Claims 30
From the paper Retrorubral field is a hub for diverse threat and aversive outcome signals · claims file · checked 2026-11-06: ready
- units-and-sessions
The paper recorded 743 single units from 11 rats over 228 fear discrimination sessions.
We recorded 743 single units from 11 rats over 228 fear discrimination sessions.
Results, Diverse cue firing patterns, paragraph 1How it will be judged
Counts of units, distinct sessions and rats. Reproduced when all three are within tolerance; the rat count must match exactly.
- behavior-discrimination
In the sessions that yielded cue-responsive units, rats fully discriminated the cues by suppression ratio (danger > uncertainty > safety), with a main effect of cue and 95% bootstrap CIs of every differential suppression ratio excluding zero.
Analysis of variance (ANOVA) for suppression ratio revealed a main effect of cue ... 95% bootstrap CIs for differential suppression ratios did not contain zero for any cue comparison
Results, opening section, paragraph 2; Figure 1BHow it will be judged
Suppression ratio per trial as (baseline poke rate - cue poke rate) / (baseline + cue), baseline the 20 s before cue onset, cue the 10 s cue (STAR Methods, Calculating suppression ratios). Sessions are those with at least one cue-responsive unit (Figure 1B legend; depends on cue-responsive-units), each counted once. Per rat, the mean over those sessions of each cue's mean ratio (uncertainty over its 8 trials); a repeated-measures ANOVA on cue over the 11 rats (the printed df, 2 and 20, fit this). Differences per rat with studentized bootstrap CIs. Tolerances are the distance from the paper's mean to the farther bound of its CI. Open is whether rat values are means over sessions (main reading) or over all pooled trials of those sessions (variant). Reproduced when the cue effect has p < 0.05, all three CIs exclude zero and the three means are within tolerance; partially reproduced when the ANOVA and the order hold but a CI or a mean does not.
- cue-responsive-units
423 units, from 190 sessions, were cue responsive, with no fewer than 11 per rat.
Four hundred and twenty-three cue-responsive single units were obtained from 190 sessions, and no fewer than 11 single units were identified per rat
Results, Diverse cue firing patterns, paragraph 1; Figure S1How it will be judged
Screen from STAR Methods, Identifying cue-responsive neurons. For each unit and each cue (danger 4 trials, uncertainty 8 trials pooled, safety 4 trials), the raw firing rate in the 10 s before cue onset is compared per trial with the mean rate in the first 1 s and in the last 5 s of the cue by a paired, two-tailed t test; a unit is cue responsive when any of the six tests has p < 0.05, uncorrected, as the paper states. Reproduced when the three counts are within tolerance. Figure S1 gives per-rat counts (RB01 35, RB03 53, RB04 24, RB06 59, RB07 29, RB09 33, RB10 18, RB11 11, RB12 47, RB13 69, RB14 45), which are reported alongside.
- kmeans-elbow
In k-means clustering of the 423 cue-responsive units, the mean squared Euclidean distance to the centroid falls as the cluster number rises from 1 to 15, and the decrease diminishes after eight clusters.
Euclidean distance from the centroid decreased as cluster number increased, with the amount of decrease diminishing after a cluster number of eight
Results, Diverse cue firing patterns, paragraph 2; Figure 1EHow it will be judged
Matrix of 423 units x 54 epochs of normalized firing (Results; the Methods say "Firing rate", so raw rate is a variant), 1-s means of the 250-ms bins; epochs 1-30 the ten 1-s cue intervals for danger, uncertainty and safety, epochs 31-54 six 1-s intervals after shock or omission for danger, uncertainty shock, uncertainty omission and safety (STAR Methods, K-means clustering). MATLAB kmeans defaults (squared Euclidean, k-means++). Open are (e-a) whether the six outcome epochs start at shock onset (12 s after cue onset) or at shock offset (12.5 s, the start of the outcome period in Figure 6), and (e-b) whether "Clustering was performed 10 times" means ten replicates per k keeping the best, or ten runs averaged; all four combinations are run. The paper judges the elbow by eye; as a standard choice the claim is reproduced when the curve decreases from k = 1 to 15 and every drop from k to k + 1 for k >= 8 is smaller than the drop from 7 to 8, under the z-a and z-b normalizations alike or one of them.
- kmeans-clusters
Eight k-means clusters were kept, of sizes k1 33, k2 59, k3 18, k4 63, k5 58, k6 61, k7 64 and k8 67.
All subsequent analyses consider eight clusters: k1, n = 33; k2, n = 59; k3, n = 18; k4, n = 63; k5, n = 58; k6, n = 61; k7, n = 64; and k8, n = 67.
Results, Diverse cue firing patterns, paragraph 2; Figure 1F; Figure S2How it will be judged
k-means with k = 8 on the matrix of kmeans-elbow, under its readings. Each cluster found is matched to one of the paper's k1 to k8 by its mean danger, uncertainty and safety firing in the intervals of Figure 3A and 3C against the patterns of Results, Diverse cue firing patterns, paragraph 3 (k1 onset danger peak, k3 sustained D > U > S, k5 ramping D > U > S, k7 delay D > U > S, k2 uncertainty inhibition, k4 equal danger and uncertainty inhibition, k6 danger inhibition with safety excitation, k8 danger inhibition). Reproduced when all eight patterns are found and every size is within tolerance; partially reproduced when all eight patterns are found and at least five sizes are within tolerance. Where the data set holds the paper's labels, their counts and the agreement (adjusted Rand index) between them and the clusters found here are reported.
- cluster-firing-characteristics
Clusters differed in baseline firing rate, coefficient of variance and coefficient of skewness, but not in anterior-posterior location, waveform half-duration or waveform amplitude ratio.
found a main effect of cluster for baseline firing rate (F7,415 = 2.48, p = 0.017, hp2 = 0.04, op = 0.88), coefficient of variance (F7,415 = 2.28, p = 0.027, hp2 = 0.04, op = 0.84), and coefficient of skewness (F7,415 = 2.65, p = 0.011
Results, Diverse cue firing patterns, paragraph 3; Figure 1F; Figure S3How it will be judged
Measures from STAR Methods, Firing and waveform characteristics. Baseline rate is the mean rate in the 10 s before cue onset; CV is SD/mean and skewness 3 x (mean - median)/SD of the inter-spike intervals in the 20 s before cue onset over the 16 trials; half duration and amplitude ratio from the mean waveform as defined there; anterior-posterior location the recording bregma level. A MANOVA with cluster as between-subjects factor and the six measures, with the univariate tests read for each. Open is whether the inter-spike intervals are pooled over the 16 trials (main reading) or the measure is averaged over trials (variant). Reproduced when baseline rate, CV and skewness have p < 0.05 and location, half duration and amplitude ratio do not; partially reproduced when at least two of the three effects hold and no other measure is significant. Effect sizes are given either way.
- k1-k3-low-baseline
Units of clusters k1 and k3 had lower baseline firing rates than all other clusters.
The single units comprising clusters k1 and k3 had lower baseline firing rates compared to all other clusters (t421 = 3.28, p = 0.001).
Results, Diverse cue firing patterns, paragraph 3; Figure 1F, bottomHow it will be judged
Baseline rate as in cluster-firing-characteristics; an independent samples t test of k1 and k3 units pooled against all other units. Reproduced when p < 0.05 with k1 and k3 lower.
- table1-cue-interaction
Every cluster showed a significant trial type x interval interaction for cue firing.
found a significant trial type x interval interaction for each cluster (Table 1)
Results, Diverse cue firing patterns, paragraph 4; Table 1 (Cue)How it will be judged
Per cluster, normalized firing in 56 250-ms bins (2 s before cue onset, the 10-s cue, the 2-s delay) for danger, uncertainty and safety; a repeated-measures ANOVA with trial type and interval within units (see the header). Reproduced when the interaction has p < 0.05 in all eight clusters; partially reproduced in six or seven.
- fig3b-excitatory-clusters
Clusters k1, k3, k5 and k7 show excitatory firing to danger and uncertainty; k1 fires most to danger at cue onset with lower, similar firing to uncertainty and safety; k3 (whole cue and delay), k5 (late cue and delay) and k7 (delay) fully differentiate danger > uncertainty > safety.
K1 neurons were strongly cue onset responsive, firing maximally to danger but firing at lower and similar levels to uncertainty and safety. K3 neurons sustained differential firing throughout cue presentation and the 2-s delay period ... danger > uncertainty > safety. K5 neurons ramped danger and uncertainty firing, fully differentiating the three cues by the last half of cue presentation ... K7 neurons were minimally cue responsive, but showed complete differential cue firing during the 2-s delay period
Results, Diverse cue firing patterns, paragraph 4; Figure 3A and 3B; Table S1How it will be judged
Per unit, mean normalized firing to each cue in the interval of Figure 3A (k1 first 1 s of cue; k3 the 10-s cue plus 2-s delay; k5 last 5 s of cue plus delay; k7 the 2-s delay); per-unit differences danger - uncertainty and uncertainty - safety with studentized bootstrap CIs (Table S1). Tolerances are the distance from the paper's mean to the farther CI bound. Reproduced when every CI agrees with the paper's on whether it excludes zero (and on sign where it does) and every mean is within tolerance; partially reproduced when the CIs agree for at least three of the four clusters.
- fig3d-inhibitory-clusters
Clusters k2, k4, k6 and k8 show sustained inhibition through the cue and delay; k2 is inhibited to uncertainty only, k4 equally to danger and uncertainty, k6 most to danger, less to uncertainty, with excitation to safety, and k8 to danger only.
K2 neurons specifically inhibited firing to uncertainty. K4 neurons specifically and equivalently inhibited firing to danger and uncertainty. K6 neurons were maximally inhibited to danger, less strongly inhibited to uncertainty, but showed excitatory firing to safety. K8 neurons specifically inhibited firing to danger.
Results, Diverse cue firing patterns, paragraph 4; Figure 3C and 3D; Table S1How it will be judged
As fig3b-excitatory-clusters, over the 10-s cue plus 2-s delay for all four clusters (Figure 3D legend).
- figs4-onset-and-ramp
From cue onset to delay, k1 firing decreases for all three cues (phasic onset firing), k5 firing increases for danger and uncertainty (ramping), and k2 firing increases for danger.
Danger (M = -1.32 ... uncertainty, (M = -0.95, 95% CI [-1.25, -0.63]), and safety (M = -0.52, 95% CI [-0.73, -0.20]) firing decreases from onset to delay were observed for the phasically responsive k1 neurons. Danger (M = 0.50, 95% CI [0.32, 0.66]) and uncertainty (M = 0.23, 95% CI [0.11, 0.35]) firing increases from onset to delay were observed for the ramping k5 neurons. A danger (M = 0.18, 95% CI [0.06, 0.31]) firing increase
Figure S4 legend; Results, Diverse cue firing patterns, paragraph 4How it will be judged
Per unit and cue, mean normalized firing in the 2-s delay minus that in the first 1 s of the cue; studentized bootstrap CI per cluster and cue. The k1 danger CI prints as "[1-.62, -0.97]", read as [-1.62, -0.97]. Tolerances are the distance from the mean to the farther bound. Reproduced when the six CIs exclude zero with the paper's sign and the means are within tolerance; partially reproduced when the k1 and k5 CIs hold but a mean or the k2 CI does not. The CIs of the other cues and clusters, which the paper shows as including zero, are reported.
- fig4abc-threat-bias
Except for k2, each cluster's uncertainty firing has the direction and time course of its danger firing at lower magnitude, and safety firing tends to oppose danger firing; k4 and k6 are inhibited to danger but excited to safety, k3 and k5 the reverse.
With the exception of cluster k2, the direction and temporal characteristics of each cluster's uncertainty firing were similar to that for danger, only diminished in magnitude. Critically, across all eight clusters, safety firing direction tended to oppose danger firing direction. Clusters k4 and k6 showed considerable firing inhibition to danger, but firing increases to safety. Just the opposite was observed for clusters k3 and k5
Results, Uniform firing bias toward threat cues, paragraph 1; Figure 4A to 4C; Figure S5How it will be judged
The paper shows this in mean traces (Figure 4A to 4C) and per-period bars (Figure S5, onset first 1 s, late cue last 5 s, delay 2 s), with no test. Judged on each cluster's mean normalized firing in the Figure 3A and 3C interval (the paper's per-cluster periods). Reproduced when (1) for k4 and k6 danger mean < 0 and safety mean > 0, and for k3 and k5 danger mean > 0 and safety mean < 0; and (2) in at least six of the seven clusters other than k2, the uncertainty mean has the sign of the danger mean and a smaller absolute value. Partially reproduced when (1) holds and (2) holds in four or five. The onset, late cue and delay means of every cluster are reported (Figure S5).
- fig4def-danger-uncertainty-correlation
Across all cue-responsive units, danger and uncertainty firing are positively related at cue onset (R2 = 0.48), late cue (R2 = 0.21) and delay (R2 = 0.21).
there was a positive relationship between danger and uncertainty firing across all clusters during each period
Results, Uniform firing bias toward threat cues, paragraph 2; Figure 4D to 4FHow it will be judged
Per unit, mean normalized firing to danger and to uncertainty in the first 1 s of the cue, the last 5 s, and the 2 s after cue offset (Figure 4 legend); Pearson correlation over the 423 units. Reproduced when all three correlations are positive with p < 0.05 and R2 within tolerance under z-a or z-b; partially reproduced when all three are positive and significant but an R2 is outside tolerance.
- fig4ghi-danger-safety-correlation
Danger and safety firing are positively related at cue onset (R2 = 0.14) but negatively related in the late cue (R2 = 0.03) and delay (R2 = 0.02).
There was a positive relationship between danger and safety firing at cue onset (R2 = 0.14 ... but negative relationships were observed during late cue (R2 = 0.03 ... and delay (R2 = 0.02, p = 0.008)
Results, Uniform firing bias toward threat cues, paragraph 2; Figure 4G to 4IHow it will be judged
As fig4def-danger-uncertainty-correlation with safety in place of uncertainty. The late cue and delay R2 are printed to two decimals, so their tolerance is the rounding half-step (0.005), wider than 10%. Reproduced when the onset correlation is positive, the late cue and delay correlations negative, all with p < 0.05, and the R2 within tolerance; partially reproduced when the signs and significance hold but an R2 does not.
- fig4-correlations-differ
The danger-uncertainty and danger-safety correlations differ from one another in each period (Z = 7.69 onset, 8.94 late cue, 8.20 delay).
Fisher r-to-z transformation revealed the danger-uncertainty and danger-safety relationships differed from one another during each period: onset (Z = 7.69), late cue (Z = 8.94), and delay (Z = 8.20)
Results, Uniform firing bias toward threat cues, paragraph 2; Figure 4 legendHow it will be judged
Meng, Rosenthal and Rubin's Z-test for two correlations sharing danger firing, using the uncertainty-safety correlation of the same units (STAR Methods; the MATLAB File Exchange routine the paper names). The test returned 0.00 for p, recorded as 0. Reproduced when each Z is within tolerance with p < 0.05; partially reproduced when all three are significant but a Z is outside tolerance.
- fig5a-tuning-peaks-exaggerated
Threat tuning curves of k1, k3 and k5 peak at an uncertainty assignment of 0.375 and k7 at 0.500, above the actual shock probability of 0.250 (exaggerated threat probability signaling).
In support, tuning curve peaks exceeded 0.250: k1, k3, and k5 = 0.375 and k7 = 0.500.
Results, Exaggerated threat probability signaling, paragraphs 1 and 2; Figure 5A and 5BHow it will be judged
Per unit, a regression (MATLAB regress, with a constant) of each trial's normalized firing on the trial's suppression ratio (fear output) and the cue's threat probability (danger 1, safety 0, uncertainty set in turn to 0, 0.125 ... 1), over the 16 trials of the unit's session, run in each 1-s bin (STAR Methods, Single-unit linear regression and Threat probability tuning curve; the Methods print 0.025 for 0.375, a typo). The tuning curve is the cluster mean of the threat probability beta per assignment; the peak is the assignment of largest mean. Open are (t-a) which bins are averaged, the 10 cue bins (Methods) or the cluster-specific periods of Figure 3A and 3C (Results), which for k3, k5 and k7 include the 2-s delay; and (t-b) how a single trial's firing is normalized, either the trial's baseline-subtracted rate z-scored with the unit's mean and SD from the shared normalization, or the trial rates z-scored across the unit's 16 trials. All four combinations are run. The grid step (0.125) is larger than 10%, so each peak must match exactly. Reproduced when, under one reading, all four peaks match; partially reproduced when all four exceed 0.250 but not all match.
- fig5a-tuning-peaks-inhibited
Among the cue-inhibited clusters, k4 (peak 0.875) and k6 (0.625) show even greater exaggeration of threat probability, while k8 peaks at the actual probability, 0.250, flattening toward 0.000.
Tuning curve peaks for each cluster were k4 = 0.875 and k6 = 0.625, now exceeding even fear output. K8 neurons were more tuned to the actual shock probability, with a peak of 0.250 that flattened toward 0.000.
Results, Exaggerated threat probability signaling, paragraph 2; Figure 5AHow it will be judged
As fig5a-tuning-peaks-exaggerated, with the betas of k4, k6 and k8 multiplied by -1 (Results). Reproduced when, under one reading, all three peaks match; partially reproduced when k4 and k6 peaks exceed 0.375 and k8 peaks at 0.375 or below.
- fig5a-tuning-anova
The tuning curves differ between clusters (main effect of cluster and cluster x assignment interaction) and show a main effect of assignment.
Consistent with different firing magnitudes and non-identical tuning, ANOVA (factors: cluster [7] and assignment [9]) revealed a main effect of cluster ... and cluster x assignment interaction ... Demonstrative of uncertainty over-responding across clusters, ANOVA found a main effect of assignment
Results, Exaggerated threat probability signaling, paragraph 2; Figure 5AHow it will be judged
Per-unit mean threat probability betas (sign-inverted for k4, k6, k8) at the nine assignments, from fig5a-tuning-peaks-exaggerated under its readings; ANOVA with assignment within units and cluster between (the printed df, 357 = 364 units - 7, fit this). Reproduced when all three effects have p < 0.05 under one reading.
- fig5e-peak-probability-over-fear-output
With peak threat probability and fear output as regressors, neither is signaled before the cue; peak threat probability is signaled from cue onset through the delay, fear output in every bin but the first 1 s of the cue, and peak threat probability signaling exceeds fear output signaling in every cue and delay bin.
Neither peak threat probability nor fear output signaling was observed prior to cue onset. Across all seven clusters (Figure 5E), peak threat probability was continuously signaled from cue onset through delay. Fear output was signaled in all bins except for the first 1 s of the cue; however, peak threat probability signaling exceeded fear output signaling at each time point.
Results, Exaggerated threat probability signaling, paragraph 3; Figure 5C to 5E; Table S2How it will be judged
Per unit of the seven clusters, the regression of fear-output and threat-probability regressors with uncertainty set to the cluster's peak (k1, k3, k5 0.375; k4 0.875; k6 0.625; k7 0.500; k8 0.250, as the paper reports them; the peaks found here are a variant) in 14 1-s bins from 2 s before cue onset through the 2-s delay. Means and studentized bootstrap CIs per bin of each beta and of their difference (Table S2; tolerances are the distance to the farther bound). Open are the trial normalization (t-b of fig5a-tuning-peaks-exaggerated) and (s-a) whether the betas of k4, k6 and k8 are sign-inverted here as for the tuning curves (Figure 5E shows positive values) or left as computed; all readings are run. The ANOVA has regressor within units; the Results name bin a between factor, but the printed df (13 and 4615 = 13 x 355) fit bin within units, which is the main reading; 355 error df imply 362 units, not 364, and the units dropped are to be reported. Reproduced when, under one reading, the CI pattern matches (both betas include zero in pre1 and pre2; peak probability excludes zero in all 12 cue and delay bins; fear output excludes zero in all but cue1; the difference excludes zero, peak probability larger, in all 12), the regressor main effect and regressor x bin interaction have p < 0.05, and at least 24 of the 28 means are within tolerance. Partially reproduced when the regressor main effect holds and peak probability exceeds fear output in at least 9 of the 12 bins.
- fig5-regressor-cluster
Signaling varied somewhat between clusters (regressor x cluster interaction) with no regressor x cluster x bin interaction.
ANOVA revealed a regressor x cluster interaction (F6,355 = 2.15, p = 0.048, hp2 = 0.04, op = 0.76) but no regressor x cluster x bin interaction (F78,4615 = 1.06, p = 0.33
Results, Exaggerated threat probability signaling, paragraph 3; Figure S6; Table S3How it will be judged
The ANOVA of fig5e-peak-probability-over-fear-output under its readings. The two-way term is a presence and the three-way an absence, judged on this ANOVA at p < 0.05. Reproduced when the regressor x cluster interaction has p < 0.05 and the three-way interaction does not; partially reproduced when only the absence holds. Both effect sizes are reported.
- table-s3-cluster-regression
Within clusters, peak threat probability exceeds fear output (main effect of regressor) in k3, k4, k6, k7 and k8 but not in k1 or k5, and only k7 shows a regressor x bin interaction.
How it will be judged
Per cluster, the ANOVA of fig5e-peak-probability-over-fear-output without the cluster factor (regressor and bin within units). The other regressor x bin interactions are printed as non-significant (k1 0.07, k3 0.09, k4 0.48, k5 0.79, k6 0.08, k8 0.27). Reproduced when the regressor main effect is significant (p < 0.05) in exactly the five clusters named, with peak probability larger, and the interaction in k7 only; partially reproduced when at least five of the seven clusters agree on both terms.
- fig5f-fear-output-vs-peak-probability
Across units there is a continuous negative relationship between fear output and peak threat probability betas (R2 = 0.43), with signaling biased toward peak threat probability, so the two are not signaled by distinct populations.
and fear output were not discretely signaled by distinct RRF populations. A continuous, negative relationship (R2 = 0.43
Results, Exaggerated threat probability signaling, paragraph 3; Figure 5F; Figure S6C to S6UHow it will be judged
Per unit of the seven clusters, the mean of each beta from fig5e-peak-probability-over-fear-output; Pearson correlation across units. Open is which bins are averaged, (f-a) the 10 cue bins (Figure S6B, "during 10-s cue interval") or (f-b) all 14 bins, together with that claim's readings. Reproduced when, under one reading, r < 0, p < 0.05 and R2 within tolerance; partially reproduced when r < 0 with p < 0.05 only. The per-cluster R2 printed in Figure S6 (0.64, 0.46, 0.55, 0.65, 0.55, 0.58, 0.64 in panel order k1, k3, k4, k5, k6, k7, k8) are reported against those found here.
- table1-outcome-anova
Every cluster showed differential firing after shock delivery or omission, a main effect of trial type and/or a trial type x interval interaction.
found a main effect of trial type and/ or a trial type x interval interaction for every cluster (Table 1)
Results, Diverse aversive outcome firing patterns, paragraph 1; Table 1 (Aversive outcome)How it will be judged
Per cluster, normalized firing in 70 250-ms bins over the 17.5 s after shock offset for danger, uncertainty shock, uncertainty omission and safety; repeated-measures ANOVA with trial type and interval within units (header). Reproduced when every cluster has at least one of the two effects at p < 0.05 and the pattern of significance matches Table 1 (both effects in all but k3, interaction only, and k6, main effect only) in at least 14 of the 16 tests; partially reproduced when every cluster has an effect but fewer tests match.
- fig6a-k1-shock
k1 units are generally shock responsive, with equivalent, transient firing after danger and uncertainty shock.
K1 neurons (Figure 6A, i and ii) were generally shock responsive, showing equivalent and transient responding following danger and uncertainty shock.
Results, Diverse aversive outcome firing patterns, paragraph 2; Figure 6A; Table S4How it will be judged
The shared outcome method for Figure 6. Per unit, mean normalized firing per trial type in the gray-box interval of Figure 6 (k1, 0 to 8 s after shock offset, read from the figure); per-cluster studentized bootstrap CIs of each trial type (Table S4) and of the per-unit differences uncertainty shock - danger, uncertainty omission - safety and danger - safety (Figure 6, column ii). Tolerances are the distance from the Table S4 mean to its farther bound. The paper's pattern for k1 is danger and uncertainty shock CIs above zero, uncertainty omission including zero, safety above zero, and of the differences only omission - safety below zero. Reproduced when all seven CIs agree with the paper's on excluding zero and sign, and the four means are within tolerance; partially reproduced when the shock - danger difference includes zero and both shock CIs are above zero.
- fig6b-k2-anti-ppe
k2 units fire most to the predicted shock on danger trials and are inhibited by the surprising shock on uncertainty trials, an anti-positive prediction error signal.
K2 neurons (Figure 6B, i and ii) fired maximally to the fully predicted shock on danger trials but showed firing inhibition to the surprising shock on uncertainty trials, consistent with an anti-positive prediction error signal.
Results, Diverse aversive outcome firing patterns, paragraph 2; Figure 6B; Table S4How it will be judged
As fig6a-k1-shock over the whole 17.5 s (gray box of Figure 6B). Paper's pattern is danger above zero, uncertainty shock below, omission including zero, safety above; shock - danger and omission - safety below zero, danger - safety above. Reproduced when all seven CIs agree and the means are within tolerance; partially reproduced when danger is above zero, uncertainty shock below zero and shock - danger below zero.
- fig6c-k3-continued-threat
k3 units sharply reduce firing after shock but sustain excitation after omission on uncertainty trials, a continued threat signal.
K3neurons(Figure 6C, i and ii) sharply reduced firing following shock but sustained excitation of firing following omission on uncertainty trials, revealing a continued threat signal.
Results, Diverse aversive outcome firing patterns, paragraph 2; Figure 6C; Table S4How it will be judged
As fig6a-k1-shock over 0 to 8 s after shock offset. Paper's pattern is danger, uncertainty shock and uncertainty omission including zero (the omission CI just does, [-0.02, 0.43]), safety below zero; only omission - safety above zero. Reproduced when the seven CIs agree and the means are within tolerance, or when they differ only in the omission CI excluding zero above (the Results call it sustained excitation); partially reproduced when omission - safety is above zero and both shock CIs include zero.
- fig6de-k4-k5-persistent-threat
k4 units sustain inhibition and k5 units sustain excitation on uncertainty omission trials, persistent threat signaling through opposite firing directions.
K4 neurons (Figure 6D, i and ii) sustained firing inhibition on uncertainty omission trials, whereas k5 neurons (Figure 6E, i and ii) sustained firing increases, revealing similar and persistent threat signaling through opposing firing directions.
Results, Diverse aversive outcome firing patterns, paragraph 2; Figure 6D and 6E; Table S4How it will be judged
As fig6a-k1-shock, k4 over the whole 17.5 s and k5 over 0 to 8 s after shock offset. Paper's pattern is k4 danger, shock and safety above zero, omission below, of the differences only omission - safety below zero; k5 danger and shock including zero, omission above, safety below, of the differences only omission - safety above zero. Reproduced when all CIs of both clusters agree and the means are within tolerance; partially reproduced when the omission CI and omission - safety have the paper's sign in both clusters.
- fig6f-k6-ppe-then-valence
k6 units show early positive prediction error through firing decreases (more inhibition after uncertainty shock than danger shock), giving way to a sustained valence signal with most firing to safety, little after omission and sustained inhibition after danger and uncertainty shock.
K6 neurons (Figure 6F, i and ii) showed early positive prediction error signaling through firing decreases: greater firing inhibition following shock delivery on uncertainty trials compared to danger trials.
Results, Diverse aversive outcome firing patterns, paragraph 2; Figure 6F; Table S4How it will be judged
As fig6a-k1-shock over 0 to 8 s after shock offset. Paper's pattern is danger and shock below zero, omission including zero, safety above; all three differences below zero. The early-to-late change is shown in the traces only; the 0-8 s window is judged and the means of the full 17.5 s are reported. Reproduced when the seven CIs agree and the means are within tolerance; partially reproduced when shock - danger is below zero and safety is the largest mean.
- fig6g-k7-valence-then-ppe
k7 units show initial valence signaling (danger shock = uncertainty shock > omission > safety) giving way to positive prediction error (uncertainty shock > danger shock), and sustain inhibition to safety through the outcome period.
K7 neurons (Figure 6G, i and ii) showedinitial valence signaling ... that gave way to positive prediction error (uncertainty shock > danger shock). Impressively, k7 neurons sustained firing inhibition to safety throughout the outcome period.
Results, Diverse aversive outcome firing patterns, paragraph 2; Figure 6G; Table S4How it will be judged
As fig6a-k1-shock over the whole 17.5 s. Paper's pattern is danger, shock and omission above zero, safety below; all three differences above zero. Reproduced when the seven CIs agree and the means are within tolerance; partially reproduced when shock - danger is above zero and safety is below zero.
- fig6h-k8-ppe-then-unsigned
k8 units show early positive prediction error (uncertainty shock > danger shock) giving way to an unsigned prediction error, with equal firing increases on uncertainty shock and omission trials exceeding those for danger and safety.
Finally, k8 neurons (Figure 6H, i and ii) showed early positive prediction error signaling (uncertainty shock > danger shock) that gavewaytoanunsignedpredictionerrorsignalinwhich equivalent firing increases on uncertainty shock and omissiontrials exceeded those for danger and safety.
Results, Diverse aversive outcome firing patterns, paragraph 2; Figure 6H; Table S4How it will be judged
As fig6a-k1-shock over the whole 17.5 s. Paper's pattern is danger below zero, shock and omission above, safety including zero; shock - danger and omission - safety above zero, danger - safety below. Reproduced when the seven CIs agree and the means are within tolerance; partially reproduced when shock - danger and omission - safety are above zero.
Pipeline
- Paper
- waiting (not checked)
- Claims
- done
- Claims check
- done (ready)
- Attempt
- todo
- Questions
- none
- Verified
- none