This paper looks for traces of extra massive particles that might have been present during cosmic inflation. Such particles would leave a specific imprint in how galaxies cluster on the largest scales. The authors analyze imaging data used for the Dark Energy Spectroscopic Instrument (DESI) targets and they make forecasts for future, more powerful surveys to see how well these signatures could be measured.

The signal the team searches for is a type of “primordial non-Gaussianity” (PNG). PNG means the initial density variations in the early universe deviate slightly from a perfect random (Gaussian) pattern. A convenient parameterization has an amplitude f_NL,Δ and a scaling index Δ. The index Δ is tied to the mass of the extra field: Δ=0 matches the usual “local” PNG (massless extra field), while larger Δ values correspond to heavier fields. These departures change the galaxy bias on large scales, producing a scale-dependent enhancement of clustering proportional to roughly k^{Δ−2} (k is the inverse length scale).

To look for this effect in real data the authors use the angular correlation function of luminous red galaxies (LRGs) from DESI Legacy Imaging. They apply a pipeline developed earlier to model how a nonzero f_NL,Δ and Δ would change galaxy clustering. With the most aggressive treatment of observational systematics they find no evidence for local PNG, and therefore cannot place a useful constraint on Δ. With a less aggressive decontamination of systematics, they recover a previously reported hint of local PNG: f_loc_NL = 27^{+10}_{-9}, consistent with earlier imaging analyses. If that imaging hint is interpreted as a beyond-local signal, the data prefer f_NL,Δ = 5.12^{+6.13}_{-3.62}×10^3 and Δ = 0.91^{+0.25}_{-0.19}. The authors emphasize this preference is not robust: it changes with different decontamination choices and is likely driven by residual observational systematics. They present the result as an illustration of the method rather than a detection.