Assess validity of GR-first global fits for LISA and their impact on tests of GR

Investigate whether a LISA data analysis strategy that first performs a global fit under general relativity to all overlapping sources and instrument noise, and only then conducts tests of general relativity a posteriori on the detected signals, yields unbiased results or is systematically biased by the omission of beyond-GR effects during the global fit.

Background

LISA data analysis requires a global fit that simultaneously models many overlapping signals and instrument noise. The paper emphasizes that robust, unbiased tests of GR within this framework have not yet been demonstrated under realistic assumptions.

If the global fit assumes GR is exactly correct, deviations present in the data might be partially absorbed by the fit, biasing subsequent a posteriori GR tests on individual sources.

Clarifying the validity of this workflow, or devising alternative strategies that avoid such biases, is essential to ensure reliable fundamental-physics tests with LISA.

References

"In this sense, our ability to perform robust tests of GR with LISA signals has not been demonstrated. For example, is still unclear if we can first perform a global fit assuming GR to be correct and perform tests of GR a posteriori on the detected signals, because in principle the omission of deviations from GR could bias the global fit."

— Black hole spectroscopy: from theory to experiment (2505.23895 - Berti et al., 29 May 2025) in Section 7.3 (Data analysis for future detectors), last paragraph