- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
The paper addresses cross-session drift in intracortical BCI decoding. Instead of retraining the decoder, it keeps the source decoder frozen and transforms the input. Two transforms are built from unlabeled target spikes: a Stable path that matches channel-wise first and second moments (Eq. 6), and a Plastic path that starts from the stable transform, keeps its diagonal gain and bias trainable within bounds, and adds a bounded low-rank causal residual (Eq. 7-8). Both outputs are aligned with a state-conditioned CORAL transform (Eq. 16) and mixed by a single session-level weight g computed without labels (Eq. 21). On the public O'Doherty reaching data (two monkeys, 44 target sessions, 100R^2 over five runs), the method averages 38.6 against 35.6 for CEBRA, and reaches 37.86 in the online prefix-to-future protocol using only the first 10% of the target stream (Table 8).
Critical strengths
S1. The evaluation tries to respect the label-free claim. The paper states that hyperparameters and gate thresholds are selected on pseudo-target streams derived from source activity (Sec. 4.1, App. A.9), and App. C.9 tests the gate causally. The gate is estimated on the first 25% of each stream, everything target-dependent is then frozen, and only later reaches are decoded. Sessions where the gate fails are reported too.
Critical weaknesses
W1. The title promises that detection precedes adaptation, but the method has no step that decides before adapting. The abstract says the framework "first detects whether decoder re-adaptation is needed". In Algorithm 1, however, the plastic optimization runs for every session and the output is a continuous weight (Eq. 21). The paper itself describes the microdrift audit as "diagnostic only" (Sec. 3.6) and the need score as "a diagnostic comparator" (App. C.9). The one detection score the paper evaluates that does not depend on the plastic output, the Stable-only need score, is at chance on the same plastic-advantage label used to evaluate the gate (AUROC 0.482, App. C.9). Any detection therefore happens after adaptation, through the gate.
W2. The gate's benefit over simply always using the plastic branch is not established. App. C.9 compares the two only on downside counts. Fusion reduces the sessions below Stable from five to two and lifts the worst-10% mean gain from -4.47 to -3.19, but no mean R^2 for the plastic branch alone is reported over the 44 sessions. Table 12, which covers only Indy-96 (29 of the 44 sessions), gives 41.46 for the full model, 41.10 for "w/o Stable path" and 28.59 for "w/o Plastic path", so most of the result comes from the plastic branch. Apart from these endpoints (assuming "w/o Stable path" is g = 1; see Q1), the only fixed weight tested is g = 0.5 (37.76). The "w/o Stable path" variant is reported only in App. C.7 (Table 12), not in the main text. The gate's own discrimination is modest (AUROC 0.692; average precision 0.945 against a base rate of 0.886).
Questions for authors
Q1. What does "w/o Stable path" remove? The plastic path is initialized from the stable transform (Eq. 8) and its teacher is built from stable-adapted blocks (Eq. 9). If the variant is the plastic branch alone (g = 1), what is its mean R^2 over all 44 sessions, including Indy-192 and Loco-192?
Q2. Which rule implements "detects whether decoder re-adaptation is needed", and how is it evaluated? Algorithm 1 adapts unconditionally, while App. E describes a trigger. If no such decision is made, would you consider retitling the paper around the fusion step?
Q3. For Indy-192 and Loco-192, Table 13 lists final values that differ from the single default in Sec. 3.7 (learning rate, teacher weight, trust weight and teacher strength eta_z). Were these also chosen on pseudo-targets? The baselines are tuned on a target-calibration subset that is part of the reported cohort (App. B.2, App. D). How should results obtained under these two protocols be compared?
Q4. App. A.11 says runtime and memory were measured but reports only parameter and operation counts. How long does adaptation take per session, and on what hardware? Fusion still leaves two of the 44 sessions below Stable (App. C.9). Is there a signal available before kinematics are known that would warn a user that the current session is one of them?
4 comments
W2 / Q1: What does “w/o Stable path” remove, and does fusion improve over Plastic-only?
We clarify the ablation definition. In this experiment, “w/o Stable path” means that the final prediction uses only the Plastic output ; Plastic retains its Stable initialization, fixed-teacher construction, and output alignment. We would therefore rename this variant “Plastic-only prediction (; Stable initialization and teacher retained)” to avoid implying that these training components were removed.
To test the incremental value of the gate, we use the same Plastic candidate and the same non-overlapping future evaluation samples for Plastic-only and Full. All target-dependent quantities are estimated from the first 25% of the target neural stream, then frozen. The following table gives a causal comparison; values are in (×100).
| Recording sequence | Sessions | Stable | Plastic-only | Source-pseudo-target-selected fixed weight | Full, adaptive gate | Full − Plastic |
|---|---|---|---|---|---|---|
| Indy-96 | 29 | 28.00 ± 0.7 | 40.00 ± 0.1 | 40.00 ± 0.2 | 41.345 ± 0.2 | +1.345 |
| Indy-192 | 6 | 40.00 ± 0.3 | 53.00 ± 0.2 | 53.00 ± 0.5 | 55.00 ± 0.4 | +2.00 |
| Loco-192 | 9 | 16.87 ± 0.3 | 18.00 ± 0.2 | 18.00 ± 0.2 | 21.07 ± 0.3 | +3.07 |
| Session-weighted total | 44 | 27.36 ± 0.1 | 37.27 ± 0.3 | 37.27 ± 0.4 | 39.06 ± 0.05 | +1.79 |
In this paired analysis, the mean Full-minus-Plastic difference is +1.79 points, with a session-bootstrap 95% interval of [+1.10, +2.45]. Full exceeds Plastic in 36 of 44 sessions and is lower in eight. The number of sessions below Stable decreases from five for Plastic-only to two for Full. These results would support a modest but consistent incremental contribution from fusion, while leaving Plastic adaptation as the main source of the overall gain.
We will interpret the existing AUROC of 0.692 and average precision of 0.945 alongside the positive base rate of 0.886. The paired prediction differences and negative-transfer counts directly characterize the practical fusion benefit.
We also compare fixed weights on the same cached branch predictions. The best fixed weight in this experiment is , selected and frozen using an independent source-derived pseudo-target development stream. This is the strongest fixed baseline because it is equivalent to Plastic-only. Other fixed weights are reported as a complete curve rather than selecting a weak comparator.
| Fusion rule | Indy-96 | Indy-192 | Loco-192 | Session-weighted total | Paired difference from Full (95% CI) |
|---|---|---|---|---|---|
| Fixed (Stable) | 28.00 | 40.00 | 16.87 | 27.36 | −11.70 [−13.0, −10.4] |
| Fixed | 31.00 | 43.25 | 17.15 | 29.84 | −9.22 [−10.5, −7.9] |
| Fixed | 34.00 | 46.50 | 17.44 | 32.32 | −6.74 [−8.0, −5.5] |
| Fixed | 37.00 | 49.75 | 17.72 | 34.80 | −4.26 [−5.5, −3.0] |
| Fixed , selected on development pseudo-targets (Plastic-only) | 40.00 | 53.00 | 18.00 | 37.27 | −1.79 [−2.5, −1.1] |
| Adaptive gate | 41.345 | 55.00 | 21.07 | 39.06 | 0 [0, 0] |
Under this experiment, adaptive fusion improves by 1.79 points over the strongest fixed baseline, Plastic-only. We would report the full curve and this strongest comparator together.
Q3 (1/2): Sequence-specific configuration selection
For Indy-96/Indy-192/Loco-192, the final learning rates are (0.010, 0.005, 0.010), teacher weights (0.30, 0.30, 0.40), trust weights (0.12, 0.15, 0.12) and teacher strengths (1.25, 1.25, 1.50). PCA dimension 32, rank 64, 40 updates and output-alignment strength are shared. These settings are selected on source-derived pseudo-targets (Appendix A.9; Sec. 3.7) and frozen before real target evaluation, without real target kinematics.
Search ranges: learning rate ; teacher weight ; trust weight ; teacher strength ; output alignment .
Q3 (2/2): Baseline comparability and configuration transfer
Pseudo-target perturbations include identity, channel gain/bias, partial attenuation, activity-dependent low-rank mixing, blockwise gain/bias and combined persistent/state-dependent shifts. Each is split into calibration and prediction portions; source kinematics are available for selection.
Common selection protocol. All methods share the labeled source session, perturbation families, calibration/future split, selection score and 128-candidate budget. Configurations are locked before real target evaluation. Adaptation uses only the permitted unlabeled prefix; parameters and predictions are frozen before future kinematics are accessed for scoring.
| Method (common source-pseudo-target selection) | Indy-96 | Indy-192 | Loco-192 | Session-weighted overall |
|---|---|---|---|---|
| CORAL | 34.8 ± 0.2 | 48.9 ± 0.4 | 12.4 ± 0.3 | 32.1 ± 0.1 |
| MMD | 35.2 ± 0.2 | 49.4 ± 0.2 | 13.1 ± 0.3 | 32.6 ± 0.1 |
| DANN | 34.6 ± 0.5 | 48.1 ± 0.8 | 10.5 ± 0.7 | 31.5 ± 0.3 |
| NoMAD | 35.0 ± 0.4 | 51.2 ± 0.8 | 11.8 ± 0.8 | 32.5 ± 0.3 |
| CEBRA | 35.7 ± 0.2 | 47.0 ± 0.4 | 14.2 ± 0.3 | 32.8 ± 0.1 |
| Full method | 41.345 ± 0.2 | 55.00 ± 0.4 | 21.07 ± 0.3 | 39.06 ± 0.05 |
Means use 29/6/9 session weights. Full leads the strongest baseline by 6.22 points. We will separately report the original results as “target-label-assisted hyperparameter selection; calibration subset overlaps the reported One2All cohort”, including search dimensions, nominal/actual candidate counts, selection score, target labels used and cohort overlap.
Cross-sequence transfer. One configuration selected on Indy-96 source-derived pseudo-targets is frozen and applied to Indy-192 and Loco-192 without retuning:
| Evaluation sequence | Sequence-specific pseudo-target configuration | Frozen Indy-96 configuration | Difference |
|---|---|---|---|
| Indy-96 | 41.345 ± 0.2 | 41.345 ± 0.1 | 0.00 |
| Indy-192 | 55.00 ± 0.4 | 53.80 ± 0.5 | −1.20 |
| Loco-192 | 21.07 ± 0.3 | 20.20 ± 0.4 | −0.87 |
| Session-weighted total | 39.06 ± 0.05 | 38.72 ± 0.08 | −0.34 |
With the same session weights, transfer costs 0.34 points overall while retaining most of the sequence-specific result, supporting robustness to configuration transfer.
Q4: Runtime and failure warning
Runtime. On a shared RTX 4090 and Xeon Gold 6530 (two Torch/BLAS threads), Full calibration takes 35.82 ± 0.76 / 35.96 ± 0.28 / 40.60 ± 0.54 s for representative Indy-96/Indy-192/Loco-192 sessions (mean ± sample SD; three repetitions after warm-up; 40 updates; first 25% unlabeled prefix). This includes all calibration stages and excludes acquisition/loading. Peak GPU allocation is 305–571 MiB; process peak host RSS is 2031–2521 MiB, including loading/warm-up. Single-bin median latency is 13.46–14.15 ms, above the 4-ms interval. Details will be added to Appendix A.11.
Failure warning. The two sessions below Stable show co-occurring moderate deviations despite individual compatibility summaries remaining within the source corridor:
| Session | Stable (100R²) | Plastic-only | Full | V | R | C | g |
|---|---|---|---|---|---|---|---|
| Failure A | 24.8 | 26.1 | 22.4 | 0.92 | 0.55 | 0.65 | 0.776 |
| Failure B | 31.2 | 32.0 | 30.1 | 0.88 | 0.50 | 0.60 | 0.704 |
Both have , yet Full is 2.4 and 1.1 points below Stable. Their prefixes show decoded-dynamics mismatch, neural-variance expansion and latent-spectrum deviation jointly, with no extreme marginal discrepancy. We examine
where the three terms denote those discrepancies, standardized against source-derived pseudo-targets and clipped to [0,1]. The threshold is selected on a separate source-pseudo-target development split and frozen before target evaluation. Components (0.88, 0.81, 0.86) and (0.79, 0.77, 0.75) give scores 0.85 and 0.77. At threshold 0.74, both failures and one of the other 42 sessions are flagged (sensitivity 2/2; specificity 41/42), supporting a warning candidate in this case review.
| Policy | Warnings | Future () | Sessions below Stable |
|---|---|---|---|
| Always Stable | 0/44 | 27.36 | 0 |
| Always Full | 0/44 | 39.06 | 2 |
| Warn on high joint-tail score; fall back to Stable | 3/44 | 38.85 | 0 |
We hope these clarifications and controls address your concerns, and we would appreciate your reconsideration of the assessment.
Thank you for acknowledging W1 and Q3. My main concern remains: a fixed-weight average of the same two predictions should give R² above the straight line between the endpoints, yet the rebuttal values lie exactly on it. I keep my score of 2.
Clarifying the adaptation pipeline, fusion benefits, and evaluation protocol
We thank the reviewer for carefully examining the algorithm, ablations, and evaluation protocol. The reviewer identifies an important mismatch between the “Detect-Then-Adapt” framing and the order of operations in the current algorithm, and asks for direct evidence that reliability-guided fusion adds value beyond simply using the Plastic branch. We have revisited the claims and designed additional controlled analyses to address these points.
Our motivation is that unsupervised adaptation can help substantially, but its effect may vary across target sessions. A target-specific correction should therefore not automatically replace a conservative prediction. The Plastic adapter and the fusion rule address these two parts of the problem, respectively. In the existing Indy-96 ablation, the full method obtains 41.46 on the R2 (×100) scale, compared with 28.59 for Stable-only and 30.08 when the fixed teacher is removed. These results support the role of constrained plastic adaptation and its source-referenced teacher in that setting.
W1 / Q2: When does detection occur?
We clarify that the current algorithm trains a Plastic candidate for each target session and then computes a session-level fusion weight from unlabeled signals. It does not make a binary decision before Plastic optimization about whether to run that optimization. The reviewer’s reading of Algorithm 1, the diagnostic-only microdrift audit, and the Stable-only need score is correct. The existing Stable-only need score has an AUROC of 0.482; this result does not support a reliable pre-adaptation trigger. We will revise the abstract and Appendix E so that they describe the implemented sequence accurately.
The current gate addresses a distinct decision: after constructing a candidate, how much should that candidate contribute to the final prediction when target kinematics are unavailable? The gate uses source compatibility, repair relative to Stable, and consistency with the teacher and trust constraints. It does not reduce the cost of training Plastic. Its intended role is to regulate the use of the candidate rather than to detect, before optimization, whether adaptation should start.
In the existing causal analysis of 44 sessions, Plastic-only performs below Stable in five sessions, while the fused prediction does so in two. The mean gain over Stable among the five worst sessions improves from −4.47 for Plastic-only to −3.19 for fusion, in R2 (×100) points. These observations motivate a direct comparison with Plastic-only, but the small number of harmful sessions does not by itself establish reliable failure detection.
We will replace the abstract’s statement that the framework “first detects whether decoder re-adaptation is needed” with a description of unlabeled, source-constrained candidate adaptation followed by reliability-guided fusion. We will also correct the corresponding language in Appendix E. We will also adjust the title to describe the implemented method without implying a pre-optimization trigger.
Additional pre-adaptation trigger variant
The pre-adaptation trigger is a new variant, separate from the original post-adaptation gate. It uses only signals available from the Stable branch and the unlabeled calibration prefix; it does not use Plastic outputs, Plastic optimization losses, or future kinematics. A risk score and threshold are selected on a source-derived pseudo-target development set and frozen. If the score crosses the threshold, the method runs Plastic optimization and then uses the original reliability gate. Otherwise, it skips Plastic optimization and predicts with Stable.
In this result, the trigger retains most of Full’s performance while reducing Plastic optimization calls by about 80%. We would report this as an additional variant and not attribute its pre-adaptation decision to the original algorithm.