CRCNS:alm-3 · Encore Science

Claims 29

From the paper Robust neuronal dynamics in premotor cortex during motor planning · claims file · checked 2026-11-07: ready

units-and-sessions
1012 single units were recorded in left ALM during behaviour across 58 sessions (20 unilateral, 38 bilateral) from 12 mice; 72 are fast-spiking, 890 putative pyramidal and 50 of intermediate width are excluded.
1012 single units were recorded during behavior across 58 recording sessions (20 sessions of unilateral experiments, Fig. 1 ; 38 sessions of bilateral experiments ... (FS) neurons (72/1012) and units with spike widths > 0.45 ms as putative pyramidal neurons (890/1012). Units with intermediate values (0.35 - 0.45 ms, 50/1012) were excluded
Methods, Electrophysiology data analysis, paragraph 1; Results, Preparatory activity in ALM
How it will be judged
Counts over the non-callosotomy sessions, units classed by the spike-width bounds of the Methods (M3). Reproduced when the counts match; partially reproduced when the unit total matches and the class counts are within 10%.
ed1-selective-fraction
Most ALM pyramidal neurons distinguish trial types (634 of 890), 176 during the sample, 337 during the delay and 493 during the response epoch.
Most ALM pyramidal neurons distinguished trial types (634/890, p<0.05, t-test; sample epoch, 176/890; delay epoch, 337/890; response epoch, 493/890, Methods)
Results, Preparatory activity in ALM; Extended Data Fig. 1b, e
How it will be judged
Two-tailed t-test on epoch spike counts, lick left against lick right, p < 0.05 (M3), on control trials, under the readings of M1, M2 and M3. Reproduced when all four counts are within tolerance under one reading; partially reproduced when the any-epoch count is.
fig1d-late-delay-ipsilateral-bias
Unilateral photoinhibition (0.5 s) of ALM ending at the go cue causes an ipsilateral response bias.
Transient (duration, 0.5 s) unilateral photoinhibition of ALM up to the ‘go’ cue (late delay) caused an ipsilateral response bias (n = 5 mice, p < 0.01, two-tailed t-test
Results, Preparatory activity after unilateral silencing, paragraph 1; Figure 1d
How it will be judged
Performance (M1), photoinhibition against control, two-tailed t-test across mice (M7). Which performance is tested is open (the Methods list separate lick right and lick left performance for ED Fig. 3, not for Fig. 1); readings (a) trials instructing a lick contralateral to the inhibited hemisphere (lick right for left ALM), (b) all trials. Reproduced when performance drops with p < 0.01 under one reading. These are behaviour experiments (Fig. 1d, ED Fig. 3); the condition is tested only if the data set holds it.
fig1d-earlier-photoinhibition-minimal
Unilateral photoinhibition ending at least 0.3 s before the go cue (early and middle delay) has minimal behavioural effect.
In contrast, photoinhibition ending at least 0.3 s prior to the ‘go’ cue produced minimal behavioral effects (middle delay, early delay; p > 0.1, two-tailed t-test)
Results, Preparatory activity after unilateral silencing, paragraph 1; Figure 1d; Extended Data Fig. 3b, d
How it will be judged
A claim of absence, judged on the paper's test. Per condition (window ending 0.8 s and 0.3 s before the cue), overall performance (M1), photoinhibition against control, two-tailed t-test across mice (M7 reading a, paired, a choice since the paper does not say; reading b shown as a variant), at the paper's significance level p < 0.05 (ED Fig. 3b asterisks). Reproduced when neither condition has p < 0.05; not reproduced when either does; the p-values are reported against the paper's p > 0.1 and the drop in performance is reported either way. The sample-epoch condition is reported as well.
ed3-recovery-interval
Performance is not significantly affected when photoinhibition (0.5 s or 1.3 s) ends more than 0.3 s before the go cue, and 0.5 s of photoinhibition just before the cue has an effect similar to photoinhibition of the whole delay epoch.
Performance was not significantly affected for T recovery >0.3s ... Photoinhibition (0.5 s) immediately before the ‘go’ cue is similar to the behavioral effect caused by photoinhibition during the entire delay epoch (1.3 s)
Extended Data Fig. 3b, d, e
How it will be judged
Behaviour-only experiments; the data set's documentation says these sessions are not included. If they are found, judged as fig1d-earlier-photoinhibition-minimal per T recovery (two-tailed t-test, p < 0.05 the threshold of the figure's asterisks). The paper gives no test for "similar" (ED Fig. 3e); as a choice stated here, the 0.5 s and 1.3 s conditions at T recovery = 0 are similar when a two-tailed t-test across mice finds no difference at p < 0.05.
ed2b-silencing
Photoinhibition silences more than 90% of putative pyramidal activity, similarly with ipsilateral and bilateral photoinhibition and in VGAT-ChR2 and PV x ReaChR mice, while contralateral photoinhibition barely changes the average rate.
Ipsilateral and bilateral photoinhibition similarly silenced neuronal activity. Average spike rate across the population was little affected by contralateral photoinhibition ... Photoinhibition was similar in the two mouse lines (>90% activity reduction)
Extended Data Fig. 2b; Methods, Photoinhibition
How it will be judged
Per neuron, spike rate during photoinhibition over the rate in the same window of control trials (ED Fig. 2b), neurons with mean rate < 1 spike/s excluded; the window is the photoinhibition period before the ramp (a), or the whole period with the ramp (b). Reproduced when the mean reduction (1 minus the mean normalized rate) is within tolerance of 0.9, that is at least 0.81, for ipsilateral and bilateral photoinhibition in each experiment and for each mouse line (ED Fig. 2b right, Experiment 2 ipsilateral), and the contralateral mean normalized rate lies within 0.8 to 1.2 (the 10% default applied to 1, doubled since the paper gives no number; stated here as a choice).
ed2d-spatial-extent
The standard photostimulus silences about 90% of spikes in a cortical area of 1 mm radius at half maximum, through all layers.
The photoinhibition silenced 90% of spikes ( Extended Data Fig. 2b ) in a cortical area of 1mm radius (at half-max) through all cortical layers
Methods, Photoinhibition; Extended Data Fig. 2d
How it will be judged
Measured in 4 untrained, non-behaving mice (ED Fig. 2d), which the data set's description does not list; tested only if those recordings are found.
ed2e-no-rebound
No rebound excitation is seen on average after photoinhibition offset; a small proportion of neurons show rebound, balanced by low-level sustained inhibition in a larger proportion.
No rebound excitation was detected after photoinhibition offset on average ( d ). A small proportion of neurons showed rebound excitation which was balanced by a low level of sustained inhibition in a larger proportion of neurons
Extended Data Fig. 2c, e; Results, Preparatory activity after unilateral silencing, paragraph 2
How it will be judged
Experiment 2, lick right trials, ipsilateral photoinhibition (ED Fig. 2e), windows the first 400 ms and 400 to 800 ms after photoinhibition. Per neuron, a two-tailed t-test of spike counts, control against photoinhibition, p < 0.01 (M5 on lick right trials). Reproduced when the population mean rate in each of the two windows after offset is not higher than control by a paired t-test over neurons at p < 0.05 (a standard choice, the paper names no population test) and fewer neurons are significantly excited than significantly suppressed in the first 400 ms after offset.
fig1f-rate-recovery
Within 400 ms after unilateral photoinhibition, spike rates are indistinguishable from control in 90% of neurons; only 10% keep a sustained change.
Within 400 ms after photoinhibition, spike rates became indistinguishable from the unperturbed condition in 90% of the neurons; only 10% of neurons retained a sustained change in spike rate
Results, Preparatory activity after unilateral silencing, paragraph 2; Figure 1f
How it will be judged
Experiment 1, per neuron and condition, the rate test of M5. The fraction is over neurons of a condition (n = 168, 168, 175 in Fig. 1f; the inclusion rule is open, M3). Which window is "within 400 ms after" is open; readings (a) the 400 ms window starting at offset, (b) the window whose time stamp, by its centre, is 400 ms after offset. Reproduced when, under one reading, the fraction is within tolerance in each condition, or in their mean; partially reproduced when it is below 0.2. The neuron counts are reported against 168, 168 and 175.
fig1g-selectivity-recovery
After unilateral photoinhibition ALM neurons recover more than 80% of their control selectivity within 514 ms; recovery times are 373 +- 260 ms (sample), 510 +- 218 ms (early delay) and 327 +- 112 ms (window ending 0.3 s before the cue).
ALM neurons recovered > 80% of their selectivity relative to the unperturbed trials within 514 ms after photoinhibition
Results, Preparatory activity after unilateral silencing, paragraph 2; Figure 1g
How it will be judged
Experiment 1, selective neurons with > 3 trials in every condition, M4 with all readings. Reproduced when, under one reading, every condition's mean selectivity reaches 80% of control within tolerance of the reported time; partially reproduced when every condition reaches 80% before the go cue.
ed4-return-to-trajectory
Neurons whose rates decrease during the delay epoch also recover after unilateral photoinhibition, returning to the time-matched control trajectory and not decaying to the end-point rate.
Spike rate difference relative to time-matched control show significantly smaller root-mean-squared error (RMS) than spike rate difference relative to end point (p<0.001, paired t-test)
Extended Data Fig. 4b, c; Results, Preparatory activity after unilateral silencing, paragraph 2; Figure 1e
How it will be judged
Experiment 1 pyramidal neurons with rate at the end of the delay lower than at its start (two-tailed t-test, 400 ms windows, trial types pooled, p < 0.05; ED Fig. 4b). Rates normalized to each neuron's mean rate, preferred and non-preferred types separately. Per neuron, RMS over the time from photoinhibition offset to the go cue of (perturbed minus time-matched control) and of (perturbed minus control rate in the last 400 ms of the delay); which photoinhibition conditions are pooled is not stated, readings (a) the early-delay condition, (b) all conditions ending before the cue pooled per neuron. Reproduced when the first RMS is smaller by a paired t-test at p < 0.001 under one reading; partially reproduced at p < 0.05. The same test on neurons with an increasing rate (Fig. 1e, neurons 1 and 2) is shown as a variant.
ed5-photoactivation
After transient photoactivation of layer 5 IT neurons, ALM selectivity recovers to 80% of control in 249 +- 68 ms (sample), 275 +- 168 ms (early delay) and 250 +- 218 ms (middle delay).
After photostimulus offset, ALM activity and selectivity recovered with a time-course that was similar to recovery after photoinhibition
Results, Preparatory activity after unilateral silencing, paragraph 2; Extended Data Fig. 5f
How it will be judged
M4 on 26 selective neurons from 2 Tlx_PL56-cre x Ai32 mice. The data set's documentation says the photoactivation data are not included; tested only if found.
fig2b-bilateral-behaviour
Unilateral photoinhibition in the early delay reduces performance only slightly (1 spot 70.3%, 4 spots 67.7%), while bilateral photoinhibition degrades it severely (1 spot per side 58.0%) or to near chance (4 spots per side 51.4%).
Behavioral performance was only slightly affected after unilateral photoinhibition with a single spot during the early delay epoch ( Fig. 2b , 70.3% correct, p = 0.009, two-tailed t-test against control); unilateral photoinhibition with four spots had small additional effect (67.7%, p=0.003). In contrast, using only two spots across both hemispheres caused performance to degrade severely ( Fig. 2b , 58.0%, p<0.001; difference from 4 spot unilateral: p<0.05, two-tailed t-test); four spots bilaterally further reduced performance to near chance level (4 laser spots: 51.4%, p<0.001)
Results, Preparatory activity after bilateral silencing, paragraph 1; Figure 2b
How it will be judged
Performance (M1) per condition over mice, tests of M7. The paper's numbers come from 13 mice including behaviour-only sessions and photoinhibition of 0.8 s and 1.3 s; the data set's recording sessions hold 0.8 s photoinhibition from delay onset. Reproduced when each value is within tolerance, each condition is below control with p < 0.05, and bilateral 1 spot is below unilateral 4 spots with p < 0.05 (two-tailed t-test, as the paper); partially reproduced when the order (control > unilateral > bilateral 1 spot > bilateral 4 spots) holds but some values are outside tolerance.
ed6-bilateral-bias-inconsistent
Bilateral photoinhibition biases movements inconsistently across mice and sessions, and the direction of a session's bias follows which control trajectory its perturbed cd trajectories approach.
Bilateral photoinhibition biased movements inconsistently across mice and sessions
Results, Preparatory activity after bilateral silencing, paragraph 1; Extended Data Fig. 6
How it will be judged
Experiment 2 sessions of M6, bilateral photoinhibition. Per session, bias = performance on lick right trials minus performance on lick left trials under photoinhibition, and the side the mean perturbed cd projection in the last 400 ms of the delay lies on; the ED Fig. 6b legend says "closer to" a control trajectory, read as (a) the side of the decision boundary, (b) the nearer of the mean control lick right and lick left projections. The paper gives no test; as a standard choice, a session is biased when the two performances differ by Fisher's exact test at p < 0.05. Reproduced when biased sessions of both signs occur and the sign of the bias agrees with the side of the decision boundary in at least 80% of biased sessions (the paper's five example sessions all agree; the 80% is a choice stated here).
fig2-bilateral-recovery
Spike rate and selectivity of left ALM neurons recover within 600 ms after ipsilateral photoinhibition, contralateral photoinhibition hardly affects activity, and after bilateral photoinhibition the rate recovers on average but selectivity does not; selectivity recovery takes 538 +- 178 ms (ipsilateral), 192 +- 114 ms (contralateral) and does not occur (bilateral).
As before ( Fig. 1 ), spike rate (> 90 %) ( Fig. 2c, d ) and selectivity (> 80%) ( Fig. 2e ) recovered 600 ms after ipsilateral perturbation. ALM activity was hardly affected by contralateral photoinhibition ... After bilateral photoinhibition neurons recovered their spike rate on average ( Fig. 2c, d ), but selectivity failed to recover
Results, Preparatory activity after bilateral silencing, paragraph 2; Figure 2c-e
How it will be judged
Experiment 2 (1 spot ipsilateral, 1 spot contralateral, 4 spots per side bilateral), M4 and M5. "Spike rate (> 90 %)" is read as (a) more than 90% of neurons without a significant rate change (M5) in the window 600 ms after offset, or (b) population mean rate within 10% of control there. Reproduced when, under one reading of each open detail, the ipsilateral rate and selectivity recover as stated with the recovery times within tolerance, the contralateral recovery time is within tolerance, and the bilateral selectivity stays below 80% of control up to the go cue while the bilateral mean rate is back within 10% of control by the cue (the 10% is a choice stated here; the paper says only that the rate recovered on average). Partially reproduced when the ordering holds (ipsilateral recovers, bilateral does not) but a time is outside tolerance.
ed7-unilateral-4-spots-bilateral-1-spot
Bilateral photoinhibition with one spot per side causes a larger persistent change in selectivity than unilateral photoinhibition with four spots; recovery takes 490 +- 280 ms (ipsilateral 4 spots), 235 +- 156 ms (contralateral 4 spots) and does not occur (bilateral 1 spot).
Bilateral photoinhibition with one laser spot produced a larger persistent change in selectivity than unilateral photoinhibiton with four laser spots
Results, Preparatory activity after bilateral silencing, paragraph 2; Extended Data Fig. 7d
How it will be judged
Experiment 3 sessions, M4. Reproduced when the recovery times are within tolerance under one reading, bilateral 1-spot selectivity does not reach 80% of control before the go cue, and the mean delta selectivity in the last 400 ms of the delay (a window chosen here; the paper names none for "persistent change") is larger in magnitude for bilateral 1 spot than for ipsilateral 4 spots.
ed7e-size-and-roc
Recovery of selectivity after unilateral photoinhibition is less robust with larger photoinhibition, and cd trajectories separate at the end of the delay with ROC 0.80 +- 0.08 (control), 0.64 +- 0.10 (ipsilateral 4 spots), 0.68 +- 0.15 (contralateral 4 spots) and 0.54 (bilateral 1 spot).
Recovery of selectivity after unilateral photoinhibition was less robust with larger photoinhibition size ( Extended Data Fig. 7e ) ... control, 0.80 ± 0.08; ipsilateral, 0.64 ± 0.10; contralateral, 0.68 ± 0.15; bilateral, 0.54 ± 0.8
Results, Preparatory activity after bilateral silencing, paragraph 2; Extended Data Fig. 7e
How it will be judged
Experiment 3 sessions with > 5 simultaneously recorded neurons tested for > 3 trials in each condition (ED Fig. 7e legend), M6 otherwise. The printed s.e.m. of the bilateral value, 0.8, cannot be an s.e.m. of an ROC value, so the tolerance is the default 10% of the value (0.054); 0.08, if the 0.8 is a misprint for it, is shown as a variant. Reproduced when all four ROC values are within tolerance under one reading; "less robust with larger size" is judged (a reading) by the ipsilateral 4-spot ROC being below the Experiment 2 ipsilateral 1-spot ROC of fig3-cd-roc.
fig3-cd-roc
Along the coding direction, trajectories for lick right and lick left recover after ipsilateral photoinhibition (ROC at the end of the delay 0.73 +- 0.02 against control 0.76 +- 0.03), are little affected by contralateral photoinhibition (0.74 +- 0.03), and stay collapsed after bilateral photoinhibition (0.58 +- 0.03).
ROC values between trajectories at the end of delay epoch: control, 0.76 ± 0.03; ipsilateral, 0.73 ± 0.02; mean ± s.e.m. across sessions, Methods). Contralateral photoinhibition had little effect ( Extended Data Fig. 8 ; ROC, 0.74 ± 0.03). As expected ( Fig. 2 ), trajectories were permanently altered after bilateral photoinhibition, resulting in small separation between the trajectories for different trial types at the time of movement onset (ROC, 0.58 ± 0.03)
Results, Robustness along the coding direction, paragraph 2; Figure 3b, c; Extended Data Fig. 8
How it will be judged
M6 with all readings. Reproduced when the 16 sessions are found and all four ROC values are within tolerance under one reading; partially reproduced when the ordering holds (control, ipsilateral and contralateral each above bilateral by more than 0.1, a margin chosen here) but a value is outside tolerance.
ed8-roc-during-photoinhibition
During photoinhibition the cd separation falls (ROC 0.54 +- 0.03 ipsilateral, 0.64 +- 0.03 contralateral, 0.54 +- 0.01 bilateral, against control 0.72 +- 0.02).
ROC values during the time period of photoinhibition: control, 0.72 ± 0.02; ipsilateral, 0.54 ± 0.03; contralateral, 0.64 ± 0.03; bilateral 0.54 ± 0.01
Extended Data Fig. 8 legend
How it will be judged
M6; the window "during photoinhibition" is open, readings (a) the 400 ms window ending at the start of the offset ramp, (b) projections averaged over the whole 0.8 s photoinhibition. Reproduced when all four are within tolerance under one reading.
ed9b-cd-stability
The direction of the lick right minus lick left response difference is stable through the sample and delay epochs (correlation between late sample and late delay 0.61 +- 0.05).
During the sample and delay epochs the direction of the difference in the mean response vectors, w t = x̄ lick right – x̄ lick left , was stable (correlation of w t 's between late sample epoch vs. late delay epoch, 0.61 ± 0.05
Methods, Analysis of population dynamics in the activity space; Extended Data Fig. 9b
How it will be judged
M6 sessions. Per session, Pearson correlation of w in the last 400 ms of the sample epoch and in the last 400 ms of the delay epoch (the "late" windows are read as these; a reading), on correct control trials; mean +- s.e.m. over sessions. Reproduced when the mean is within tolerance.
ed8a-mode-variance
The projection on cd captures 65.6 +- 5.1% of population selectivity and 36.4 +- 6.3% of task-related activity variance over the sample and delay epochs; modes 1 to 5 together explain 98.5 +- 0.5% of the variance and 95.8 +- 1.2% of the selectivity.
The projection along the cd captured 65.6±5.1% of the population selectivity ... and 36.4±6.3% of the total variance in ALM task-related activity ... Mode 1-5 together explained 98.5±0.5% of the total variance of task-related activity and 95.8±1.2% of population selectivity over the sample and delay epochs
Methods, Analysis of population dynamics in the activity space; Extended Data Fig. 8a
How it will be judged
M6 sessions; modes as in the Methods (mode 2, perturbed minus unperturbed lick right mean response at the go cue; mode 3, the same for lick left, Gram-Schmidt orthogonal to mode 2; modes 4 and 5, the first two singular vectors of the baseline-subtracted lick right and lick left PSTHs concatenated, Gram-Schmidt orthogonal to modes 1 to 3), from correct control and ipsilateral trials. Selectivity is the RMS over time of the lick right minus lick left rate difference, variance the RMS of baseline-subtracted activity, over sample and delay. Open details. The baseline; readings (a) the mean rate before sample onset, (b) the mean rate over the whole trial. How a mode's share is computed; readings (a) the RMS of the projection over the RMS of the full activity, (b) the ratio of squared norms. "At the go cue" read as the last 400 ms of the delay. Reproduced when all four values are within tolerance under one reading; partially reproduced when the two cd values are.
fig3-cd-predicts-movement
Deviations along the coding direction predict the upcoming movement trial by trial, in control trials and after unilateral and bilateral perturbation, even when average selectivity is destroyed.
Individual trials with a deviation toward the “lick right” trajectory along the cd predicted more frequent “lick right” responses and vice versa ( Fig. 4 , Extended Data Figs 6b and 9 ). This analysis shows that even after average selectivity is destroyed by perturbations, ALM population dynamics still dictate upcoming movements
Results, Robustness along the coding direction, paragraphs 2 and 4; Figure 3c; Figure 4b; Extended Data Fig. 8b, 9e
How it will be judged
M6, trials pooled over sessions (Methods). For each condition (control, ipsilateral, contralateral, bilateral), the fraction of trials whose lick matches the side of the decision boundary, and performance on lick right and lick left trials binned by distance from the boundary (bin 4 on the cd distance scale, ED Fig. 9e, read in the units of cd^T x after the session offset; scaling of cd is open, readings (a) cd as the raw mean difference, (b) cd of unit length). The paper gives no test; as a standard choice, the fraction of lick right responses rises with distance toward the lick right side by logistic regression of the lick on the signed distance, p < 0.05, and the prediction accuracy exceeds 0.5 by a binomial test, p < 0.05. Reproduced when both hold in all four conditions (the statement covers unilateral perturbation, Fig. 3c and ED Fig. 9e); partially reproduced when they hold in control and bilateral trials but not in a unilateral condition, or in control only.
fig3-other-modes
The persistent mode is altered by perturbation up to and beyond movement onset but neither discriminates trial type nor predicts behaviour; the ramping mode shows non-selective ramping in the delay, is resistant to unilateral and bilateral perturbation and does not predict behaviour.
However, this projection did not discriminate trial type nor predict behavior on control trials. Third, a mode that maximally captured the remaining activity variance, showed non-selective ramping during the delay epoch, did not predict behavior, and was resistant to unilateral and bilateral perturbations
Results, Robustness along the coding direction, paragraph 3; Figure 3d, e; Extended Data Fig. 8b-e
How it will be judged
Modes 2 and 4 of ed8a-mode-variance (persistent and ramping), decision boundaries as for cd (Methods). The paper gives no test; standard choices stated here. Not discriminating or predicting is judged by an ROC of lick right against lick left projections in the last 400 ms of the delay within 0.5 +- 0.1 averaged over sessions and a prediction accuracy on control trials whose binomial test against 0.5 has p > 0.05. Robustness of the ramping mode is judged by the perturbed projection in the last 400 ms of the delay lying within the s.e.m. band over sessions of the control projection for ipsilateral and bilateral trials, while the persistent-mode projection lies outside it for ipsilateral trials. Reproduced when all hold.
ed9f-reaction-time
Reaction times are faster on trials whose cd trajectory lies far from the decision boundary at the go cue.
Reaction times are faster on trials in which the trajectory is far from the decision boundary at the time of the ‘go’ cue
Extended Data Fig. 9f
How it will be judged
M6, 16 sessions; reaction time relative to the session mean, lick right and lick left trials pooled (ED Fig. 9f). The reaction time is not defined in the paper; read as the time from go cue onset to the first lick. Which trials are used is open; readings (a) control trials, (b) all conditions. The paper gives no test; as a standard choice, Spearman correlation of relative reaction time with absolute distance from the boundary, negative with p < 0.05.
fig5b-callosotomy-behaviour
Corpus callosum bisection leaves performance unaffected, but makes performance highly sensitive to unilateral photoinhibition, with a significant interaction, dropping to the level of bilateral photoinhibition in control mice.
Remarkably, behavioral performance was unaffected ( Fig. 5b , control trials, before vs. after callosotomy, p>0.05, two-tailed t-test) ... However, behavioral performance was now highly sensitive to transient unilateral photoinhibition ( Fig 2b , 1 laser spot: p=0.0019, two-tailed t-test against control). There was a significant interaction between callosotomy and unilateral photoinhibition (p=0.0035, repeated measure two-way ANOVA)
Results, Contralateral input is required for recovery, paragraph 1; Figure 5b; Extended Data Fig. 10b
How it will be judged
Performance (M1) per mouse; tests of M7 and a repeated-measures two-way ANOVA on performance with factors callosotomy (before, after) and photoinhibition (control, unilateral 1 spot). Reproduced when control performance after callosotomy does not differ from before (p > 0.05), unilateral photoinhibition lowers it with p < 0.01, and the interaction has p < 0.05, and the after-callosotomy unilateral performance is within 5.8 points (10%) of 58.0% (bilateral 1 spot in control mice, Fig. 5b cyan cross, fig2b-bilateral-behaviour). If the data set holds no sessions before the cut, the before-and-after comparison and the interaction cannot be tested and the claim is judged on the after-cut test alone, as partially testable.
fig5-callosotomy-recovery
After callosotomy, ALM neurons recover their average spike rate after ipsilateral photoinhibition but not their selectivity, selectivity recovers after contralateral photoinhibition (217 +- 228 ms) and not after bilateral, and selectivity along cd is reduced.
After ipsilateral photoinhibition, ALM neurons recovered their average spike rate ( Fig. 5c, d ), but selectivity failed to recover ( Fig. 5e ). Selectivity in the cd direction was reduced ( Fig. 5f , 16 sessions)
Results, Contralateral input is required for recovery, paragraph 2; Figure 5c-f
How it will be judged
Callosotomy sessions (conditions ipsilateral, contralateral and bilateral, Fig. 5d, e; the number of spots is not stated for these sessions, and the conditions are taken as the data set labels them), M4, M5 and M6 (the session rule of M6 applied to these sessions). Reproduced when the ipsilateral mean rate returns within 10% of control before the go cue (a margin chosen here) while selectivity stays below 80% of control up to the cue, the contralateral recovery time is within tolerance, the bilateral selectivity does not recover, and the ipsilateral cd ROC at the end of the delay is reduced; what it is reduced against is open, readings (a) the callosotomy control ROC, (b) the Experiment 2 ipsilateral ROC of fig3-cd-roc.
ed10e-callosotomy-preparatory-activity
Preparatory activity after callosotomy is similar to that of control mice, with 254 of 496 putative pyramidal neurons selective.
Preparatory activity in callosotomized mice (n=7 mice) was similar to control mice ( Fig. 5c ; Extended Data Fig. 10 ), providing additional evidence that the two ALM hemispheres can maintain preparatory activity independently
Results, Contralateral input is required for recovery, paragraph 2; Extended Data Fig. 10e, f
How it will be judged
Selectivity test of M3 on callosotomy control trials. "Similar" is read as the selective fraction within 10 percentage points of the control mice's 634/890 under the same reading (a choice stated here), and the population selectivity (M4) during the delay epoch present and rising toward the go cue as in ED Fig. 1c. Reproduced when the counts are within tolerance and the fraction is similar.
models-monolithic-fail
Monolithic network models (fixed-point attractor, integrators, trained recurrent networks) do not recover their unperturbed trajectory after transient silencing of half of their neurons.
Overall, all “monolithic” models consisting of one network were unable to explain robustness
Results, Robust model networks, paragraph 1; Figure 1c; Extended Data Fig. 1f-i, 4c
How it will be judged
A simulation that uses no recorded data (the data set's title says it holds simulations as well); parameters from the Methods and ED Fig. 1 legend (simple integrator N = 100, tau 100 ms, one eigenvalue 0.99; corrective-feedback integrator with the synaptic time constants given; FORCE network N = 400, tau 200 ms, g 1.5, p 0.1, 30 iterations; tamed chaos, 20 iterations). Silencing holds half the neurons at zero for 0.2 s < t < 1.0 s (50 random neurons in the simple integrator, 200 in the FORCE network); in the corrective-feedback integrator it is a large current into the inhibitory neuron with negative rates disallowed (ED Fig. 1g), its amplitude open. The paper gives no parameters for a monolithic fixed-point attractor; it is read as one module of the Fig. 6b attractor without inter-module connections (a reading). Open details, each a standard choice that should not change the outcome, Euler step 1 ms and random seeds; the connectivity of the simple integrator beyond its spectrum, the weight-pruning threshold of the tamed chaos network and the FORCE learning rate are open and could matter, so three random draws are run. Reproduced when, in each model, the output at t = 2 s after perturbation differs from the unperturbed output by more than 20% of the unperturbed selectivity (a threshold chosen here) in at least two of three draws.
fig6-modular-models-robust
Modular networks of two coupled modules (attractor, integrator, trained recurrent network) restore activity after transient silencing of one module but drift to a random fixed point when both are silenced.
Selective commissural connections restored activity on the other side after unilateral transient silencing ( Fig. 6b ). When the two sides are silenced, the network drifts to one of the fixed points randomly
Results, Robust model networks, paragraph 2; Figure 6b-d; Methods, Modeling and simulation
How it will be judged
A simulation with the parameters of the Methods (modular attractor, self weight 0.5235, mutual -0.5235, inter-module 0.3, tau 100 ms, the given g(x), input 0.1, tonic 0.5, noise variance 0.01; modular integrator with the printed 8 x 8 matrix, tau 10 ms, inputs +-0.04, tonic 40; modular FORCE network). Integration step 1 ms (a standard choice; for the noise, variance per step as printed, and per unit time as a second reading). Unilateral silencing holds one module at zero for 0.2 s < t < 1.0 s; bilateral silencing holds both. Reproduced when, for the attractor and integrator, the silenced module returns within 10% of its unperturbed state by t = 2 s in at least 90% of 100 noisy runs, and after bilateral silencing the attractor ends at the wrong fixed point in between 25% and 75% of runs (thresholds chosen here; "randomly" read as near one half). The FORCE network is reported, as its training is not fully specified.

Pipeline

Paper: waiting (not checked)
Claims: done
Claims check: done (ready)
Attempt: todo
Questions: none
Verified: none