Starburst galaxies, which are known as “reservoirs” of high-energy cosmic-rays, can represent an important high-energy neutrino “factory” contributing to the diffuse neutrino flux observed by IceCube. In this paper, we revisit the constraints affecting the neutrino and gamma-ray hadronuclear emissions from this class of astrophysical objects. In particular, we go beyond the standard prototype-based approach leading to a simple power-law neutrino flux, and investigate a more realistic model based on a data-driven blending of spectral indexes, thereby capturing the observed changes in the properties of individual emitters. We then perform a multi-messenger analysis considering the extragalactic gamma-ray background (EGB) measured by Fermi-LAT and different IceCube data samples: the 7.5-year High-Energy Starting Events (HESE) and the 6-year high-energy cascade data. Along with starburst galaxies, we take into account the contributions from blazars and radio galaxies as well as the secondary gamma-rays from electromagnetic cascades. Remarkably, we find that, differently from the highly-constrained prototype scenario, the spectral index blending allows starburst galaxies to account for up to $40\%$ of the HESE events at $95.4\%$ CL, while satisfying the limit on the non-blazar EGB component. Moreover, values of $\mathcal {O}(100~\mathrm{PeV})$ for the maximal energy of accelerated cosmic-rays by supernovae remnants inside the starburst are disfavoured in our scenario. In broad terms, our analysis points out that a better modeling of astrophysical sources could alleviate the tension between neutrino and gamma-ray data interpretation.

Star-forming and starburst galaxies, which are well-known cosmic-ray reservoirs, are expected to emit gamma-rays and neutrinos predominantly via hadronic collisions. In this Letter, we analyze the 10 yr Fermi-LAT spectral energy distributions of 13 nearby galaxies by means of a physical model which accounts for high-energy proton transport in starburst nuclei and includes the contribution of primary and secondary electrons. In particular, we test the hypothesis that the observed gamma-ray fluxes are mostly due to star-forming activity, in agreement with the available star formation rates coming from IR and UV observations. Through this observation-based approach, we determine the most likely neutrino counterpart from star-forming and starburst galaxies and quantitatively assess the ability of current and upcoming neutrino telescopes to detect them as pointlike sources. Remarkably, we find that the cores of the Small Magellanic Cloud and the Circinus galaxy are potentially observable by KM3NeT/ARCA with 6 yr of observation. Moreover, most of the nearby galaxies are likely to be just a factor of a few below the KM3NeT and IceCube-Gen2 pointlike sensitivities. After investigating the prospects for detection of gamma-rays above TeV energies from these sources, we conclude that the joint observations of high-energy neutrinos and gamma-rays with upcoming telescopes will be an objective test for our emission model and may provide compelling evidence of star-forming activity as a tracer of neutrino production.

The gamma-ray emission from Starburst and Starforming Galaxies (SBGs and SFGs) strongly suggest a correlation between star-forming activity and gamma-ray luminosity. However, the very nature of cosmic ray (CR) transport and the degree of their confinement within SBG cores are still open questions. We aim at probing the imprints left by CR transport on gamma-ray and neutrino observations of point-like SFGs and SBGs, looking into quantitative ways to discriminate among different transport models. We analyse the 10-year Fermi-LAT spectral energy distributions of 13 nearby galaxies with two different CR transport models, taking into account the corresponding IR and UV observations. We also generate mock gamma-ray data to simulate the CTA performance in detecting these sources. In this way, we propose a test to discriminate between the two CR models, quantifying the statistical confidence at which one model can be preferred over the other. We point out that current data already give a slight preference to CR models which are dominated by advection. Moreover, we show that CTA will allow us to firmly disfavour models dominated by diffusion over self-induced turbulence, compared to advection-dominated models, with Bayes factors which can be as large as 107 for some of the SBGs. Finally, we estimate the diffuse gamma-ray and neutrino fluxes of SFGs and SBGs, showing that they can explain $25{{\%}}$ of the diffuse HESE data, while remaining consistent with gamma-ray limits on non-blazar sources.