Calcium fluorescence

A compressed neuron-development visualization, followed by a 3-D neuron with illustrative firing. GCaMP fluorescence reports each spike; OptiCeil determines the post-spike acquisition window required for a stated accuracy.

t = 0.000 s

—

Growing

Drag to rotate

SPIKE
—

Calcium fluorescence (what the microscope sees)True spikes (electrode)Kept and neededKept, not neededNeeded, not kept

Firing pattern

Speed

Legend

Bright dot at the center: cell body (soma). The neuron's core, where inputs are added up and a spike is triggered.
Thick line upward: trunk (apical dendrite). The main cable collecting input from far away.
Fine lines: dendrites and branches. Where the neuron receives signals from other cells.
Small glowing ends: growing tips. Branches extending during the growth time-lapse.
Grey thread: axon. Carries the spike to other neurons.
Brief orange flash and pulse: the spike (action potential). A millisecond electrical event. A microscope does not see this directly; it is drawn here so you know when it happened.
Green glow spreading out: calcium fluorescence. Calcium enters after the spike and GCaMP lights up. This slower glow is what the microscope actually records.
Faint drifting specks: decoration for depth only.

1 · Sensor and preparation

2 · Required fraction of attainable performance

A fraction of the best performance this recording can reach, as in the study.

3 · Post-spike acquisition window

Acquisition window—

Recoverability at your window (approx.)—

Accuracy at your window (approx.)—

4 · OptiCeil output

STOP_TEMPORAL_ACQUISITION

—

Recoverability retained (held-out)—

Balanced accuracy retained (held-out)—

Residual ambiguity at this stop—

CEILING_UNRESOLVED · OptiCeil compares acquisitions; it does not claim an absolute physical limit for this recording.

5 · Information arriving after each spike

Recoverability indexHeld-out balanced accuracy● measured · line interpolated

The marker follows the latest spike. Green: information still arriving. Grey: beyond what your target needs. Red: past peak accuracy, where longer imaging made the answer worse.

Longer acquisition, lower held-out accuracy

Background: GFP and GCaMP

1962Osamu Shimomura isolates green fluorescent protein (GFP) from the jellyfish Aequorea victoria.
2001Junichi Nakai and colleagues fuse a rearranged GFP to calmodulin and M13, creating GCaMP: it glows when calcium binds.
2008Nobel Prize in Chemistry to Shimomura, Martin Chalfie and Roger Tsien for the discovery and development of GFP.
2026OptiCeil: the minimum post-spike acquisition window required for a stated inference.

Method and provenance

Measured values (dots, stop points, retained fractions, residual ambiguity, accuracy drops) come from Ewing, "Information arrival sets the minimum acquisition time in calcium imaging," Research Square, 2026 (doi:10.21203/rs.3.rs-10898326/v1): 914 CASCADE segments, 312 neurons, neuron-disjoint discovery and held-out validation, 30 Hz (one volume ≈ 33 ms).
Curves between measured points are shape-preserving interpolations to show direction. The recoverability index after its peak is drawn declining in proportion to the reported accuracy drop; the study reports that it declined but not by how much. Early accuracy for the two GCaMP6s regimes follows the index shape.
The neuron, spike trains and traces are illustrative, not CASCADE recordings. Indicator rise and decay times are approximate, chosen to show that GCaMP6s is slower than GCaMP8m. "High-frequency" is an illustrative burst pattern.
Growth is a compressed time-lapse. In the developing brain, calcium signalling triggered by activity helps regulate dendritic growth and branching over days, alongside activity-independent cues such as neurotrophins (e.g. Wong & Ghosh, Nat Rev Neurosci 2002). The OptiCeil results concern spike detection, not growth.
Flashing is limited to at most three flashes per second, and motion is reduced when the device requests it.
The spinal regime is shown with lower signal-to-noise for illustration. The study found that the spinal label itself did not explain lower recoverability; a matched residual difference remained and is described there as observational.
The recoverability index is comparative, not an absolute physical bound, hence CEILING_UNRESOLVED.
History: Nobel Prize in Chemistry 2008 (nobelprize.org); GCaMP: Nakai, Ohkura & Imoto, 2001.