The four-parameter logistic (4PL) curve fits dose-response data as y = Bottom + (Top − Bottom) / (1 + 10^((LogEC50 − x) · HillSlope)).
The 4PL is the standard dose-response model in immunoassays and drug screens. Four parameters: Bottom (lower asymptote), Top (upper asymptote), LogEC50 (the log-dose at which y is half-way between Bottom and Top), HillSlope (the slope at the midpoint — positive for increasing curves).
For data on log-dose axis it sigmoid-fits cleanly. On linear-dose axis a wide dose range is needed to identify Top and Bottom; with a narrow range the model is unidentifiable and you should either narrow to 2-parameter fitting (fix Top / Bottom) or use the Hill / Michaelis-Menten alternative.
Absolute IC50/EC50 (the dose at which y reaches a specific value, e.g. 50% of control) is computed in the result panel by inverse interpolation — different from LogEC50, which is the relative midpoint.
If a fit looks visibly wrong, enable Multi-start: it refits from many seeded starting points (deterministically) and keeps the lowest residual-sum-of-squares result, which reliably escapes the local minima that trap a single starting guess — most useful for the flexible models (logistic growth, 5PL asymmetry) whose parameters sit on flat likelihood ridges.
Numeric X (dose / log-dose) + Y (response).
Estimate, SE, t, p, 95% CI per parameter + fitted curve with 95% CI + 95% PI bands (delta method) + ROUT-on-residuals outlier flags at Q = 1% + absolute IC50/EC50 calculator + LD50 etc.
A HillSlope very different from 1 ⇒ the curve is steeper or shallower than the classical Hill equation — try the Hill model and compare AICc.