = 233 POm RSUs)
    • S1 FSUs: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 4 × 10⁻⁸, Bonferroni-corrected (n = 54 S1 FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. As an untyped variant that decides nothing (notebook, Untyped responses), all units with more than 2000 spikes, reading N1 and B1: at 3000 Hz all S1 units are at 119% (134 to 201 ms) and 178% (201 to 268 ms), all POm units at 113% and 122%, each above baseline at corrected p < 0.01.

  • BlockedResults, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4 legend· Data missing from the dataset

    With 30 Hz PRF (0.6% DC), S1 RSUs show an inhibitory response after sonication (67 to 134 ms) and POm RSUs show delayed excitatory activity (134 to 268 ms).

    At 30 Hz PRF, the S1 RSUs showed an inhibitory response after sonication (67–134 ms, n = 148, p = 0.0029) and POm RSUs expressed delayed excitatory activities (n = 189, p = 2 × 10−9).

    Results, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4 legend

    What the paper reports

    • S1 RSUs, 67 to 134 ms, decrease: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 0.0029, Bonferroni-corrected (n = 148 S1 RSUs)
    • POm RSUs, delayed increase: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 2 × 10⁻⁹, Bonferroni-corrected (n = 189 POm RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. As an untyped variant that decides nothing (notebook, Untyped responses), all units with more than 2000 spikes, reading N1 and B1: at 30 Hz all S1 units are at 96% in 67 to 134 ms (below baseline at corrected p < 0.01) and all POm units at 111% and 110% in 134 to 201 and 201 to 268 ms (above baseline).

  • BlockedResults, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4· Data missing from the dataset

    S1 FSUs show no time-locked response to 3000 Hz PRF, S1 FSUs show no activity change to 30 Hz PRF, and POm FSUs show no change to either PRF.

    There was no activity change among S1 FSUs (n = 61) to 30 Hz PRF and POm FSUs to both PRFs (3,000 Hz, n = 17; 30 Hz, n = 43).

    Results, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 1; Figure 4

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made.

  • BlockedResults, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 2; Figure 5A· Data missing from the dataset

    In sham sessions (no tFUS), S1 and POm RSUs and FSUs show no time-locked and no delayed (201 to 268 ms) activity change.

    In sham condition (no tFUS), S1 neurons and POm neurons showed no time-locked (S1 RSUs: n = 31, p = 1.0; S1 FSUs: n = 53, p = 0.55; POm RSUs: n = 67, p = 1.0; POm FSUs: n = 67, p = 1.0)

    Results, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 2; Figure 5A

    What the paper reports

    • S1 FSUs, time-locked: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 0.55, Bonferroni-corrected (n = 53 S1 FSUs)
    • S1 RSUs, POm RSUs and POm FSUs, time-locked and delayed, and S1 FSUs delayed: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 1, Bonferroni-corrected (n = 31 S1 RSUs, 53 S1 FSUs, 67 POm RSUs, 67 POm FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. No recording is marked as sham or as tFUS aimed at CG1: none of the 816 distinct recordings has an identifier, general/notes or general/data_collection naming sham, no tFUS, CG1 or cingulate; every recording gives a PRF and a PD (in data_collection or in the identifier), and the targets named are S1, M1 and POm.

  • BlockedResults, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 2; Figure 5B· Data missing from the dataset

    With 3000 Hz PRF tFUS aimed at CG1, S1 and POm RSUs and FSUs show no time-locked and no delayed (201 to 268 ms) activity change.

    In control condition (tFUS targeted at CG1), S1 neurons and POm neurons showed no time-locked (S1 RSUs: n = 12, p = 1.0; S1 FSUs: n = 34, p = 1.0; POm RSUs: n = 55, p = 1.0; POm FSUs: n = 42, p = 0.33) and delayed responses (201–268 ms, S1 RSUs: p = 1.0; S1 FSUs: p = 0.10; POm RSUs: p = 0.32, POm FSUs: p = 1.0).

    Results, Synchronizing neuronal responses to tFUS in CTC pathway, paragraph 2; Figure 5B

    What the paper reports

    • POm FSUs, time-locked: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 0.33, Bonferroni-corrected (n = 42 POm FSUs)
    • S1 FSUs, delayed: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 0.1, Bonferroni-corrected (n = 34 S1 FSUs)
    • POm RSUs, delayed: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 0.32, Bonferroni-corrected (n = 55 POm RSUs)
    • S1 RSUs, S1 FSUs and POm RSUs time-locked, and S1 RSUs and POm FSUs delayed: Kruskal-Wallis, post hoc one-tail two-sample Wilcoxon against baseline, p = 1, Bonferroni-corrected (n = 12 S1 RSUs, 34 S1 FSUs, 55 POm RSUs, 42 POm FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. No recording is marked as sham or as tFUS aimed at CG1: none of the 816 distinct recordings has an identifier, general/notes or general/data_collection naming sham, no tFUS, CG1 or cingulate; every recording gives a PRF and a PD (in data_collection or in the identifier), and the targets named are S1, M1 and POm. The sessions with tFUS aimed at M1 are not the CG1 control of the paper (Methods, CG1 at AP 3, ML 0), and the paper does not describe them.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6A· Data missing from the dataset

    With constant 200 us PD, time-locked S1 RSU responses are elicited only by high PRFs (1500, 3000 and 4500 Hz), the highest at 1500 Hz (123.34 +/- 4.3%), and the during-sonication response differs across PRFs.

    time-locked S1 RSUs’ responses could only be elicited from high PRFs: 1,500 Hz (n = 284, p = 2.74 × 10−6 compared with baseline), 3,000 Hz (n = 312, p = 3.13 × 10−7), and 4,500 Hz (n = 152, p = 0.0028), among which 1,500 Hz evoked highest response level of 123.34 ± 4.3%

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6A

    What the paper reports

    • across PRFs, 0 to 67 ms: Kruskal-Wallis, p = 1.5 × 10⁻¹⁵ (n = 1050 S1 RSUs)
    • 1500 Hz against baseline: post hoc one-tail two-sample Wilcoxon, p = 2.7 × 10⁻⁶, Bonferroni-corrected (n = 284 S1 RSUs)
    • 3000 Hz against baseline: post hoc one-tail two-sample Wilcoxon, p = 3.1 × 10⁻⁷, Bonferroni-corrected (n = 312 S1 RSUs)
    • 4500 Hz against baseline: post hoc one-tail two-sample Wilcoxon, p = 0.0028, Bonferroni-corrected (n = 152 S1 RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6A· Data missing from the dataset

    Delayed (201 to 268 ms) S1 RSU responses occur at PRFs above 30 Hz, the highest at 4500 Hz (148.39 +/- 5.41%), and are PRF dependent.

    The delayed responses (201–268 ms) of RSUs could be observed at PRFs higher than 30 Hz and the highest response occurred at 4,500 Hz (148.39 ± 5.41%).

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6A

    What the paper reports

    • Kruskal-Wallis across PRFs, postsonication, p = 5.1 × 10⁻¹³ (n = 1050 S1 RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6B· Data missing from the dataset

    S1 FSUs show no time-locked response to any PRF and delayed responses only at PRFs above 1500 Hz, highest at 4500 Hz (158.75 +/- 10.75%), PRF dependent postsonication.

    FSUs did not show time-locked responses to any PRFs and only presented delayed responses at PRFs higher than 1,500 Hz, with the highest response level at 4,500 Hz (158.75 ± 10.75%).

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6B

    What the paper reports

    • Kruskal-Wallis across PRFs, postsonication, p = 5.2 × 10⁻¹³ (n = 247 S1 FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6C· Data missing from the dataset

    POm RSUs show time-locked responses to 1500 and 3000 Hz PRF and delayed responses at PRFs from 1500 Hz, highest at 1500 Hz (141.68 +/- 2.38%); both windows are PRF dependent.

    time-locked responses (Fig. 6B) from POm RSUs could be elicited by 1,500 Hz (n = 153, p = 0.003) and 3,000 Hz PRFs (n = 233, p = 0.0003). However, the delayed responses (201–268 ms) of POm RSUs were induced by PRFs higher than 1,500 Hz and were highest (141.68 ± 2.38%) at 1,500 Hz.

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6C

    What the paper reports

    • across PRFs, 0 to 67 ms: Kruskal-Wallis, p = 2.1 × 10⁻²¹ (n = 913 POm RSUs)
    • across PRFs, 201 to 268 ms: Kruskal-Wallis, p = 6 × 10⁻¹⁷ (n = 913 POm RSUs)
    • 1500 Hz against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 0.003, Bonferroni-corrected (n = 153 POm RSUs)
    • 3000 Hz against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 3 × 10⁻⁴, Bonferroni-corrected (n = 233 POm RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6D· Data missing from the dataset

    POm FSUs show no time-locked response to any PRF (though 30 and 1500 Hz differ from each other) and a delayed response only to 4500 Hz; postsonication activity is PRF dependent.

    The FSUs did not show any time-locked responses and only expressed the delayed responses to 4,500 Hz (n = 26, p = 4.2 × 10−6).

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 1; Figure 6D

    What the paper reports

    • 4500 Hz against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 4.2 × 10⁻⁶, Bonferroni-corrected (n = 26 POm FSUs)
    • 30 against 1500 Hz, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 0.004 (n = 106 POm FSUs)
    • across PRFs, 201 to 268 ms: Kruskal-Wallis, p = 4 × 10⁻⁴ (n = 106 POm FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 7A· Data missing from the dataset

    At 3000 Hz PRF, S1 RSU time-locked responses depend on DC and occur at 30% and 60%; delayed responses occur at 6%, 30% and 60%.

    with 3,000 Hz PRF, the time-locked responses of S1 RSUs were dependent on DCs which can be elicited with DC of 30% (n = 194, p = 3.38 × 10−6) and 60% (n = 312, p = 5.96 × 10−6). Delayed activities of RSUs were found at 6% (n = 56, p = 0.001), 30% (n = 194, p = 9.64 × 10−32), and 60% (n = 312, p = 1.38 × 10−15).

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 7A

    What the paper reports

    • across DCs, 0 to 67 ms: Kruskal-Wallis, p = 1.3 × 10⁻¹¹ (n = 645 S1 RSUs)
    • across DCs, 201 to 268 ms: Kruskal-Wallis, p = 2.5 × 10⁻²⁹ (n = 645 S1 RSUs)
    • 30% against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 3.4 × 10⁻⁶, Bonferroni-corrected (n = 194 S1 RSUs)
    • 60% against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 6 × 10⁻⁶, Bonferroni-corrected (n = 312 S1 RSUs)
    • 6% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 0.001, Bonferroni-corrected (n = 56 S1 RSUs)
    • 30% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 9.6 × 10⁻³², Bonferroni-corrected (n = 194 S1 RSUs)
    • 60% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 1.4 × 10⁻¹⁵, Bonferroni-corrected (n = 312 S1 RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The duty cycle series sessions are in the dataset (the S1 and POm recordings with tFUS at S1 have 19 PD values from 1.4 us to 30000 us, data check), but without cell types.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 7B· Data missing from the dataset

    At 3000 Hz PRF, S1 FSUs show delayed responses at 6%, 30% and 60% DC, the same range as S1 RSUs, and 0.6% differs from 30% postsonication.

    Delayed responses of FSUs could be elicited by 6% (n = 22, p = 1.29 × 10−8), 30% (n = 49, p = 2.32 × 10−11), and 60% (n = 54, p = 2.38 × 10−8) DCs (Fig. 7B), which were the same range of S1 RSUs.

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 7B

    What the paper reports

    • 6% against baseline: post hoc one-tail two-sample Wilcoxon, p = 1.3 × 10⁻⁸, Bonferroni-corrected (n = 22 S1 FSUs)
    • 30% against baseline: post hoc one-tail two-sample Wilcoxon, p = 2.3 × 10⁻¹¹, Bonferroni-corrected (n = 49 S1 FSUs)
    • 60% against baseline: post hoc one-tail two-sample Wilcoxon, p = 2.4 × 10⁻⁸, Bonferroni-corrected (n = 54 S1 FSUs)
    • 0.6% against 30%, postsonication: post hoc one-tail two-sample Wilcoxon, p = 0.002 (n = 171 S1 FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The duty cycle series sessions are in the dataset (the S1 and POm recordings with tFUS at S1 have 19 PD values from 1.4 us to 30000 us, data check), but without cell types.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 7C· Data missing from the dataset

    At 3000 Hz PRF, POm RSU time-locked responses show no DC dependence (highest at 60%, 106 +/- 0.77%), while delayed responses are DC dependent and occur at DCs of 6% and above.

    The POm RSUs’ time-locked responses did not show DC dependence and were highest at 60% (106 ± 0.77%; Fig. 7C). Delayed responses could be activated by DC higher than 6% (6%, n = 22, p = 1.55 × 10−6; n = 7, p = 0.001; 60%, n = 233, p = 1.28 × 10−6; 90%, n = 18, p = 0.006).

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 7C

    What the paper reports

    • across DCs, 201 to 268 ms: Kruskal-Wallis, p = 3.6 × 10⁻⁴ (n = 286 POm RSUs)
    • 6% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 1.5 × 10⁻⁶, Bonferroni-corrected (n = 22 POm RSUs)
    • unlabelled DC (read as 30%) against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 0.001, Bonferroni-corrected (n = 7 POm RSUs)
    • 60% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 1.3 × 10⁻⁶, Bonferroni-corrected (n = 233 POm RSUs)
    • 90% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 0.006, Bonferroni-corrected (n = 18 POm RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The duty cycle series sessions are in the dataset (the S1 and POm recordings with tFUS at S1 have 19 PD values from 1.4 us to 30000 us, data check), but without cell types.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 7D· Data missing from the dataset

    At 3000 Hz PRF, POm FSUs show no time-locked response against baseline, though 30% DC is higher than 0.6% and 6% during sonication; delayed responses occur at 30%, 60% and 90%.

    There was no time-locked responses of POm FSUs compared with baseline while 30% DC showed higher activities compared with 0.6% (p = 0.0001) and 6% DC (p = 0.0003). The delayed activities of POm FSUs occurred at 30% (n = 18, p = 0.003), 60% (n = 17, p = 0.005), and 90% (n = 52, p = 6.8 × 10−5) DCs

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 7D

    What the paper reports

    • across DCs, 0 to 67 ms: Kruskal-Wallis, p = 3.3 × 10⁻⁴ (n = 129 POm FSUs)
    • 30% against 0.6%, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 10⁻⁴ (n = 129 POm FSUs)
    • 30% against 6%, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 3 × 10⁻⁴ (n = 129 POm FSUs)
    • 30% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 0.003, Bonferroni-corrected (n = 18 POm FSUs)
    • 60% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 0.005, Bonferroni-corrected (n = 17 POm FSUs)
    • 90% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 6.8 × 10⁻⁵, Bonferroni-corrected (n = 52 POm FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The duty cycle series sessions are in the dataset (the S1 and POm recordings with tFUS at S1 have 19 PD values from 1.4 us to 30000 us, data check), but without cell types.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 8A· Data missing from the dataset

    At 30 Hz PRF, S1 RSUs show time-locked responses at 6%, 30% and 90% DC and a delayed response at 90%; both windows are DC dependent.

    With 30 Hz PRF stimulation (Fig. 8A), the S1 RSUs showed time-locked responses to DC of 6% (n = 81, p = 1.19 × 10−8), 30% (n = 61, p = 0.001), and 90% (n = 66, p = 2.17 × 10−6) and delayed responses to 90% DC (n = 66, p = 1.5 × 10−24).

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 8A

    What the paper reports

    • across DCs, 0 to 67 ms: Kruskal-Wallis, p = 7 × 10⁻¹² (n = 442 S1 RSUs)
    • across DCs, 201 to 268 ms: Kruskal-Wallis, p = 10⁻²⁷ (n = 442 S1 RSUs)
    • 6% against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 1.2 × 10⁻⁸, Bonferroni-corrected (n = 81 S1 RSUs)
    • 30% against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 0.001, Bonferroni-corrected (n = 61 S1 RSUs)
    • 90% against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 2.2 × 10⁻⁶, Bonferroni-corrected (n = 66 S1 RSUs)
    • 90% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 1.5 × 10⁻²⁴, Bonferroni-corrected (n = 66 S1 RSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The duty cycle series sessions are in the dataset (the S1 and POm recordings with tFUS at S1 have 19 PD values from 1.4 us to 30000 us, data check), but without cell types.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 8B· Data missing from the dataset

    At 30 Hz PRF, 90% DC induces a time-locked increase in S1 FSUs and the only delayed S1 FSU response; FSU responses are DC dependent during sonication but not postsonication.

    Moreover, 90% DC with 30 Hz PRF induced positive time-locked response in S1 FSUs (Fig. 8B; n = 28, p = 4.57 × 10−11). Like S1 RSUs, S1 FSUs only showed delayed responses to 90% DC (n = 28, p = 0.007) with 30 Hz PRF.

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 8B

    What the paper reports

    • across DCs, 0 to 67 ms: Kruskal-Wallis, p = 1.9 × 10⁻⁸ (n = 147 S1 FSUs)
    • 90% against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 4.6 × 10⁻¹¹, Bonferroni-corrected (n = 28 S1 FSUs)
    • 90% against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 0.007, Bonferroni-corrected (n = 28 S1 FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The duty cycle series sessions are in the dataset (the S1 and POm recordings with tFUS at S1 have 19 PD values from 1.4 us to 30000 us, data check), but without cell types.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 9A· Data missing from the dataset

    At constant 60% DC, during sonication only 30 Hz PRF differs from the other PRFs in S1 RSUs, and postsonication activity is highest at 4500 Hz; both windows are PRF dependent.

    only 30 Hz PRF showed different effects to other PRFs (300 Hz, p = 0.008; 1,500 Hz, p = 0.006; 3,000 Hz, p = 0.007; 4,500 Hz, p = 0.0003) on S1 RSUs during sonication, while postsonication activities preferred 4,500 Hz

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 9A

    What the paper reports

    • across PRFs, 0 to 67 ms: Kruskal-Wallis, p = 2.5 × 10⁻⁴ (n = 759 S1 RSUs)
    • across PRFs, 201 to 268 ms: Kruskal-Wallis, p = 1.6 × 10⁻²⁵ (n = 759 S1 RSUs)
    • 30 against 300 Hz, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 0.008
    • 30 against 1500 Hz, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 0.006
    • 30 against 3000 Hz, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 0.007
    • 30 against 4500 Hz, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 3 × 10⁻⁴

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The duty cycle series sessions are in the dataset (the S1 and POm recordings with tFUS at S1 have 19 PD values from 1.4 us to 30000 us, data check), but without cell types.

  • BlockedResults, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 9B· Data missing from the dataset

    At constant 60% DC, S1 FSUs respond to 1500 Hz PRF during sonication and prefer 4500 Hz postsonication; both windows are PRF dependent.

    S1 FSUs responded to 1,500 Hz PRF (n = 48, p = 3.35 ×10−6) during sonication and preferred 4,500 Hz post sonication (n = 18, p = 5.84 × 10−10; Fig. 9B).

    Results, tFUS parameters on affecting neuronal activities of CTC pathway, paragraph 3; Figure 9B

    What the paper reports

    • across PRFs, 0 to 67 ms: Kruskal-Wallis, p = 0.0079 (n = 151 S1 FSUs)
    • across PRFs, 201 to 268 ms: Kruskal-Wallis, p = 5 × 10⁻⁵ (n = 151 S1 FSUs)
    • 1500 Hz against baseline, 0 to 67 ms: post hoc one-tail two-sample Wilcoxon, p = 3.4 × 10⁻⁶, Bonferroni-corrected (n = 48 S1 FSUs)
    • 4500 Hz against baseline, 201 to 268 ms: post hoc one-tail two-sample Wilcoxon, p = 5.8 × 10⁻¹⁰, Bonferroni-corrected (n = 18 S1 FSUs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The duty cycle series sessions are in the dataset (the S1 and POm recordings with tFUS at S1 have 19 PD values from 1.4 us to 30000 us, data check), but without cell types.

  • BlockedResults, tFUS elicited correlation change in CTC pathway, paragraph 1; Figure 10A and its legend· Data missing from the dataset

    With 3000 Hz PRF, 60% DC, the directional correlation from S1 RSUs to POm RSUs and from POm RSUs to S1 FSUs increases during sonication, while the other two directional correlations do not change then; postsonication, S1 RSU to POm FSU and POm RSU to S1 RSU correlation decreases.

    With 3,000 Hz PRF, 60% DC stimulation, S1 RSUs showed enhanced correlation to POm RSUs during sonication (Fig. 10A) ... POm RSUs showed an increased correlation to S1 FSUs during sonication ... Postsonication, we found the correlation from S1 RSUs to POm FSUs and POm RSUs to S1 RSUs significantly decreased

    Results, tFUS elicited correlation change in CTC pathway, paragraph 1; Figure 10A and its legend

    What the paper reports

    • S1 RSU to POm RSU, during sonication: one-tail two-sample Wilcoxon against prestimulation correlation, p = 0.01
    • POm RSU to S1 FSU, during sonication: one-tail two-sample Wilcoxon against prestimulation correlation, p = 0.02

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The dandiset's STTC files (370, under sub-STTC-*) hold STTC values between the units of two areas, but carry no unit identities or cell types, and their arrays have no dimension for the nine dt values their descriptions list, so they cannot stand in either.

  • BlockedResults, tFUS elicited correlation change in CTC pathway, paragraph 1; Figure 10B· Data missing from the dataset

    With 30 Hz PRF, S1 RSU to POm FSU directional correlation increases and POm RSU to S1 FSU correlation decreases during sonication.

    When delivering 30 Hz ultrasound, S1 RSUs to POm FSUs higher correlated (p = 0.0058), while POm RSUs showed decreased correlation to S1 FSUs during sonication (p = 0.0058).

    Results, tFUS elicited correlation change in CTC pathway, paragraph 1; Figure 10B

    What the paper reports

    • S1 RSU to POm FSU, increase: one-tail two-sample Wilcoxon against prestimulation correlation, p = 0.0058
    • POm RSU to S1 FSU, decrease: one-tail two-sample Wilcoxon against prestimulation correlation, p = 0.0058

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The dandiset's STTC files (370, under sub-STTC-*) hold STTC values between the units of two areas, but carry no unit identities or cell types, and their arrays have no dimension for the nine dt values their descriptions list, so they cannot stand in either.

  • BlockedResults, tFUS elicited correlation change in CTC pathway, paragraph 2; Figure 10C· Data missing from the dataset

    With constant 200 us PD, the S1 RSU to POm RSU correlation change during sonication shows no PRF dependence overall, though 1500 Hz gives a higher change than 30 Hz.

    When applying tFUS with constant 200 μs PD, varying the DC and PRF, the correlation did not show PRF-dependent change. A total of 1,500 Hz PRF did show a higher correlation change compared with 30 Hz PRF (Fig. 10C; p = 0.02).

    Results, tFUS elicited correlation change in CTC pathway, paragraph 2; Figure 10C

    What the paper reports

    • 1500 against 30 Hz: post hoc one-tail two-sample Wilcoxon, p = 0.02

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The dandiset's STTC files (370, under sub-STTC-*) hold STTC values between the units of two areas, but carry no unit identities or cell types, and their arrays have no dimension for the nine dt values their descriptions list, so they cannot stand in either.

  • BlockedResults, tFUS elicited correlation change in CTC pathway, paragraph 2; Figure 10D· Data missing from the dataset

    At constant 60% DC, the S1 RSU to POm RSU correlation change is PRF dependent, with a preference for 1500 and 3000 Hz.

    Then we fixed the DC to 60% and varied PRF. Unlike the neuronal responses (Fig. 9), the correlation showed the preference to 1,500 and 3,000 Hz (Fig. 10D).

    Results, tFUS elicited correlation change in CTC pathway, paragraph 2; Figure 10D

    What the paper reports

    • Kruskal-Wallis across PRFs, p = 0.0012

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The dandiset's STTC files (370, under sub-STTC-*) hold STTC values between the units of two areas, but carry no unit identities or cell types, and their arrays have no dimension for the nine dt values their descriptions list, so they cannot stand in either.

  • BlockedResults, tFUS elicited correlation change in CTC pathway, paragraph 2; Figure 10E· Data missing from the dataset

    At 30 Hz PRF, the S1 RSU to POm RSU correlation change depends on DC, changing most at middle DCs (6, 30, 60%) and less at 0.6% and 90%.

    For 30 Hz PRF (Fig. 10E), the correlation tended to change most at a middle range (6, 30, 60%) and respond less to lower (e.g., 0.6%) and higher DC (e.g., 90%).

    Results, tFUS elicited correlation change in CTC pathway, paragraph 2; Figure 10E

    What the paper reports

    • Kruskal-Wallis across DCs, p = 0.0054

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The dandiset's STTC files (370, under sub-STTC-*) hold STTC values between the units of two areas, but carry no unit identities or cell types, and their arrays have no dimension for the nine dt values their descriptions list, so they cannot stand in either.

  • BlockedResults, tFUS elicited correlation change in CTC pathway, paragraph 2; Figure 10F· Data missing from the dataset

    At 3000 Hz PRF, the S1 RSU to POm RSU correlation change does not depend on DC.

    For 3,000 Hz PRF, the pathway correlation showed DC independence (Fig. 10F).

    Results, tFUS elicited correlation change in CTC pathway, paragraph 2; Figure 10F

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The dandiset's STTC files (370, under sub-STTC-*) hold STTC values between the units of two areas, but carry no unit identities or cell types, and their arrays have no dimension for the nine dt values their descriptions list, so they cannot stand in either.

  • BlockedResults, tFUS elicited correlation change in CTC pathway, paragraph 3; Discussion, The feedback and feedforward cortico-thalomo-cortical circuits, paragraph 2· Data missing from the dataset

    The PRFs that elicit stronger local S1 RSU responses are more likely to also activate the CTC pathway, so pathway modulation is PRF dependent.

    our results demonstrate that this pathway modulation is PRF dependent. To be more specific, the PRFs that could elicit stronger local neuronal responses were more likely to also activate CTC pathway activity.

    Results, tFUS elicited correlation change in CTC pathway, paragraph 3; Discussion, The feedback and feedforward cortico-thalomo-cortical circuits, paragraph 2

    Depends on fig6a-s1-rsu-time-locked-prf and fig10c-sttc-constant-pd, both blocked because the dataset holds no waveforms or cell types.

  • BlockedResults, Auditory confound on CTC pathway modulation using tFUS, paragraph 2; Figure 11C legend· Data missing from the dataset

    With 30 Hz PRF, the directional correlation from AC neurons to S1 RSUs increases during sonication and lasts about 150 to 200 ms.

    During tFUS sonication with 30 Hz PRF, the correlation from 878 paired AC neurons to S1 RSUs significantly increased (p = 0.0014) which lasting for ∼150 ms (left).

    Results, Auditory confound on CTC pathway modulation using tFUS, paragraph 2; Figure 11C legend

    What the paper reports

    • one-tail two-sample Wilcoxon against prestimulation correlation, p = 0.0014 (n = 878 AC to S1 RSU pairs)

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The AC units are in the dataset, but the target S1 RSUs and the S1 FSU part need the cell types. The dandiset's STTC files (370, under sub-STTC-*) hold STTC values between the units of two areas, but carry no unit identities or cell types, and their arrays have no dimension for the nine dt values their descriptions list, so they cannot stand in either.

  • BlockedResults, Auditory confound on CTC pathway modulation using tFUS, paragraph 2; Figure 11D legend· Data missing from the dataset

    With 3000 Hz PRF, AC neurons show no directional correlation change with S1 RSUs, and their correlation with S1 FSUs changes only about 200 to 300 ms after onset.

    However, when delivering 3,000 Hz PRF ultrasound energy to S1, auditory neurons did not show directional correlation change with S1 RSUs. The correlation change of AC neurons with S1 FSUs only occurred ∼200–300 ms.

    Results, Auditory confound on CTC pathway modulation using tFUS, paragraph 2; Figure 11D legend

    The paper types units as RSUs or FSUs by k-means on the initial phase and afterhyperpolarization of the spike waveform (Methods, Neuronal type classification). No file holds waveforms (the units tables hold spike_times and celltype_label only, and acquisition, analysis and processing are empty in all 1550 unit files), and celltype_label, described as label 1 = RSU, label 2 = FSU, is 0 in all 1516 files that have it and absent from the other 34. The RSU/FSU split cannot be made. The AC units are in the dataset, but the S1 RSU and S1 FSU targets need the cell types. The dandiset's STTC files (370, under sub-STTC-*) hold STTC values between the units of two areas, but carry no unit identities or cell types, and their arrays have no dimension for the nine dt values their descriptions list, so they cannot stand in either.

  • BlockedResults, Auditory confound on CTC pathway modulation using tFUS, paragraph 2· Data missing from the dataset

    Low PRF (30 Hz) induces a direct auditory confound by enhancing AC to S1 correlation during sonication, while high PRF (3000 Hz) shows less auditory confound on time-locked modulation.

    These results indicate that low PRF would induce direct auditory confound by enhancing the correlation between auditory and S1 neuronal activities, while high PRF showed less auditory confound on time-locked neuronal modulation.

    Results, Auditory confound on CTC pathway modulation using tFUS, paragraph 2

    Depends on fig11c-ac-to-s1-30hz and fig11d-ac-to-s1-3000, both blocked because the dataset holds no waveforms or cell types for the S1 units.

Findings about the data

What attempts found about the dataset itself, for its owner and the archive.

  • Data missingThe spike waveforms are not in the dataset, and the RSU/FSU cell types are not either: no file has waveforms or mean waveforms, and acquisition, analysis and processing are empty in all 1550 unit files. The paper types every S1 and POm unit from its waveform (Methods, Neuronal type classification), so 30 of the 33 claims (all those on RSU or FSU responses or their correlations) cannot be tested, and the RSU/FSU split of the two count claims cannot be tested either.
  • Metadata errorThe units column celltype_label is described as label 1 = RSU, label 2 = FSU, but it is 0 for every unit in the 1516 unit files that have it; 34 unit files have no such column.
  • Other734 of the 1550 unit files are second (or third) copies of another file, with identical spike times and trial starts and the same identifier, under session start times up to 57 s apart and different file names; 816 distinct recordings remain. Counting the files as they are would double most unit counts.
  • Data missingNo recording is marked as sham (no tFUS) or as tFUS aimed at CG1, the control conditions of Figure 5; every recording gives a PRF and a PD and a target of S1, M1 or POm. If the sham and CG1 sessions are in the dataset, they cannot be found.
  • Data missingThe dataset has 10 rats with POm recordings and 7 with AC recordings, where the paper reports 16 S1-POm rats and 12 S1-AC rats (Methods); 20 rats in all.
  • Selection not documentedThe AC recordings at PD 200 us give 633 units with more than 2000 spikes (112, 89, 138, 167 and 127 at 30 to 4500 Hz), where Figure 11 reports 453 (83, about 59, 107, 104 and 100), although the dataset has fewer AC rats than the paper. The POm totals exceed the paper's RSU + FSU counts by 36 to 44% under every selection of recordings by identifier (at 3000 Hz 352 against 250, at 30 Hz 315 against 232, over the five PRFs 1470 against 1019), with 10 POm rats where the paper reports 16. They come within 10% only if two of the POm rats are left out (9 of the 45 pairs, for example BH511 and BH513, which nothing marks for exclusion) or if only units with more than 8000 spikes are counted. The paper's unit selection cannot be repeated from its description.
  • OtherThe dataset also holds recordings of M1 and sessions with tFUS aimed at M1 (152 distinct recordings) or at POm (65), and a few recordings labelled S1BF (3), VPM (1), S1_R, S1_1, M1_R (1 each) and AC_1, AC_2 (1 each), none of which the paper describes; the dandiset description mentions only S1, POm and AC with tFUS at S1.
  • Metadata errorIn the 60 sessions at PD 200 us with tFUS at S1 that have files from two of S1, POm and AC (the only sessions checked), both files have identical trial start times, so the areas were recorded together, but their session_start_time values differ (median 15.5 min, up to 3964 min); session_start_time does not give the start of the recording.
  • Description missingThe recorded area and the tFUS target are given only in free text in general/notes, in two spellings; the electrodes table has location unknown and the electrode group location is 'brain area'. Five recordings of BH500 leave the area empty, and no other recording does. general/data_collection, which carries the PRF and PD, is empty in 33 recordings (BH494 25, BH499 6, BH500 2), where the identifier gives them; where both are given (782 recordings) they agree.
  • File format errorThe arrays of the STTC files do not have the shape their descriptions give: the descriptions list a dimension for nine dt values, while the arrays have the 10 time windows, the trials and the units of the two areas (for example STTC is 10 x 499 x 20 x 9) and no dt dimension. Which dt was used, and which units the rows and columns are, is not stated.
  • Methods not availableQuestion for the paper's authors: was the normalized spiking rate the ratio of a unit's mean rate in the window to its mean rate in the 67 ms before stimulation (over trials), or a per-trial ratio averaged over trials; and in the post hoc Wilcoxon test against baseline, what was the baseline sample (the normalized pre-stimulus rates, which are 100% under the first reading, or another pre-stimulus window)? The Statistical analysis says only that rates were normalized to the 67 ms before stimulation. For Figure 11B the readings give different outcomes, and the claim is blocked.

Questions

Where the author or a verifier believed the claims file misread the paper, they asked, citing the paper, and a reader who had not seen the data answered from the paper alone, amending the claims file where it was wrong. The author then ran the claim as answered. Attempts before protocol pilot-11 also asked here for method details the claims file did not give.

  • openasked 2026-11-01about fig11b-ac-responses

    The header names N2 (the per-trial ratio of window to baseline count, averaged over the trials with a nonzero baseline) as a reading of the normalized spiking rate. Does the paper allow it? Averaging per-trial ratios over the trials with a nonzero baseline keeps the trials whose baseline is larger than average, so N2 sits well below 100% with no stimulus. If the paper does not allow it, should N2 be dropped from the header? The same reading is named for every spiking claim (Figs. 4 to 9 and 11B).

    Passage. The spiking rates were normalized to the pre-stim period (67 ms pre-stim) for comparison across different conditions using Kruskal-Wallis H test. (Materials and Methods, Statistical analysis)

    Why it matters. On the AC units with trial starts shifted by 1 s, where there is no stimulus, N2 gives 49.8% to 61.7% and N1 gives 94.3% to 99.0%; a verifier's simulation of Poisson spiking with no effect gives about 45% under N2 against 86% under N1. N2 gives a Kruskal-Wallis p of 0.0213 during sonication for Fig. 11B, where N1 gives 7.83e-06. Dropping N2 alone would not change the outcome of fig11b, since the two baseline readings under N1 also disagree, but it could change the outcome of a spiking claim in a later attempt on data with cell types.

    No reader has answered it yet.
  • openasked 2026-11-01about fig11b-ac-responses

    Under N1 the header's baseline reading B1 (the units' normalized rates in the 67 ms before stimulation) is exactly 100% for every unit, so the two-sample Wilcoxon test the paper names becomes a test of the window's values against a constant. Is B1 a reading the paper allows, or does the paper's two-sample test imply a baseline sample that varies across units (such as B2, or the units' raw pre-stimulus rates)?

    Passage. Post hoc one-tail two-sample Wilcoxon tests with Bonferroni's correction were employed. (Materials and Methods, Statistical analysis); The spiking rates were normalized to the pre-stim period (67 ms pre-stim) (same paragraph)

    Why it matters. For Fig. 11B, N1 with B1 gives 1500, 3000 and 4500 Hz above baseline during sonication (partially reproduced) and N1 with B2 adds 300 Hz (not reproduced); the claim is blocked because the readings disagree. A verifier's simulation with no effect found a false-positive rate of 2.3% at alpha 0.01 for B1 and 0% for B2.

    No reader has answered it yet.
  • openasked 2026-11-01about fig4-unit-counts

    The criteria of fig4-unit-counts and fig6-unit-counts call a claim reproduced when the counts are within tolerance under one reading of the k-means typing, while the header says that the outcome is the one every reading gives and that a claim whose readings differ is blocked. Which rule should the two count criteria follow?

    Passage. The units are first classified as either RSUs or FSUs based on the initial phase and afterhyperpolarization of the waveform by k-means clustering (Materials and Methods, Neuronal type classification)

    Why it matters. A verifier noted the two count criteria disagree with the header's rule. It changes nothing in this attempt, since the area totals are the same under every typing reading and decide both claims, but it would decide the outcome of an attempt on data with waveforms, where the readings could give different RSU/FSU splits. The same applies to fig6-unit-counts.

    No reader has answered it yet.

Notebooks

Every registered entry that uses this dataset. An entry can hold several notebooks, and can use more than one dataset.

Reproduction Ultrasound responses in the cortico-thalamo-cortical pathway (Gao et al. 2025)

Tests the findings of Gao, Ramachandran, Yu and He (2025) on dandiset 000946, which holds sorted spike times from S1, POm, auditory cortex and M1 of anesthetized rats during transcranial focused ultrasound at several pulse repetition frequencies and duty cycles. The files hold no spike waveforms and no cell types, so only the auditory cortex responses of Figure 11B, which the paper does not split by type, can be tested. All files

Verification

  • Round 1
    2026-11-01    Agreed with 30 of 33 outcomes. Independent agent run by CatalystNeuro (same harness, separate context, did not write the notebook). Reviewed: Round 1. The executed notebook of commit e64dd2bb (author-answers) with the conclusion cells and the summary removed, the helper module and the claims file, with the paper. The verifier also read the archive's metadata and full asset list and tested the helper functions on synthetic input. agrees compares its outcome with the outcome of the record it reviewed.
    • fig4-unit-counts: the verifier’s outcome was “Not reproduced”. The notebook calls this blocked, but its own totals in cell 17 decide it. POm units at 3000 Hz are 352 against 233+17 = 250, and at 30 Hz 315 against 232. Both are far outside 10%, so neither the reproduced outcome (eight counts in tolerance) nor the partial one (area totals within 10%) can be reached under any RSU/FSU split. This is subject to the session selection (item 6): identifiers with suffixes may be repeat or off-target placements the paper left out.
    • fig6-unit-counts: the verifier’s outcome was “Not reproduced”. Cell 17 again. S1 totals are 986, 1726 or 2414 depending on the S1 selection, against 1297 +/-10% (1167 to 1427). POm is 1470 against 1019 +/-10%. No split can rescue either. This is subject to the session selection (item 6).
    • fig11b-ac-responses: the verifier’s outcome was “Blocked”. Cell 24: the four method readings give different outcomes. N1/B1 is partial (during-sonication set {1500, 3000, 4500}, one PRF too many). N1/B2 is not reproduced (300 Hz also above baseline during sonication). N2/B1 and N2/B2 are not reproduced (Kruskal-Wallis in 0-67 ms p = 0.0213). The claims header says that when readings differ, the claim is blocked as methods_underspecified with every result shown. The notebook's judge gives a verdict per reading but never combines them under that rule. No reading reproduces the claim. The unit count is 633 against 453 +/-45.3, outside tolerance. N2 is not a reasonable reading of normalized to the pre-stim period (biased low, 50-62% at shifted onsets), but dropping it would not change fig11b, because N1/B1 and N1/B2 still disagree. Under N1 the B1 baseline sample is exactly 100 for every unit, so the rank-sum test is effectively a one-sample location test. Each PRF comes from only 3 or 4 of 6 rats, so the Kruskal-Wallis across PRFs confounds PRF with rat.
  • Round 2
    2026-11-01    Agreed with 33 of 33 outcomes. Independent agent run by CatalystNeuro (same harness, separate context, did not write the notebook). Reviewed: Round 2. The full executed notebook of commit 797a90f9 (author-revise-1), with the conclusion cells and the summary, the helper module, the claims file, the README and the attempt record, with the paper. agrees compares its outcome with the outcome of the record it reviewed.

Method differences

How the tests differed from the paper’s method or data, and the claims each applies to.

  • Recordings deposited twice (identical spike times and trial starts) are counted once.fig4-unit-counts fig6-unit-counts fig11b-ac-responses
  • The RSU/FSU split cannot be made (no waveforms or cell types), so the claim is judged on the area totals, which are the same under every typing reading of the claims file header.fig4-unit-counts fig6-unit-counts
  • Which sessions count beyond the sessions of the S1 and POm rats is not fixed; every selection the identifiers allow is shown, and the outcome is the same under all of them because the POm totals are outside 10% under each.fig4-unit-counts
  • Which sessions count beyond the sessions of the S1 and POm rats is not fixed; every selection the identifiers allow is shown, and none brings either area total within 10%.fig6-unit-counts
  • The normalization (N1 or N2) and the baseline sample of the post hoc test (B1 or B2) are not fixed by the paper (Statistical analysis, "normalized to the pre-stim period (67 ms pre-stim)"); all four readings were run and they give different outcomes.fig11b-ac-responses
  • The Bonferroni factor is the five PRF levels of the panel (claims file header, reading C1); reading C2 applies only to a panel with a single comparison.fig11b-ac-responses
  • The dataset holds AC recordings at PD 200 us from 6 rats (7 AC rats in all) where the paper reports 12 S1-AC rats, and it gives 633 units with more than 2000 spikes where the paper reports 453; the paper's selection of AC units could not be matched.fig11b-ac-responses
  • Each PRF comes from 3 or 4 of the 6 rats, and the rats differ by PRF, so the Kruskal-Wallis test across PRFs mixes PRF with rat (for example BH538 averages 654.1% at 3000 Hz, 201 to 268 ms, against 310.5% pooled); the paper also pools units across rats.fig11b-ac-responses
  • Trial start_time is taken as sonication onset; the clock check shows the firing changes after it and not 1 s later, which is consistent with this but does not prove there is no fixed delay.fig11b-ac-responses

Provenance

Attempt record
000946-gao-2025-01
Paper
Transcranial Focused Ultrasound Modulates Feedforward and Feedback Cortico-Thalamo-Cortical Pathways by Selectively Activating Excitatory Neurons
Date
2026-10-31
Made by
An agent, run by CatalystNeuro
Dataset
DANDI:000946
Version
0.250829.1249 (doi:10.48324/dandi.000946/0.250829.1249)
Run on
GitHub Actions
Harness
Claude Code
Model
claude-opus-5-5
Protocol
pilot-13

Transcripts of the agent runs

Each run’s prompt, messages and tool calls, with long output cut and the text of the paper left out, and its final report.

Other datasets that name the same paper

DANDI:000463 (not attempted) DANDI:000637 DANDI:001199
← DANDI:000945DANDI:000947 →