LN Model · STRFs · Reverse Correlation · Auditory Cortex
🧠 Tier: Standard UG/Graduate · Auditory Neuroscience
Version:
§ 01
Interactive Simulation
Variable 1
—
—
Variable 2
—
—
Input
—
—
Output
—
—
Count
0
—
State
—
—
Time
0
ms
Playback
Preset
Parameters
Parameter 15.0
Parameter 25.0
Input I5.0
T sim (ms)500
Overlays
§ 02
The Idea, Step by Step
▸ From "what sound does this neuron love?" to the STRF
Picture a friend at a noisy party who only looks up when their favorite song comes on. A neuron in your auditory cortex behaves the same way: out of every possible sound, it "wakes up" — fires more spikes — for one particular pattern of pitches arriving with a particular timing. A spectrotemporal receptive field (STRF) is simply a picture of that favorite sound.
START — the neuron's favorite sound
Sound has two ingredients that matter here: which pitches (frequency) and when they happen (timing). An STRF is a 2-D map with pitch on one axis (say 0–8 kHz) and "how long ago" on the other (time lag, 0–200 ms). Warm patches mean "this pitch, this long before now, makes me fire"; cool patches mean "this one quiets me down."
BUILD — averaging the sound before each spike
To find that map, play random noise — a jumble of all pitches — and mark the instant the neuron fires. Then look back at the slice of sound just before each spike and average all those slices. The random parts cancel; whatever pattern keeps showing up is the filter. This is the spike-triggered average (STA). If a cell fires $N_{spk}=1000$ spikes and you average the 200 ms of sound before each, that average is its STRF fingerprint. The neuron's drive is how strongly recent sound overlaps the filter — multiply and add: $x(t)=\sum STRF(f,\tau)\,s(f,t-\tau)$.
DEEPEN — the linear–nonlinear (LN) model
Turn the sum into integrals over frequency $f$ and lag $\tau$: $x(t)=\int_0^\infty\!\!\int_0^\infty STRF(f,\tau)\,s(f,t-\tau)\,d\tau\,df$ — the linear stage. Real neurons can't fire negative rates, so pass $x$ through a static nonlinearity $g$ (a threshold or sigmoid): $r(t)=g(x(t))$. That two-step LN model captures roughly 60–80% of an A1 neuron's response. For Gaussian white noise the raw STA is an unbiased estimate of the filter; for correlated stimuli you "whiten" it, $\hat{k}=C_{ss}^{-1}C_{sr}$. On the panel, Parameter 1 and Parameter 2 stand in for the filter's spectral and temporal tuning width, Input I scales the drive into the nonlinearity, and T sim sets how long you record.
TRY THIS — in the sim above
Push Input I up and watch the Output card climb as the nonlinearity is driven harder — then back to zero to see it fall silent below threshold. Stretch T sim to record a longer window (more spikes → a cleaner STA in real data). Press Play and watch the clock advance, then Reset to start a fresh recording.
Present broadband (white) noise → record spikes → average stimulus in window before each spike. For Gaussian noise: STA = true filter (unbiased). For non-Gaussian: regularised regression (ridge, LASSO). STRF in 2D: frequency axis (0–8 kHz) × time lag (0–200 ms).
STEP 2 — LN Model
Linear stage: x(t) = ∫ STRF(f,τ) s(f,t-τ) dfdτ. Static nonlinearity: r(t) = g(x(t)) — sigmoid or half-wave rectification. LN captures ~60–80% of response variance for V1, A1 neurons. Explains selectivity (filter) and threshold (nonlinearity).
STEP 3 — Auditory Cortex STRFs
A1 STRFs: separable (F(f)×T(τ)) or non-separable (frequency sweep selectivity). Best temporal modulation: 4–16 Hz. Best spectral modulation: 0.2–2 cycles/octave. STRFs predict natural speech responses with R² ≈ 40–60%. Non-separable STRFs more common in higher auditory cortex (A2, belt areas).
▸ Primary References
Depireux et al. (2001) J Neurophysiol; Simon et al. (2007) Neuron
🧠 ConceptualWhat is the core mathematical insight of Spectrotemporal Receptive Fields?▼
The Spectrotemporal Receptive Fields framework provides a rigorous bridge between microscopic neural parameters and macroscopic observables. The key insight is that complex emergent behaviours — oscillations, memory, coding precision — can be derived from simple mathematical rules governing individual neurons and their connections. The quantitative framework enables testable predictions about circuit function from experimentally measurable quantities.
Key: Mathematical rigour connects single-neuron parameters to network behaviour. Every emergent property (oscillation, attractor, code) is derivable from the microscopic rules.
🌍 AppliedHow is Spectrotemporal Receptive Fields used in neurotechnology?▼
The Spectrotemporal Receptive Fields framework is directly applied in brain-computer interfaces and computational medicine. Mathematical models derived from these principles optimise stimulation parameters for DBS, predict drug effects on network dynamics, decode neural signals for BCI applications, and guide the design of neuromorphic hardware. The quantitative framework enables model-based optimisation rather than empirical trial-and-error approaches in clinical settings.
Key: Direct applications in BCI design, DBS parameter optimisation, pharmacological modelling, and neuromorphic chip architecture.
🔬 SimulationHow do I use the simulation for Spectrotemporal Receptive Fields?▼
Configure the parameter sliders on the right panel, select a preset to set up a canonical example, then press Play. The Main View shows the primary mathematical relationship; the Phase Plane shows geometric structure; the Time Series shows temporal evolution; the Parameter Sweep reveals sensitivity to inputs. All displayed quantities correspond directly to terms in the equations in Section 2. Export the data as CSV for further analysis.
Key: Each tab maps to a specific aspect of the mathematics. Use Parameter Sweep to find bifurcation points and Phase Plane to understand stability.
💡 Non-ObviousWhat is counterintuitive about Spectrotemporal Receptive Fields?▼
The most surprising result in Spectrotemporal Receptive Fields research is that small parameter changes can produce qualitatively different dynamics (bifurcations). A system that is stable for one parameter value can suddenly oscillate, burst, or become chaotic with a tiny perturbation. This sensitivity to parameters is the hallmark of nonlinear systems and has profound implications: the same neural circuit can perform completely different computations depending on its neuromodulatory state, which shifts parameters near bifurcation boundaries.
Key: Bifurcations: tiny parameter changes → qualitative behavioural change. Neuromodulators exploit this by operating near bifurcation points to switch neural circuit function.
📐 ComputationalWhat numerical methods and pitfalls apply to Spectrotemporal Receptive Fields?▼
The appropriate numerical method depends on stiffness: for near-spike dynamics (HH, AdEx) use RK4 with dt ≤ 0.025 ms or implicit methods; for mean-field equations (WC, Amari) use Euler or RK4 with dt ≤ 0.5–1 ms. Always verify accuracy by halving dt and checking solution convergence. The most common errors: wrong units (mixing mV, ms, nS), missing initial conditions (not setting w₀=b×v₀ in Izhikevich), and incorrect boundary conditions in Fokker-Planck or cable equation simulations.
Key: Verify stability by halving dt. Most errors: wrong units, incorrect initial conditions, violated model assumptions. Check the model validity range before trusting any result.
§ 04 Resources
Gerstner et al. — Neuronal Dynamics. Free: neuronaldynamics.epfl.ch
Dayan & Abbott — Theoretical Neuroscience, MIT Press, 2001
Izhikevich — Dynamical Systems in Neuroscience. Free: dynamicalsystems.org
Neuromatch Academy — neuromatch.io (free comp neuro tutorials)
§ 05
Misconceptions & Common Errors
Sub-block A — Conceptual
❌"The Spectrotemporal Receptive Fields model is only theoretical with no experimental support."
✅The Spectrotemporal Receptive Fields framework has been extensively validated. Key predictions have been confirmed in electrophysiology, imaging, and pharmacological experiments. Modern neuroscience uses these models as quantitative tools integrated with the experimental cycle — for prediction, interpretation, and guidance of experiments.
📖 Depireux et al. (2001) J Neurophysiol; Simon et al. (2007) Neuron
Sub-block B — Numerical
❌Applying Spectrotemporal Receptive Fields equations outside their range of validity (wrong N, dt, or approximation regime).
✅Every model has a validity regime. Always check: (1) Is N large enough for mean-field? (2) Is dt small enough (halve dt and verify solution converges)? (3) Are units consistent (mV, ms, nS throughout)? (4) Are initial conditions correct? When in doubt, compare to direct simulation of the full system.
🔍 Why: Violating assumptions gives quantitatively or qualitatively wrong results — and the error is often invisible without explicit validation.
§ 05 References
Depireux et al. (2001) J Neurophysiol; Simon et al. (2007) Neuron
Gerstner et al. — Neuronal Dynamics, Cambridge, 2014