Review papers with raw data transparency

• Model and approach: A robust PDMS 3D Neuron-on-Chip with Geltrex-embedded, hPSC-derived prefrontal cortical neurons matured ~2 weeks then injured with a focal 9 mJ weight-drop (6 g, 15 cm). Calcium imaging (Fluo-4 AM) with 13 single-cell and network metrics + secretome (48-plex/10-plex/4-plex) and biochemical endpoints provide multimodal longitudinal readouts Neuron-on-Chip device and protocols
• Key reproducible findings: biphasic functional trajectory — early (0.5–72 h) hyper-synchrony, prolonged rise/fall times, elevated wND/wPL, calpain-1 autolysis and caspase-3 activation; late (5–8 d) depolarization, loss of long-range connectivity, increased modularity/NoC, extracellular pT181/tTau elevations and intracellular AT8+/NFT+ Tau by day 8 Biphasic functional and proteopathy timeline
• Strengths: human-derived neurons in 3D — more physiological than 2D; multimodal longitudinal endpoints (functional → biochemical → morphological); clearly described device and injury parameters; use of graph-theory metrics and sCCA to link secretome to functional states; sample replication across devices/timepoints stated Methods and analysis pipeline
• Main limitations & blindspots (critical):
  1. Neuron-only system: purposely excludes astrocytes/microglia/endothelium — so claims about in vivo neuroinflammation causality must be qualified (authors acknowledge this). The neuronal secretome may differ markedly when glia are present Neuron-only limitation noted by authors
  2. Longitudinal tracking of identical cells was not performed for all endpoints — calcium recordings and end-point assays came from separate devices/timepoints, which complicates single-cell causality between activity patterns and biochemical changes.
  3. Secretome measurements were from pooled replicates using commercial multiplex panels; detection limits, normalization strategy, and raw values are not publicly provided in the preprint (data/code/CAD are available on request), limiting independent re-analysis.
  4. Sample sizes vary by assay and are modest (3–5 devices/timepoint; 2–3 secretome experiments); statistical power for less-robust changes (e.g., non-significant calpain 32 kDa fragment trend) may be limited.
  5. Weight-drop onto PDMS/hydrogel approximates focal impact but does not recapitulate complex in vivo strain/rate/pressure fields—translational mapping to human TBI biomechanics requires caution.

1. Device CADs: authors state CAD files available on request — release of CAD + PDMS curing logs would allow replication of mechanical properties (elastic modulus depends on curing/ratios) Data availability statement
2. Cell source/differentiation: differentiation recipe described; reproducibility across iPSC lines and passages is unknown — recommend reporting donor lines, karyotype, and batch QC metrics (e.g., synaptophysin expression quantification per device).
3. Calcium imaging pipeline: authors used Pnevmatikakis deconvolution and defined activity thresholds (≥5 spikes/min) — publishing MATLAB code and example raw TIFFs would permit reanalysis and spike-estimation benchmarking.
4. Secretome panel: raw concentration values, LOD/LOQ, and normalization (per-device cell number or total protein) should be released for meta-analysis.
5. Statistical methods: PERMANOVA/t-SNE used appropriately for multivariate separation; provide full p-values, permutation parameters, and effect sizes in supplementary data.

• Perform the same injury in neuron-only cultures from independent iPSC donors and fail to reproduce the biphasic calcium/network dynamics or the extracellular Tau rise — that would argue against generalizability.
• Co-culture with astrocytes/microglia: if inclusion of glia prevents Tau secretion/aggregation under identical injury conditions, the neuron-autonomous claim would be weakened.
• Longitudinal single-cell tracking combining live Ca2+ imaging + later ICC on the same cells that do NOT show the predicted link between early hyper-synchrony and later AT8+/NFT+ accumulation would undermine proposed causal chain.

1. Single-cell longitudinal assay: Combine genetically encoded calcium indicator (GCaMP) + phototagging to record activity at acute timepoints, then fix the same device for ICC to directly link early hyperactivity at single-cell level to later AT8/NFT positivity. This tests causality (feasible and decisive).
2. Glia reconstitution series: Stepwise addition of human astrocytes, microglia, and endothelial cells to test whether neuronal secretome and Tau aggregation are augmented, unchanged, or suppressed — distinguishes neuron-autonomous vs non-autonomous mechanisms.
3. Protease inhibition time-window: Treat cultures with calpain inhibitor (e.g., calpeptin) immediately after injury vs delayed (24 h) to map when protease activity is necessary for extracellular Tau release and intracellular NFT formation.
4. Biomechanics sensitivity: Use device with calibrated strain/strain-rate readouts (or finite-element modeling of PDMS deformation) and test whether rate vs energy matters for the observed cascade (addresses translational relevance to different TBI types).
5. Secretome perturbation: Neutralize IP-10/IL-10/IFNα2 in conditioned media (antibodies) or block their neuronal receptors to test whether early secreted cytokines mediate late network remodeling/tauopathy.

• Use the device to screen interventions aimed at preventing early excitotoxicity (e.g., NMDA antagonists, calpain inhibitors) and test downstream impact on Tau release/aggregation.
• Report full raw calcium traces, spike inference parameters, and secretome raw concentrations to enable independent meta-analyses and standardization across labs.
• Cross-validate findings in at least two independent iPSC donor lines (and report donor metadata) before extrapolating to population-level biology.

The study delivers a compelling, human-relevant in vitro platform and convergent functional/biochemical evidence that neuron-intrinsic responses to focal mechanical insult can produce an early excitotoxic phase followed by network decomposition and tauopathy-like features. Strengths are the integration of network metrics with biochemical secretome profiling and morphological Tau readouts. Key caveats remain: neuron-only culture limits in vivo generalizability; some endpoints use pooled/limited replicates and longitudinal single-cell causality is not yet established. Reproducibility will hinge on public release of CADs, raw calcium movies, and secretome data; the platform is a strong tool for mechanistic and screening studies when used with the proposed follow-up validations Authors