Neuroscope Science Dashboard Guide

Neuroscope is MoscaQuant's read-only visualization, replay, and scientific inspection layer. It combines hybrid MaleCNS geometry, frozen replay telemetry, preserved causal-path evidence, graph traversal, responder analysis, and MQ-5 intervention replay from completed experimental artifacts.

SAME NEURONS. DEEPER QUESTIONS.

Quick Start

ROTATE Drag on the main viewer.

PAN Shift + drag.

ZOOM Mouse wheel.

INSPECT Click a neuron.

REPLAY Scrub or press Play.

TRACE Select a causal node, then Trace Path.

Viewer Invariants

NEURONS 12,475 REAL SOMA 7,486 FALLBACK 4,989 CAUSAL EDGES 13 RESPONDERS 9 REPLAY FRAMES 192

Unexpected changes to these values should be treated as a build or data issue until explained.

Color Cheat Sheet

BLUE — real soma geometry

GRAY — topology fallback

ORANGE — causal responder / causal overlay

WHITE — spike or selection emphasis

CYAN — effective activity / upstream trace

GREEN — downstream trace

Geometry

Real soma geometry

Nodes marked somaLocation use real MaleCNS soma coordinates. These are anatomical data.

Topology fallback

Nodes marked topologyFallback lack usable soma geometry in the frozen Neuroscope geometry artifact. Their displayed positions preserve useful topology only.

Topology fallback coordinates are not anatomical coordinates and must not be interpreted as anatomical evidence.

Display Controls

Real soma geometry — show/hide real anatomical soma positions.
Topology fallback — show/hide non-anatomical fallback placement.
Causal overlay — show/hide the frozen MQ-3.2 causal graph.
Responders only — restrict display to responder neurons.

Role Filters

Retina 3,241 Relay 2,430 Graded 5,490 Descending neurons 1,314 Responders 9

A neuron can belong to more than one role. Filtering affects presentation only.

Replay

Neuroscope contains 192 recorded replay frames, numbered 0 through 191.

Previous frame
Play / Pause
Next frame
2, 4, 8, 12, or 24 fps playback
Direct timeline scrubbing

Manual timeline movement automatically pauses playback.

Frame summary

Retinal spikes
Relay spikes
Graded-active count
DN mean activity
DN maximum activity
DN spikes
Maximum responder value

Replay is recorded telemetry. Pressing Play does not execute a new simulation.

Activity Visualization

Voltage magnitude increases node brightness and size.
Effective activity receives cyan emphasis.
Spikes receive bright white emphasis.
Responders remain amber regardless of momentary activity.
Responder magnitude can increase responder size.

Neuron Inspector

Click a neuron to inspect:

Model index
MaleCNS body ID
Roles
Graded type
DN cluster
Soma side
Soma neuromere
Geometry source
Causal-node status
Responder status

Causal Inspection

Causal nodes expose their preserved incoming and outgoing connections.

IN · model 43417 weight 6.6445e-4 frame 145

Edge direction, weight, and observed frame originate from the frozen MQ-3.2 causal artifact.

Causal Traversal

Select a causal node and press TRACE PATH.

Cyan — upstream
Green — downstream
Trace root — selected origin
Unrelated anatomy — dimmed

The trace panel reports node count, edge count, and maximum graph depth.

Verified example

56393 → 68045 → 1273

Trace model 1273 to reconstruct the full three-node chain.

Responder Ensemble

92 656 317 137122 126002 55 129 51 1273

The Responder Ensemble panel displays each responder's current value, peak value, peak frame, and preserved causal-input count.

Use ISOLATE 9 RESPONDERS to hide the rest of the population.

Recorded peak frames

92 F191 656 F165 317 F187 137122 F183 126002 F191 55 F191 129 F191 51 F191 1273 F191

Synchronized Responder Plot

The compact responder graph shows all nine recorded responder series across frames 0–191.

White line — current frame
Colored lines — individual responder series
Selected responder — thicker trace
Click the graph — jump to that frame

Guided Exercises

Exercise 1 — Responder 656

Find model 656 in the responder panel and click F165.

Inspect its incoming causal edge from model 43417 and compare the edge observation at frame 145 with the later response peak at 165.

Exercise 2 — Multi-hop path

Navigate to model 1273 and select TRACE PATH.

Expected:

TRACE ROOT 1273 NODES 3 EDGES 2 MAX DEPTH 2 UP model 68045 depth 1 UP model 56393 depth 2

Exercise 3 — Late responder ensemble

Jump to frame 191, isolate the nine responders, and compare their synchronized activity traces.

MQ-5 Intervention Replay

The INTERVENTION REPLAY panel exposes completed MQ-5 intervention evidence inside Neuroscope.

It reads the frozen downstream artifact mq5-interventions-v1.json. Neuroscope does not execute the intervention runtime and does not generate new experimental outcomes.

Generalized single-edge replay

The generalized family exposes all 13 frozen causal edges and the recorded 25%, 50%, 75%, and 100% attenuation conditions.

Where available, the dashboard also displays the frozen timing and matched-control evidence. A missing control remains unavailable; the viewer does not construct a replacement.

Two-hop pathway replay

56393 → 68045 → 1273 upstream causal frame F146 intermediate causal frame F147

The pathway view can compare completed upstream and intermediate perturbations against their recorded baseline traces.

Convergence replay

Neuroscope exposes the three frozen convergence systems:

44274 + 55925 → 55 55548 + 87441 → 51 92657 + 93484 → 129

A-only, B-only, and combined conditions are descriptive replays of completed MQ-5 results.

Combined effect greater than either single input is not, by itself, a statistical synergy claim. Neuroscope does not assign synergy, antagonism, subadditivity, or superadditivity.

Responder 51 effects occur at an extremely small numerical scale. Their magnitude must be interpreted cautiously.

Robustness and provenance

Provenance indicators may report completed CONF-004A normalization robustness and MQ-5.6 independent-runtime replication evidence. These are existing experimental results, not evidence generated by the dashboard.

Trace interpretation

Baseline traces are frozen recorded telemetry.
Intervention traces are frozen recorded telemetry.
Dose selection changes which completed result is displayed.
Jumping to a causal frame changes viewer position only.
Missing telemetry is displayed as unavailable rather than reconstructed.

INTERVENTION REPLAY IS READ-ONLY. IT DOES NOT RUN MOSCAQUANT.

Scientific Boundaries

Neuroscope is a read-only interpretation layer over frozen MoscaQuant experimental telemetry.

No synthetic fallback is promoted to anatomy.
Replay does not rerun the experiment.
Intervention Replay does not execute interventions.
Traversal does not create new causal edges.
Visualization does not strengthen the underlying evidence.
Responder magnitude does not establish causal importance.
Filtering never changes the underlying dataset.
Biological causality is not claimed.

Financial semantics

NOT ASSIGNED

Troubleshooting

Replay artifact missing

source .venv/bin/activate python -m neuroscope.build_web_replay

Old interface appears

Hard-refresh the browser or add a cache-busting query parameter:

http://127.0.0.1:8080/?fresh=1

Check JavaScript

node --check neuroscope/web/app.js git diff --check

Check what the web server is serving

curl -s http://127.0.0.1:8080/app.js | grep -n "RESPONDER ENSEMBLE"

Recommended Operator Workflow

1. Inspect anatomy 2. Adjust role filters 3. Select a neuron 4. Inspect metadata 5. Scrub replay 6. Inspect local causal evidence 7. Trace causal graph 8. Compare responder ensemble 9. Open Intervention Replay 10. Select family / experiment / dose 11. Compare baseline and intervention traces 12. Inspect provenance and limitations 13. Return to full anatomy

Principle

Visualization may make the evidence easier to understand, but it must not silently change what the evidence means.