Determine the bias induced by the quasi-static occupancy approximation

Determine the magnitude and direction of the residual bias in harmonic-spectrum estimates caused by applying the stationary occupancy estimator to moving cue geometries in the closed-loop Drosophila virtual-reality recordings.

Background

The occupancy estimator treats the heading distribution as equilibrated at a fixed cue configuration, although the animal’s movement continuously changes cue bearings. Closed-loop simulations and open-loop replay produced biases of opposite signs, with the former underestimating and the latter overestimating the first harmonic. Because neither simulation directly identifies the experimental bias, the paper leaves both its magnitude and direction unresolved.

References

However, the closed-loop simulation uses an idealized model animal, whereas the replay breaks the real feedback loop; neither directly identifies the bias in the experimental estimate. We therefore apply no correction: all reported spectra are the original fits, and both the magnitude and direction of any residual bias remain open (S4).

— Selection rules for the harmonic spectroscopy of animal decisions  (2610.00990 - Salahshour, 1 Oct 2026) in Section 3.3.2, “The heading dynamics, measured and cross-checked” (S4)