Review Papers Claim by Claim

Summary judgement

The preprint provides converging anatomical, electrophysiological and behavioural evidence that a spinal excitatory interneuron class (dI3) receives multimodal sensory inputs and efference copies (notably from Renshaw cells and motor neurons), projects to motor pools (flexor biased, some divergent to antagonists), and is necessary for fast corrective reactions during locomotion — supporting the claim that dI3 neurons can act as spinal comparator modules for rapid corrections of movement (dI3 comparator preprint

Detailed, critical review and synthesis

1) What the paper claims (concise)

The authors propose that a population of spinal excitatory interneurons — dI3 neurons — form cellular comparator modules that integrate instructive sensory feedback (proprioceptive and cutaneous) with predictive efference copies (notably via Renshaw cells and direct motor neuron collaterals) to generate rapid online corrective commands to motor neurons during gait and perturbations (dI3 comparator preprint

2) Strengths

• Multimodal convergence: The paper combines high-resolution anatomy, monosynaptic rabies premotor tracing, patch-clamp physiology (DR and VR stimulations), and behaviour to connect circuit anatomy to function — strengthening causal inference (multimodal evidence
• Convergence at single-cell level: The combination of DR and VR stimulations in the same dI3 neurons (with paired analyses) provides direct evidence that individual cells can receive both instructive and predictive signals — crucial for comparator logic (single cell convergence
• Behavioural causality: Reversible chemogenetic silencing (hM4D(Gi) with JHU37160) and acute pharmacology (mecamylamine) affecting performance in ladder/beam and stumbling paradigms tie the circuitry to ecologically relevant rapid corrections (behavioural causality
• Careful nested statistics: The authors computed ICCs and used hierarchical bootstrapping where appropriate, reducing the risk of pseudo-replication common in slice/cell datasets (statistical rigor

3) Key limitations, weaknesses, and alternative interpretations

1. Specificity of the Renshaw manipulation: Mecamylamine blocks nicotinic receptors broadly and may affect cholinergic transmission beyond motor neuron->Renshaw synapses (including presynaptic modulation of sensory terminals) — the authors acknowledge possible off-targets (mecamylamine caveat
2. Tracing sensitivity and underestimation of convergence: Monosynaptic ΔG-rabies tracing is known to have low/transmission-limited efficiency — the fraction of premotor dI3 cells receiving proprioceptive inputs or being premotor may be underestimated, so quantitative claims about proportions should be interpreted cautiously (rabies efficiency
3. Species and developmental stages: Many in vitro electrophysiology datasets use neonatal/juvenile spinal cords (P5-P15) while behavioural manipulations use adults; maturation changes in synaptic latencies and connectivity could alter the functional interpretation in adult animals and in other species including humans (age differences
4. Flexor bias and functional specificity: The paper documents a flexor bias in dI3 outputs, but does not fully dissect whether comparator submodules are organized by effector (flexor versus extensor) or phase of gait; without cell-type-specific projection mapping of functional subpopulations, the exact computations remain partly underspecified (flexor bias caveat
5. Alternative predictive inputs: Renshaw cells are a plausible source of efference copy but descending inputs (corticospinal, reticulospinal) or other spinal premotor interneurons could provide predictive information; the paper shows Renshaw involvement but cannot exclude parallel predictive channels (other predictive inputs

4) How convincing is the comparator interpretation?

Comparator logic requires (a) availability of a predictive signal (efference copy), (b) availability of an instructive signal (sensory feedback), (c) their convergence on the same processing units, (d) an output that signals mismatch and produces corrective action. The paper addresses each: (a) Renshaw and motor neuron collaterals supply a predictive channel; (b) proprioceptive and cutaneous afferents supply instructive feedback with subcellular segregation; (c) DR and VR stimulations evoked opposing conductances in the same dI3 neurons; (d) silencing dI3 or disrupting Renshaw signaling degrades rapid corrections. Taken together, the evidence supports the comparator hypothesis at the level of spinal microcircuits, with the caveat that the precise computational operation (subtraction, gating, gain modulation, rebound-triggered excitation) remains mechanistically underspecified and likely heterogeneous across dI3 subpopulations (comparator assessment

5) Reproducibility and transparency

Methods are described in substantial detail: transgenic lines, AAV constructs, ΔG-RV rabies protocols, electrophysiology solutions, behavioral protocols, and statistical pipelines are given. That supports reproducibility; remaining barriers are specialized transgenic lines and technical expertise for longitudinal spinal cord electrophysiology and neonatal rabies injections. The paper declares funding and a conflict of interest (cofounder role for an author), which is properly disclosed (methods and COI

6) Key experiments that would strengthen or falsify the model

• Cell-type specific optogenetic activation/inhibition of Renshaw->dI3 synapses during behavior (rather than global mecamylamine) to show causality specifically via that predictive channel.
• Single-cell in vivo recordings of identified dI3 neurons during perturbations in behaving animals to show predictive/instructive signal timing and error-related firing correlated with corrective output.
• Targeted projection-specific manipulation of flexor vs extensor-projecting dI3 subpopulations (intersectional genetics) to test domain-specific comparator computations during phase-dependent perturbations.

7) Practical/clinical implications and broader impact

If spinal comparator modules exist and can be modulated, they offer a proximate target for interventions to improve rapid corrective reflexes after spinal cord injury or in gait disorders — for example, spatiotemporal neuromodulation (tSCS) that enhances predictive integration or targeted stimulation of dI3-like networks to improve stumble recovery (speculative; translational gap remains large) (translational note

8) Concise recommendations for readers

1. Interpret proportion estimates (e.g., % premotor dI3 with proprioceptive inputs) as lower bounds due to tracing efficiency limits.
2. Value the multi-method convergence: anatomy + physiology + behavior provides stronger evidence than any single technique.
3. Follow-up should prioritize cell-type- and synapse-specific causal tests, and in vivo single-cell recordings during behavior to nail timing/algorithmic computations.

9) Short, actionable critique summary

High-quality, multi-level work that credibly demonstrates dI3 neurons integrate sensory and efference-copy signals and are necessary for fast corrections; next steps should be specific manipulation of the Renshaw->dI3 synapse, in vivo identified single-unit recording during perturbations, and projection-specific functional dissection to confirm computational details and generality beyond mice (final synthesis

10) Reproduced visual summary (simple quantitative visualization)

References used (primary)

All claims in this review are tied to the primary preprint:

• dI3 comparator preprint

Interactive next steps

If you want: (a) run a causal power analysis to estimate the sample sizes needed for an Renshaw-cell specific optogenetic manipulation; (b) get a detailed protocol for intersectional viral targeting of flexor vs extensor-projecting dI3 subpopulations; (c) generate synthetic spike timing predictions for in vivo dI3 activity under comparator models — click below to run the BGPT biology agent to iteratively compute and produce protocols/data analyses.

Author review links

Open bespoke author reviews: