Each passing vehicle produces a brief indoor noise peak L₁ = L_Amax − Insulation. Whether it wakes the sleeper is modelled as a logistic dose-response curve — the same functional form used in field studies of aircraft- and road-noise sleep disturbance (e.g. Basner et al.):
P(arousal) = 1 / (1 + exp(−b · (L₁ − L₅₀[stage])))
b ≈ 0.12 per dB (steepness of the dose-response curve)
L₅₀[stage] = indoor level at which P = 50%:
N1 (light) → 50 dB REM → 52 dB
N2 → 62 dB N3 (deep) → 75 dB
Deep sleep (N3) has the highest arousal threshold — the brain's thalamic gating suppresses sensory relay most strongly there — while N1 and REM are far more easily disturbed. Each pass-by event is drawn as a Poisson process at the chosen rate; a random draw against P(arousal) decides whether that event produces a micro-arousal, which resets the sleeper to a brief waking stage and interrupts the ~93-minute ultradian N1→N2→N3→REM cycle that would otherwise run through the night.
- LAmax — the peak outdoor sound level of a single pass-by, in dB(A).
- Events / hour — traffic density outside the window.
- Insulation — façade/window sound reduction; WHO's 2018 Environmental Noise Guidelines recommend keeping bedroom Lnight below ≈40 dB.
- Sleep efficiency — the fraction of elapsed night actually spent asleep rather than in an arousal.
This 2D rendition drives the exact same stage-scheduling and dose-response equations as the 3D scene, plotted directly as a scrolling hypnogram rather than a 3D bedroom — drag to pan the strip, scroll/pinch to zoom the time axis.
This is a simplified, literature-inspired teaching model, not a clinical instrument — real dose-response coefficients vary with individual sensitivity, prior sleep debt and noise spectrum.