F-I curve¶
The firing frequency a neuron produces across applied current levels under a stated protocol and rate convention.
Core Idea¶
An F-I curve shows how the firing frequency of a specified neuron or neuron model changes as applied current, I, is varied.[ref-94b4c43713a0][ref-eb9e48e68791] Its points depend on an explicit current protocol and a rule for calculating rate from spikes, such as a finite observation window or a model steady-state convention. Different neurons and conditions can produce different shapes. No one threshold, slope, or sigmoid is part of the definition.
Scope of Application¶
Use the name when current is varied for an individual neuronal bearer and firing frequency is paired with each tested level.
- Cortical-cell experiments: injected pulses and spike-train rates describe measured excitability.[^ref-94b4c43713a0]
- Neuron models: simulated current steps yield rates under stated model parameters.[^ref-eb9e48e68791]
- Cell comparisons: compatible protocols reveal different onset behavior across cell types.
- Conductance studies: compare F-I curves as channel settings change; conductance is a condition, not the current axis.[^ref-eb9e48e68791]
- Protocol comparisons: retain pulse duration, rate window, and step direction when interpreting differences.
Tateno distinguishes a conductance-input f-g plot from f-I.[^ref-94b4c43713a0] A population-only rate function or isolated threshold also needs its own identity test.
Clarity¶
Ask whose firing, which applied current, and which rate convention? Tateno and colleagues distinguished initial from later firing estimates under 600-ms current pulses.[^ref-94b4c43713a0] A curve reported simply as “firing rate versus input” hides that choice. Drion and colleagues report model steady-state rates, a different convention.[^ref-eb9e48e68791] Specifying it prevents a change of method from being mistaken for a change of neuron.
Manages Complexity¶
A curve compresses many spike trains or voltage trajectories into current-rate pairs. This helps compare excitability, onset and gain while discarding exact spike timing and subthreshold behavior. Record the preparation or model, current schedule and rate estimator beside the curve. In Drion's model, upward and downward current steps can expose hysteresis; one sweep can hide the difference.[^ref-eb9e48e68791]
Abstract Reasoning¶
Fix the bearer and protocol; vary current; calculate rate consistently; then compare curves. A new shape still belongs to the F-I class. If the input becomes conductance, the output ceases to be neuronal firing frequency, or only one current level is known, the claimed relation has changed or is incomplete. Check method and conditions before attributing a curve difference to a biological mechanism.
Knowledge Transfer¶
The same current-to-rate test works for measured cells and specified biophysical neuron models. Tateno's cortical cells showed different near-threshold onset types; Drion's Connor–Stevens model changed onset class as A-type potassium conductance changed.[ref-94b4c43713a0][ref-eb9e48e68791] Those are distinct realizations of one relation, not evidence that their mechanisms or timescales match. Beyond single-neuron current responses, an input-output plot is only an analogy unless the neuronal carrier, applied-current axis, firing-rate convention and protocol boundary remain literal. Measured cells and simulations provide different kinds of evidence; a model curve should not be described as another laboratory observation.
[^ref-94b4c43713a0]: T. Tateno, A. Harsch and H. P. C. Robinson, “Threshold Firing Frequency–Current Relationships of Neurons in Rat Somatosensory Cortex: Type 1 and Type 2 Dynamics”, Journal of Neurophysiology 92(4) (2004), 2283–2294. Full original article inspected at an external PDF mirror, matched to publisher DOI; especially printed pp. 2284 (Methods) and 2286–2287 (Fig. 4 and discussion). [^ref-eb9e48e68791]: Guillaume Drion, Timothy O’Leary and Eve Marder, “Ion channel degeneracy enables robust and tunable neuronal firing rates”, Proceedings of the National Academy of Sciences 112(38) (2015), E5361–E5370. Full original PDF inspected in an institutional author repository; especially printed E5361–E5362 (Results and Fig. 1), E5364 (Fig. 4), and E5368 (Methods).
Neighborhood in Abstraction Space¶
F-I curve sits in a sparse region of the domain-specific corpus (95th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Unclustered & Miscellaneous (2721 abstractions)
Nearest neighbors
- Coincidence detection in neurobiology — 0.80
- Hyperpolarization (biology) — 0.77
- Rheobase — 0.77
- Pace layers — 0.77
- Correlation integral — 0.76
Computed from structural-signature embeddings · 2026-10-17