DANDI:000576

DELTA power was reduced in 12 of 14 amygdalae (negative MI in Table 2), irrespective of whether the amygdalae were within the SOZ or not

Results 2.1, paragraph 3; Table 2

Of the 6 readings that define all 14 delta MIs, 5 give 11 or 12 negative with negative MIs inside and outside the SOZ (baseline z-scoring with both group windows and the fixed window, 11, positive in amygdalae 2, 10 and 11; z-scoring over the amygdala with the all-14 group window and the fixed window, 12, positive in 2 and 10). Z-scoring over the trial with the fixed window, the first reading the criterion lists that defines them and given as measured, gives 10 (positive in 2, 9, 10 and 11), so the count is not in range under every reading. Table 2 has amygdalae 2 and 4 positive. The per-amygdala window defines 1 or 2 delta MIs. The variants give 10 to 13.

How the test differed from the paper’s method

• MI readings as in table2-mi-values; z-scorings as in fig1b-delta-decrease.
• Assumption. The only reading that fails uses window (c), fixed windows of 0 to 2 s for HGP and 0 to 24 s for delta, taken from the Results' text (Results 2.1, Methods 4.11) and not from the claims file. The claims file's window readings that define all 14 delta MIs (group clusters under baseline z-scoring in both sets and under z-scoring over the amygdala in all 14) give 11, 11 and 12, all in range. So the outcome rests on counting window (c) as a reading. The question is for the authors (methods_unavailable, MI window).

We identified 50 putative clusters of neurons (Niediek et al., 2016) from 9 amygdalae and included those with firing rate >1 Hz (N 1⁄4 41 neurons) in further analysis

Results 2.3, paragraph 1; Methods 4.9

The dandiset holds 42 units, 42 above 1 Hz over the epochs each was recorded in (reading a, given as measured) and 40 over the whole 442 s (reading b); both within 4.1 of 41. The 50 putative clusters cannot be tested, as only units above about 1 Hz were deposited. Units are in 8 amygdalae, not 9, so the third expected value fails. The measured value is the first expected value only. The units come from 5 patients (1, 2, 3, 4, 7), which matches the abstract's 5 participants, but not the amygdalae of Table 1, which lists MUA for patients 5 and 8 and none for 7. Their mean rate over the observed epochs is 4.65 +- 3.91 Hz (median 3.13), against the data descriptor's 1.66 +- 2.65 Hz; 31 units are above 2 Hz.

How the test differed from the paper’s method

• Epochs in which a unit has no spike at all are taken as not recorded (16 units of subject 7 after epoch 11, 4 units of subject 2 in epoch 1), as their rates elsewhere are 1 to 16 Hz.

Across all neurons, the covariation of neuronal firing with HGP was not significantly different between the two experimental conditions (Spearman’s r 1⁄4 0.17 ... �0.28 in AVERSIVE, 0.18 �0.18 in NEUTRAL; Wilcoxon paired test, p > 0.05)

Results 2.3, paragraph 3; Figure 3B, lower boxplot

What the paper reports

• Wilcoxon paired test, p > 0.05 (n = 41 neurons)

The readings of the claims file disagree; the claimed absence holds in 4 of 6. Per block and averaged no time base finds a difference (p 0.72 to 0.83), nor does the 1 s time base over the blocks laid end to end (r = 0.085 +- 0.130 in AVERSIVE and 0.094 +- 0.124 in NEUTRAL, p = 0.54, given as measured). Over the blocks laid end to end on the 1.5 s and 15 s time bases, each block smoothed on its own, the test finds NEUTRAL higher (p = 0.012 and 0.023, median differences -0.087 and -0.131). Smoothing over the whole session first (the first revision's form, now a variant) gave p = 0.002 and 0.009. The means are within the stated spreads under every reading. Partially reproduced and not blocked, because the six readings are named by the claims file, so no method detail is left to choose; the test ran under each, and the claimed absence holds under some and fails under others, that is, it holds only in part. Unlike the criteria of fig1b-delta-decrease and fig1f-hgp-more-specific-than-delta, this criterion does not say to block when its readings disagree. The time base is still asked of the authors (methods_unavailable).

How the test differed from the paper’s method

• In the readings with the blocks laid end to end, each block is smoothed (or binned) on its own before concatenation, so no sample carries another block or a baseline clip.
• HGP is the Methods 4.8 envelope without per-trial normalisation; a per-block Spearman r is unchanged by any per-trial normalisation.

Across all neurons and participants, the ongoing firing rate positively correlated with μHGP (Spearman’s r 1⁄4 0.27 �0.15), HGP (Spearman’s r 1⁄4 0.19 � 0.17) and BOLD (Spearman’s r 1⁄4 0.23 �0.20) ... Across all recording modalities, the correlation of the firing rate and other measures was lowest with HGP (Wilcoxon’s paired test, p < 1e-2)

Results 2.3, paragraph 5; Figure 3D and its inset

What the paper reports

• correlation with HGP against those with muHGP and with BOLD: Wilcoxon paired test, p < 0.01 (n = neurons, number not given (BOLD exists only for the amygdalae with fMRI))

The firing-HGP Spearman r is 0.136 +- 0.241 with 15 s smoothing (reading a, given as measured) and 0.139 +- 0.195 on the 1.5 s time base, positive in 29 and 32 of 42 units, within 0.19 +- 0.17. The muHGP and BOLD correlations and the two tests that HGP is lowest cannot be computed, as the microwire signals and the BOLD time courses are not available. The measured value is the HGP mean, the second expected value; the first and third cannot be measured.

How the test differed from the paper’s method

• Only the HGP part is tested.

The HGP was enhanced in response to aversive stimuli during the first 2 s of the AVERSIVE sequence presentation (Fig. 1C, unpaired t-tests, corrected for multiple comparisons using random permutations, p < 0.05)

Results 2.1, paragraph 1; Figure 1C

What the paper reports

• unpaired t-tests, corrected for multiple comparisons using random permutations (Results); linear mixed model with cluster-based permutation (Methods 4.11), p < 0.05

None of the 12 mixed-model readings and none of the 12 unpaired t-test readings of the claims file finds a significant HGP cluster, nor do the two z-scoring variants. The largest mixed-model t in the first 2 s is 3.01 at 0.95 s after onset (baseline z-scoring, nSOZ amygdalae), above the threshold at 13 time points, and its cluster has permutation p 0.062; the smallest cluster p of all mixed-model readings is 0.056. With the HGP band limited to 60-150 or 60-250 Hz (variants of the open upper edge) the only significant HGP clusters are AVERSIVE above NEUTRAL from about 6 to 8 s (p 0.020 and 0.042) and, with baselinepool, 3.1 to 4.7 s, not in the first 2 s. Among the baselines of Methods 4.8 (variants), the previous epoch as baseline with z-scoring over the amygdala gives an HGP cluster from 0.6 to 5.5 s in the nSOZ amygdalae (p 0.002); as blocks alternate, that baseline is always a block of the other condition. With the trialsd variant two more HGP clusters begin in the first 2 s, both in the nSOZ amygdalae, 0.6 to 1.9 s with the mean of all epochs as baseline (p 0.028) and 0.5 to 8.4 s with the previous epoch (p 0.002). These three are the only HGP clusters that begin in the first 2 s, and all use a baseline other than the 1 s clip. Per-amygdala time courses are in the notebook.

How the test differed from the paper’s method

• As fig1b-delta-decrease. HGP is a 60 Hz high-pass (the claims file's reading); two band-pass variants are shown and do not decide the outcome.

Since no effect was observed in the other spectral bands (see Methods 7.8), in the following we focus on DELTA power and HGP

Results 2.1, paragraph 1; bands in Methods 4.8

The readings of the claims file disagree. Under the z-score over the whole trial no band has a cluster, nor has delta. Under baseline z-scoring the mixed model finds clusters with AVERSIVE below NEUTRAL in alpha in the 7 nSOZ amygdalae (13.3 to 15.9 s, p 0.016) and in theta and alpha in all 14 (p 0.002 to 0.028); under z-scoring over the amygdala all four bands have clusters in all 14 and none in the nSOZ amygdalae. Every reading that shows the delta effect of Fig. 1B (6 of 12) shows a cluster in another band. Of the variants, baselinepool in the nSOZ amygdalae shows the delta effect and no other-band cluster; in all 14 it has theta and alpha clusters.

How the test differed from the paper’s method

• Assumed that the paper's pipeline is one under which its own delta result (Fig. 1B) appears, as the paper reports the delta effect and the absence in other bands from the same analysis (Results 2.1, paragraph 1). The outcome is judged on the readings of the claims file that show the delta effect; the trial reading shows no cluster in any band, delta included. The first revision adopted baseline z-scoring alone; the verifier found that the author's own choice, and it is no longer used.
• Outcomes are judged on the three z-scorings of the claims file (trial, baseline, amygdala). Two variants are shown and enter no count of readings, trialsd (the baseline-corrected envelope divided by its SD in the trial) and baselinepool (the epoch's baseline mean and the SD of the baseline clips pooled over the amygdala's epochs, a stable form of baseline z-scoring). The two other baselines of Methods 4.8 (mean of all epochs, previous epoch) are also run as variants.
• As fig1b-delta-decrease for the test.

Not quoted: Values read from Table 2, which no sentence states one by one. Minus signs are garbled in the extracted text; the signs used here agree with the counts the Results give.

No reading gives all 28 values within tolerance (at best 4 of 28 among the claims file's z-scorings, z-scoring over the amygdala with fixed windows; 6 with the trialsd variant) or all 28 signs (at best 23 of 28). The measured values are those of the first reading the criterion lists that defines all 28, z-scoring over the trial with fixed windows (1 of 28 within tolerance, 16 signs). The group cluster window defines no HGP MI, as there is no HGP group cluster; the per-amygdala clusters define 1 to 3 MIs per band. Re-mapping subjects 2 and 7, or 7 to the paper's 5 and 8, does not bring the values closer under the claims file's z-scorings (0 to 4 within tolerance); the trialsd variant goes from 6 to 5 and 7.

How the test differed from the paper’s method

• Window readings (a) group clusters and (b) per-amygdala clusters are the claims file's; (c) fixed windows, 0 to 2 s for HGP and 0 to 24 s for delta, is taken from the Results' text ("occurred at stimulus onset in HGP and was sustained throughout the trial in DELTA", Results 2.1 paragraph 2; "HGP enhancement in the first 2 s or sustained delta decrease", Methods 4.11) and counted as a reading.
• Where a window reading finds no significant window for an amygdala, its MI is undefined (the first revision set it to 0); a claim is judged on the readings that define every MI it needs.
• Outcomes are judged on the three z-scorings of the claims file (trial, baseline, amygdala). Two variants are shown and enter no count of readings, trialsd (the baseline-corrected envelope divided by its SD in the trial) and baselinepool (the epoch's baseline mean and the SD of the baseline clips pooled over the amygdala's epochs, a stable form of baseline z-scoring). The two other baselines of Methods 4.8 (mean of all epochs, previous epoch) are also run as variants.

MI and HGP-MI were negatively correlated across all amygdalae

Results 2.1, paragraph 2; Figure 1D

What the paper reports

• Spearman correlation, r = -0.67, p = 0.01 (n = 14 amygdalae)

Only window (c) defines both MIs. Under it r = -0.30 (p = 0.30), -0.12 (p = 0.69) and -0.12 (p = 0.68) for z-scoring over the trial, against the baseline and over the amygdala; the variants give -0.33 and -0.04. None is within tolerance or has p < 0.05. The printed Table 2 values give r = -0.74 (p = 0.003).

How the test differed from the paper’s method

• MI readings as in table2-mi-values; z-scorings as in fig1b-delta-decrease.
• Assumption. Window (c), fixed windows of 0 to 2 s for HGP and 0 to 24 s for delta, is the only window reading that defines the MIs this claim needs, so the outcome rests on it. It is taken from the Results' text (Results 2.1, Methods 4.11, "HGP enhancement in the first 2 s"), while the paper defines the MI over "the interval where the two curves are significantly separated" (Table 2 legend), which these data give for no HGP MI. The question is for the authors (methods_unavailable, MI window).

Specifically for amygdalae not included in the SOZ, we observed that the stronger the HGP enhancement, the stronger the reduction in DELTA power (Table 2)

Results 2.1, paragraph 2; Table 2

Over the 7 nSOZ amygdalae, window (c) gives r = -0.57 (trial), -0.14 (baseline) and +0.07 (amygdala), negative in 2 of the 3 readings that define the MIs, so not under every reading. Windows (a) and (b) define no HGP MI for these amygdalae. The variants give -0.07 and +0.21. The printed Table 2 values give r = -0.68 (p = 0.094).

How the test differed from the paper’s method

• MI readings as in table2-mi-values; z-scorings as in fig1b-delta-decrease.
• Assumption. Window (c), fixed windows of 0 to 2 s for HGP and 0 to 24 s for delta, is the only window reading that defines the MIs this claim needs, so the outcome rests on it. It is taken from the Results' text (Results 2.1, Methods 4.11, "HGP enhancement in the first 2 s"), while the paper defines the MI over "the interval where the two curves are significantly separated" (Table 2 legend), which these data give for no HGP MI. The question is for the authors (methods_unavailable, MI window).

All amygdalae outside SOZ showed a positive HGP-MI in response to AVERSIVE videos, while this response was absent in amygdalae within the SOZ (p 1⁄4 0.009)

Results 2.1, paragraph 3; Table 2 and its legend

What the paper reports

• HGP MI, nSOZ against SOZ amygdalae: not named, p = 0.009 (n = 7 nSOZ and 7 SOZ amygdalae)

Window (c) gives 4, 4 and 5 of 7 nSOZ amygdalae with a positive HGP MI and 3, 5 and 7 of 7 SOZ amygdalae with a negative one (trial, baseline, amygdala); no reading gives both counts. The measured values are those of the trial reading. Under z-scoring over the amygdala Fisher's test gives p = 0.021 but the nSOZ count is 5. Windows (a) and (b) define no HGP MI for all amygdalae.

How the test differed from the paper’s method

• MI readings as in table2-mi-values; z-scorings as in fig1b-delta-decrease.
• Assumption. Window (c), fixed windows of 0 to 2 s for HGP and 0 to 24 s for delta, is the only window reading that defines the MIs this claim needs, so the outcome rests on it. It is taken from the Results' text (Results 2.1, Methods 4.11, "HGP enhancement in the first 2 s"), while the paper defines the MI over "the interval where the two curves are significantly separated" (Table 2 legend), which these data give for no HGP MI. The question is for the authors (methods_unavailable, MI window).

HGP but not DELTA distinguishes between amygdalae within SOZ and outside SOZ (p 1⁄4 0.019, not corrected)

Figure 1E legend

What the paper reports

• HGP MI, SOZ against nSOZ: not named, p = 0.019, not corrected-corrected (n = 7 nSOZ and 7 SOZ amygdalae)

Under no reading that defines both MIs (window c) and neither test does the HGP MI differ between nSOZ and SOZ amygdalae at p < 0.05 (smallest p 0.084, baseline z-scoring, t-test); the DELTA MI does not differ either (p 0.08 to 0.95). With the baselinepool variant the rank-sum test gives HGP p = 0.035 and DELTA p = 0.48. Run on the printed Table 2 values the same tests give HGP p = 0.004 and 0.002 but also DELTA p = 0.035 and 0.013; asked in questions/000576.yaml.

How the test differed from the paper’s method

• MI readings as in table2-mi-values; z-scorings as in fig1b-delta-decrease.
• Assumption. Window (c), fixed windows of 0 to 2 s for HGP and 0 to 24 s for delta, is the only window reading that defines the MIs this claim needs, so the outcome rests on it. It is taken from the Results' text (Results 2.1, Methods 4.11, "HGP enhancement in the first 2 s"), while the paper defines the MI over "the interval where the two curves are significantly separated" (Table 2 legend), which these data give for no HGP MI. The question is for the authors (methods_unavailable, MI window).

With sensitivity (7/7 1⁄4 100% CI [66–100]%), high specificity (6/7 1⁄4 86% CI [42 99]%) and accuracy 93% (13/14, CI [66–99]%), HGP enhancement indicated amygdalar functional integrity

Results 2.1, paragraph 3; Methods 4.11, paragraph 4

What the paper reports

• sensitivity, 7 of 7: 95% CI 66 to 100 %
• specificity, 6 of 7: 95% CI 42 to 99 %
• accuracy, 13 of 14: 95% CI 66 to 99 %

Window (c) gives sensitivity 4/7, 4/7 and 5/7, specificity 3/7, 5/7 and 7/7 and accuracy 7/14, 9/14 and 12/14 (trial, baseline, amygdala; the trial reading given as measured); no reading gives 7/7, 6/7 and 13/14. The Clopper-Pearson 95% intervals of the paper's counts are 59-100, 42-100 and 66-100%, against the printed 66-100, 42-99 and 66-99%; no single interval method run (one-sided exact, Wilson, Jeffreys) gives all three printed intervals. Asked in questions/000576.yaml.

How the test differed from the paper’s method

• MI readings as in table2-mi-values; z-scorings as in fig1b-delta-decrease.
• Assumption. Window (c), fixed windows of 0 to 2 s for HGP and 0 to 24 s for delta, is the only window reading that defines the MIs this claim needs, so the outcome rests on it. It is taken from the Results' text (Results 2.1, Methods 4.11, "HGP enhancement in the first 2 s"), while the paper defines the MI over "the interval where the two curves are significantly separated" (Table 2 legend), which these data give for no HGP MI. The question is for the authors (methods_unavailable, MI window).

All 41 clusters of neurons increased their firing rate during the AVERSIVE condition (Firing Rate Variation, FRV 1⁄4 0.26 �0.15)

Results 2.3, paragraph 2; Figure 3A

Over the 24 s blocks and observed epochs 27 of 42 units have FRV > 0, mean 0.033 +- 0.690 Hz (Wilcoxon against 0, p = 0.50); over the 26 s trials, or with all epochs, 23 to 26 positive and means 0.007 to 0.032 Hz. By subject the mean FRV runs from -0.21 to 0.62 Hz. Relative to each unit's mean rate it is 0.061 +- 0.231 (a variant).

How the test differed from the paper’s method

• Observed epochs as in neuron-count; the variant with all epochs is shown.

Contrary to our expectation of an immediate spiking onset, the neurons increased their firing gradually and sustained enhanced firing throughout the AVERSIVE video sequences

Results 2.3, paragraph 2; Figure 3A

The population AVERSIVE minus NEUTRAL rate (100 ms bins, 1 s smoothing) is 0.310 Hz over 0 to 2 s and 0.018 Hz over 2 to 24 s, with its first peak at 0.15 s and 4 of the twelve 2 s bins negative. Neither reading of the criterion holds. It is -0.078 Hz over the baseline clip.

We found a strong positive correlation across all participants between the SSI and FRV (Spearman’s r 1⁄4 0.78, p < 1e-7)

Results 2.3, paragraph 4; Figure 3C; Figure 4A

What the paper reports

• Spearman correlation, r = 0.78, p < 10⁻⁷ (n = neurons with an SSI)

Over the 41 units with another unit in their amygdala, Spearman r between SSI and FRV is 0.26 (p = 0.11) on the 1 s time base and 0.40 (p = 0.009) on the 15 s one, against 0.78 with p < 1e-7. Neither reading gives the value, and the 1 s reading misses p < 0.05, so the partial result of the criterion does not hold either. Dropping the 4 pairs likely on one wire gives 0.24 (p = 0.13) and 0.35 (p = 0.023).

How the test differed from the paper’s method

• SSI uses every other unit of the same amygdala. Methods 4.11 say "on different microwires"; the units table has no microwire column (columns offset, spike_times, obs_intervals, electrodes, electrode_group, waveform_mean, waveform_sd), each unit points only to the bipolar macro channel, and the file holds one electrode group. So the Results' definition, "all neighboring neurons", is assumed. A variant drops pairs likely on one wire by a spike train test (no spikes within 1 ms of each other where the 2 to 12 ms flanks predict at least 3); it flags 4 of 116 testable pairs, and the others fire within 1 ms as often as the flanks predict (median ratio 1.19).

We found that the co-modulation between neuronal firing and HGP is described by both FVR and SSI (Fig. 4B, r 1⁄4 0.59, p < 1e-4)

Results 2.4, paragraph 1; Figure 4B

What the paper reports

• multiple linear regression, r = 0.59, p < 10⁻⁴

With 15 s time bases for both the firing-HGP r and SSI, R = 0.635 (within 0.059 of 0.59) and model p = 5.6e-5; over the four readings R = 0.56 to 0.68 and model p from 6e-6 to 7e-4. SSI has p <= 2e-4 in every reading; FRV has p 0.07 to 0.65, with a negative coefficient in three, so the regression is described by SSI alone and the claim's "both FRV and SSI", which the criterion requires, fails in every reading. The SSI variant without likely same-wire pairs gives the same.

How the test differed from the paper’s method

• SSI as assumed in fig3c-ssi-frv-correlation.

In nSOZ amygdalae, DELTA band power was consistently reduced throughout the duration of the AVERSIVE sequence (Fig. 1B)

Results 2.1, paragraph 1; Figure 1B

What the paper reports

• linear mixed model per time point, cluster-based permutation correction (Fig. 1 legend calls it cluster analysis), p < 0.05

The criterion blocks the claim when the readings disagree, and they do; it holds in 6 of the 12 mixed-model readings of the claims file. With the z-score over the whole 26 s trial no reading finds a delta cluster (smallest cluster p 0.88 in all 14 amygdalae, 1.0 in the 7 nSOZ). With z-scoring against the 2 s baseline, AVERSIVE is below NEUTRAL with significant clusters covering 14.3 s of the 24 s block in the nSOZ amygdalae and 15.6 s in all 14. With z-scoring over all trials of the amygdala, 17.1 s in all 14 and none in the nSOZ amygdalae (smallest p 0.072). The block mean of AVERSIVE minus NEUTRAL is negative under every reading (-0.03 to -1.39). The variants agree with the baseline reading (baselinepool 13.3 and 16.8 s, trialsd 7.0 and 16.1 s). The unpaired t-test variant gives the same pattern. The question for the authors is in the findings (methods_unavailable, z-scoring).

How the test differed from the paper’s method

• The paper's mixed model Power ~ condition + (1|Patients) is computed in its within-patient form (patient fixed effects), which gives the same condition estimate for this balanced design (8 AVERSIVE and 9 NEUTRAL trials per amygdala); the notebook compares it with statsmodels' MixedLM at four time points (t -2.125 against -2.123 and so on).
• Labels are permuted within each patient, the same permutation for both amygdalae of a patient (Methods 4.11, "permuted the labels of neutral and aversive trials across all participants", 500 permutations); the unpaired t-test variant permutes over all trials.
• The test is run over the 24 s block (0 to 24 s from onset, Fig. 1 "locked to stimulus onset"); the baseline clip is left out. A choice, stated in the notebook.
• Outcomes are judged on the three z-scorings of the claims file (trial, baseline, amygdala). Two variants are shown and enter no count of readings, trialsd (the baseline-corrected envelope divided by its SD in the trial) and baselinepool (the epoch's baseline mean and the SD of the baseline clips pooled over the amygdala's epochs, a stable form of baseline z-scoring). The two other baselines of Methods 4.8 (mean of all epochs, previous epoch) are also run as variants.
• Envelopes are filtered per 26 s epoch with 2 s mirror padding, as the epochs are not contiguous in the files, and kept at 10 Hz after the 1 s smoothing (standard choices).

The HGP response exceeds the DELTA response in specificity when indicating the amygdala’s dysfunction within SOZ (p~0.2 HGP 5/5 DELTA 1/5, chi2-test)

Figure 1F legend

What the paper reports

• chi-square test, p = 0.2 (n = HGP 5/5, DELTA 1/5 (as printed))

The readings disagree on which response is more specific, which the criterion blocks. Window (c), the only one that defines the MIs of all 7 SOZ amygdalae, gives HGP more specific under baseline z-scoring (5/7 against 2/7, chi-square p = 0.11) and z-scoring over the amygdala (7/7 against 1/7, p = 0.001), and a tie under z-scoring over the trial (3/7 each). The variants give HGP (6/7 against 1/7 and 3/7). Reading (b), the printed counts 5/5 against 1/5, gives HGP more specific with p = 0.0098 (0.053 with Yates' correction), not about 0.2; no reading gives a p near 0.2.

How the test differed from the paper’s method

• MI readings as in table2-mi-values; z-scorings as in fig1b-delta-decrease.
• Assumption. Window (c), fixed windows of 0 to 2 s for HGP and 0 to 24 s for delta, is the only window reading that defines the MIs this claim needs, so the outcome rests on it. It is taken from the Results' text (Results 2.1, Methods 4.11, "HGP enhancement in the first 2 s"), while the paper defines the MI over "the interval where the two curves are significantly separated" (Table 2 legend), which these data give for no HGP MI. The question is for the authors (methods_unavailable, MI window).