How Significant are Cosmological Collider Signals in the Planck Data?
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.
Vera: We're looking at the paper: "How Significant are Cosmological Collider Signals in the Planck Data?".
Jocelyn: Here's the gist of the paper. The search for primordial non-Gaussianities (PNG) is theoretically well motivated but remains observationally challenging In this work, we provide a class of simplified collider templates for these particles that encompasses a broader range of masses, sound speeds, and interactions The most significant result reaching 2.4σ for spin-0, after taking into account the look-elsewhere effect These results indicate that, with refined collider templates and improved data analysis strategies, there are credible prospects with forthcoming observations to detect PNG and also rule out single field inflation In modern cosmology, there is an increasing recognition that the primordial Universe can serve as a natural laboratory for probing fundamental physics at extremely high energies Physical processes during inflation may leave imprints in today’s cosmological observations, such as surveys of the Cosmic Microwave Background (CMB) and Large Scale Structure (LSS) A major observational window for probing new physics during inflation is primordial non-Gaussianity (PNG) Capturing deviations from Gaussian statistics of primordial fluctuations, the PNG at leading order is depicted by the scalar bispectra in Fourier space, with amplitude parameter fNL As the bispectra are functions of three Fourier modes k1, k2, k3, there can be many possibilities for the momenta dependence, also known as shapes Therefore, for the detection of PNG, it is very important to identify various bispectrum shapes from well-motivated theories and prepare the corresponding templates Template Construction For the theory setup, we assume that inflation can be described by a scale-invariant and weakly coupled effective field theory (EFT) with the inflaton field ϕ and additional massive particles σ The dimensionless shape function of a generic scalar bispectrum takes the following form S(k1, k2, k3) = xSequil + ySortho + Xn znS(n) s.e., (1) The presence of the massive particles leads to the single-exchange shapes, and we use the sum over S(n) s.e.
Vera: to represent possible contributions from various masses, spins, sound speeds and interactions For a given mass, spin, interaction and sound speed ratio, Scol.
Jocelyn: takes the following form Scol.
Vera: = Ps(kˆ 2 · kˆ 3) / k3 k1 + k2 1/2 Xa fa(ki; µ) (2) A general bispectrum template can be rewritten as S(k1, k2, k3) = ˜xSequil + ˜ySortho + Xn znS col.
Title and authors: Vera: So, to start, let's talk about the title and who wrote this—How Significant are Cosmological Collider Signals in the Planck Data? The core idea is testing what we expect from heavy particles during inflation using actual Planck data.
Jocelyn: And the authors are Petar Suman, Dong-Gang Wang, Wuhyun Sohn, James R. Fergusson and E P Shellard at the Centre for Theoretical Cosmology and Hong Kong University of Science and Technology in Paris. They’re bringing together cosmology and particle physics perspectives here.
Subrahmanyan: It tells us that even though PNG is a challenging search right now with low signal-to-noise ratios, the strategy needs to change because the shapes we test are highly dependent on the specific theory you start with.
Vera: Exactly. The paper is about moving beyond just looking at standard single-field shapes and building templates for massive particles to see if they show up in current data. It’s about finding what might be hiding in that low signal region.
Jocelyn: They are using the Planck data to test these new collider signals, and they are focusing on making sure their analysis strategy is as efficient as possible with the tools we have available.
Subrahmanyan: This paper sets the stage for how we can systematically explore these different particle signatures, which is crucial because it gives us a structured way to look at the theoretical possibilities before we even start running complex analyses.
The paper's summary: Vera: So what’s the actual summary? Basically, they show that the best strategy for testing these PNG signals right now isn't just using one template; it’s orthogonalizing those collider templates so they don't overlap too much with the tightly constrained single-field shapes.
Jocelyn: They propose constructing an orthogonalized template, which is like taking the basic massive exchange shape and forcing it to be completely uncorrelated with the standard equilateral and orthogonal shapes. They fix two coefficients, alpha and beta, to make sure this new template has zero correlation with those other things.
Subrahmanyan: That idea of orthogonality is key because it means that if we see a signal in this new template, we know it’s not just some weird fluke or an overlap with the single-field predictions; it's a real signature from massive particles.
Vera: And they use the Modal pipeline for this analysis, which is a principal bispectrum method used in Planck analysis, and they say their results are consistent with previous analyses when using these new templates.
Jocelyn: They’re trying to identify the most significant collider signals in the latest data using this orthogonalized approach and also figuring out the most efficient data analysis strategies for testing them at this stage.
Subrahmanyan: The paper is focused on finding those specific signals, which they point toward a massive scalar exchange template with a mass parameter mu equal to one point eight five as being the most significant signal in the current CMB data <ref:2511.17500#pg1>.
The paper's improvements: Vera: So what are the actual improvements this work suggests over previous methods? They focus on three things: first, they include a more complete menu of templates with various masses, spins and sound speeds.
Jocelyn: Second, they orthogonalized the collider templates specifically to minimize their overlaps with the single-field equilateral-type shapes so we can extract those unique signatures of massive particles.
Subrahmanyan: And third, they use the Modal pipeline for data analysis instead of their previous framework, which makes it independent and allows them to test these new templates against existing Planck data in a consistent way.
Vera: So this method is designed to be more robust because by breaking that degeneracy with the single-field shapes through orthogonalization, you get a clearer picture of what's truly out there.
Jocelyn: It also means that the resulting constraints from these orthogonalized templates are easier to interpret, and you can apply this same template set across different observational datasets like CMB or LSS.
Subrahmanyan: The paper explicitly states that this approach highlights the characteristic signatures of the cosmological collider because the resulting template cannot be mimicked by any single field predictions, which is a big theoretical win for separating those signals.
Conclusion: Vera: So to wrap up, the main implication from How Significant are Cosmological Collider Signals in the Planck Data? is that there's a specific massive scalar exchange template with a mass parameter mu equal to one point eight five that stands out as the most significant signal in current Planck data <ref:2511.17500#pg1>.
Jocelyn: They also found another important signal, which is a spin-two template with mu equal to one point zero eight and sound speed cs equal to zero point six eight, giving an adjusted signal-to-noise ratio of one point nine three after accounting for the look-elsewhere effect.
Subrahmanyan: These results might hint at the existence of unknown particles, specifically a massive scalar and possibly KK gravitons in the very early Universe, which would be a concrete direction for developing our theoretical understanding of inflation.
Vera: The paper’s main takeaway is that with refined collider templates and better data analysis strategies like orthogonalization, there are credible prospects with forthcoming observations to detect these PNG signals.
Jocelyn: They also suggest that if these signals are confirmed with more precise measurements in the future, it would rule out single-field inflation entirely and give us a clear path forward in inflation physics.
Subrahmanyan: We need those next-generation experiments like Euclid and LSST to deliver the wealth of data needed to test these collider signals, and this paper provides that common platform for joint constraints across CMB and LSS surveys.
Petar Suman, Dong-Gang Wang, Wuhyun Sohn, James R. Fergusson, E. P. S. Shellard
Centre for Theoretical Cosmology, Department of Applied Mathematics and Theoretical Physics, University of Cambridge · Institute for Advanced Study and Department of Physics, Hong Kong University of Science and Technology · AstroParticule et Cosmologie
astro-ph.CO, gr-qc, hep-ph, hep-th
Submitted: 2025-11-21
Updated: 2026-11-05
Comments: 6 pages, 2 figures, 2 tables
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 91/100
The gist: The search for primordial non-Gaussianities (PNG) is theoretically well motivated but remains observationally challenging In this work…
Key concepts
- Primordial Non-Gaussianity (PNG)
- This refers to deviations from the standard Gaussian statistics expected in the early Universe's fluctuations. These deviations are quantified by parameters like $f_{NL}$ and are studied by analyzing the shape of three-point correlation functions, known as bispectra.
- Collider Templates
- These are simplified theoretical models representing potential imprints of massive particles (like heavy scalars or gravitons) during inflation. They provide a systematic way to search for specific non-Gaussian shapes that single-field theories cannot easily mimic.
- Orthogonalization Strategy
- This method involves constructing new templates ($ ilde{S}_{col}$) by ensuring they are uncorrelated with the tightly constrained single-field predictions ($S_{equil}$ and $S_{ortho}$). This isolates the unique signature of massive particle interactions during inflation, making it easier to detect.
- Look-Elsewhere Effect
- This is a statistical correction applied when searching for signals across many possible parameters. In this context, it adjusts the significance of the observed signal by accounting for the fact that multiple templates are being tested simultaneously.
Terminology
Summary
The search for primordial non-Gaussianities (PNG) is theoretically well motivated but remains observationally challenging In this work, we provide a class of simplified collider templates for these particles that encompasses a broader range of masses, sound speeds, and interactions The most significant result reaching 2. p. 04σ for spin-0, after taking into account the look-elsewhere effect These results indicate that, with refined collider templates and improved data analysis strategies, there are credible prospects with forthcoming observations to detect PNG and also rule out single field inflation
Motivation for the Search
In modern cosmology, there is an increasing recognition that the primordial Universe can serve as a natural laboratory for probing fundamental physics at extremely high energies Physical processes during inflation may leave imprints in today’s cosmological observations, such as surveys of the Cosmic Microwave Background (CMB) and Large Scale Structure (LSS) A major observational window for probing new physics during inflation is primordial non-Gaussianity (PNG) Capturing deviations from Gaussian statistics of primordial fluctuations, the PNG at leading order is depicted by the scalar bispectra in Fourier space, with amplitude parameter fNL As the bispectra are functions of three Fourier modes k1, k2, k3, there can be many possibilities for the momenta dependence, also known as shapes Therefore, for the detection of PNG, it is very important to identify various bispectrum shapes from well-motivated theories and prepare the corresponding templates
Template Construction
For the theory setup, we assume that inflation can be described by a scale-invariant and weakly coupled effective field theory (EFT) with the inflaton field ϕ and additional massive particles σ The dimensionless shape function of a generic scalar bispectrum takes the following form S(k1, k2, k3) = xSequil + ySortho + Xn znS(n) s.e., (1) The presence of the massive particles leads to the single-exchange shapes, and we use the sum over S(n) s. p. 1e. to represent possible contributions from various masses, spins, sound speeds and interactions For a given mass, spin, interaction and sound speed ratio, Scol. p. 1 takes the following form Scol. p. 1 = Ps(kˆ 2 · kˆ 3) / k3 k1 + k2 1/2 Xa fa(ki; µ) (2) A general bispectrum template can be rewritten as S(k1, k2, k3) = ˜xSequil + ˜ySortho + Xn znS col. p. 1 (3)
Orthogonalization Strategy
Before moving to the CMB data analysis, let’s take a closer look at the degeneracy present in the observational tests of cosmological collider To break the degeneracy, one can adopt the approach used in [22], where the authors included the spin-0 massive-exchange templates in (1) and then marginalized over all possible values of the amplitudes x and y of the equilateral and orthogonal shapes Instead, we propose the orthogonalization approach to directly measure the characteristic imprints of massive particles during inflation We construct the orthogonalized template for the collider shape as S˜col. = Scol. + αSequil + βSortho, (5) where we impose the following conditions to fix the two coefficients α and β ⟨S˜col. p. 2 p. 2, Sequil⟩ = 0 and ⟨S˜col. p. 2 p. 2, Sortho⟩ = 0. p. 2 p. 2 (6) This approach highlights the characteristic signatures of the cosmological collider, as the resulting template cannot be mimicked by any single field predictions
Data Analysis Pipeline
The Modal estimator uses two such expansions, one for the primordial bispectrum in k-space, and another for the CMB bispectrum in l-space The Modal and CMB-BEST pipelines have been internally cross-validated for both single field shapes [30] and multiple collider templates [23] We find that the results from the two are consistent for all templates where the modal expansion remains accurate Providing first-ever Modal constraints on this new class of templates is a crucial step towards this goal
Key Findings
The most significant signal of massive particles during inflation in the current CMB data is given by a massive scalar exchange template with µ = 1.85 This may be hinting towards a non-zero scalar bispectrum in the Universe that we live in If this is confirmed with more precise measurements in the future, it would rule out single field inflation and provide a concrete direction for developing our theoretical understanding of the physics of inflation The adjusted SNR is 2. p. 435σ after taking the look-elsewhere effect into account This shape function originates from the massive scalar exchange bispectrum S II col. p. 3 with (∂iϕ)2σ and ϕσ˙ interactions The second most significant signal is given by a spin-2 template S s=2 col. p. 3 with µ = 1.08 and cs = 0.68, which has an adjusted SNR of 1. p. 3 p. 393 The results might provide interesting hints for the existence of unknown particles – one for massive scalars, another possibly for KK gravitons in the very early Universe
Future Directions
The orthogonalized collider templates developed in this work would be useful for future PNG studies with both CMB and LSS, and would provide a common platform to obtain joint constraints It will also be interesting to extend our analysis to collider signals in the primordial trispectrum, which contains richer structures arising from spinning particles during inflation Multiple LSS experiments, such as SphereX, Euclid and LSST, will deliver a wealth of data in the next decade that can be used for testing cosmological collider signals It would be very informative to find out if more accurate and more general templates here would lead to stronger constraints in the forecast The orthogonalized collider templates developed in this work would be useful for future PNG studies with both CMB and LSS, and would provide a common platform to obtain joint constraints
Acknowledgments
We are grateful to Santiago Agüi Salcedo, Xingang Chen, Giovanni Cabass, Ciaran McCulloch, Xi Tong and Bowei Zhang for stimulating discussions JF and PS are supported by Science and Technology Facilities Council (STFC) grant ST/X508287/1 DGW is partially supported by a Rubicon Postdoctoral Fellowship from the Netherlands Organisation for Scientific Research (NWO), the EPSRC New Horizon grant EP/V048422/1, and the Stephen Hawking Centre for Theoretical Cosmology WS is supported by the SCIPOL project funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Grant agreement No. 101044073) Part of this work was undertaken on the Cambridge CSD3 part of the STFC DiRAC HPC Facility (www. p. 4dirac.ac.uk) funded by BEIS capital funding via STFC Capital Grants ST/P002307/1 and ST/R002452/1 The paper is 2511. p. 417500.
--- Page 6 ---
How Significant are Cosmological Collider Signals in the Planck Data? The search for primordial non-Gaussianities (PNG) is theoretically well motivated but remains observationally challenging Tight constraints with low significance for the standard non-Gaussian shapes suggest that detection may lie beyond the reach of near-future experiments However, tests of PNG are highly template-dependent From a theory perspective, a whole new family of bispectrum shapes arise in the cosmological collider program, with distinct signatures of heavy particles during inflation In this work, we provide a class of simplified collider templates for these particles that encompasses a broader range of masses, sound speeds, and interactions We propose that, given the current state of observations, the most effective strategy to search for PNG signals is through orthogonalizing the collider templates, such that they are uncorrelated with the tightly constrained single field predictions Using the Modal pipeline and Planck CMB data, we perform a systematic parameter scan of the collider templates with the most significant result reaching 2. p. 04σ for spin-0, after taking into account the look-elsewhere effect; indicative results for spin-1 and spin-2 peak near 2σ
--- Page 3 ---
A complete set of scalar bispectra– Let’s first prepare the analytical templates for bispectrum shape functions For the theory setup, we assume that inflation can be described by a scale-invariant and weakly coupled effective field theory (EFT) with the inflaton field ϕ and additional massive particles σ[10, 11, 24–26] Then at the leading order, the dimensionless shape function of a generic scalar bispectrum takes the following form S(k1, k2, k3) = xSequil + ySortho + Xn znS(n) s.e., (1) The next step is to derive analytical templates for the massive-exchange correlators First, the EFT background Sbkg which consists of rational polynomials of the kinematic variables and resembles the equilateral or orthogonal templates Secondly, the non-analytic collider signal Scol. p. 1 which exhibits squeezedlimit oscillations The latter provides the characteristic imprint of the new massive particles that cannot be mimicked by any single field shapes With this knowledge, we may absorb the Sbkg parts of massive-exchange shapes into the two single field PNG templates in (1) Then a general bispectrum template can be rewritten as S(k1, k2, k3) = ˜xSequil + ˜ySortho + Xn znS col. p. 1 (3)
--- Page 4 ---
cos(S(1), S(2)) ≡⟨S(1), S(2)⟩/p⟨S(1), S(1)⟩⟨S(2), S(2)⟩, where the inner product is defined as ⟨S(1), S(2)⟩ ≡ yVk w(ki)S(1)(ki)S(2)(ki)dVk, (4) This weight function is chosen to reflect the l−scaling of the fNL estimator in the CMB If this is the case, then the significance of cosmological collider-like PNG signals depends on the parameters governing the single-field selfinteractions Instead, we propose the orthogonalization approach to directly measure the characteristic imprints of massive particles during inflation In practice, we construct the orthogonalized template for the collider shape as S˜col. = Scol. + αSequil + βSortho, (5) where we impose the following conditions to fix the two coefficients α and β ⟨S˜col., Sequil⟩ = 0 and ⟨S˜col., Sortho⟩ = 0. (6)
--- Page 4 ---
The fNL corresponds directly to the PNG of the shape S˜col. First, the resulting constraints are more easily interpretable Second, the full set of templates designed in this approach can be easily applied in any observational dataset, and thus constraints from different surveys can be compared on an equal footing Third, for the linear bispectrum model in the CMB context, the constraints obtained from orthogonalized templates are easier to compute, while remaining equivalent to the marginalized constraints up to a normalization factor Last, this approach highlights the characteristic signatures of the cosmological collider, as the resulting template cannot be mimicked by any single field predictions
--- Page 4 ---
We perform a vast parameter search in the log-uniform range of cs between 10−2 and 102, and in the uniform mass range of µ between 1 and 6 Raw SNR maximized over cs as a function of mass parameter is shown in Figure 1 Any masses heavier than this would yield highly oscillating (and exponentially suppressed) shapes for which the accuracy of Modal decomposition drops low The most significant signals correspond to Scalar-II (3.34σ) and Spin-2 (2. p. 395σ)
--- Page 4 ---
The primordial template in our analysis that is most favoured by the Planck data is shown in Figure 2 After taking into account the look-elsewhere effect, the adjusted SNR is 2.35σ This shape function originates from the massive scalar exchange bispectrum S II col. p. 3 with (∂iϕ)2σ and ϕσ˙ interactions It is the unorthogonalised template which has some overlap with the single field orthogonal shape, while also demonstrating the characteristic oscillatory signals of the massive scalar particle This is consistent with the previous result of the CMB-BEST analysis [21], where the most significant signal of 1. p. 38σ comes from a similar template with µ = 2. p. 313 The second most significant signal is given by a spin-2 template S s=2 col. p. 3 with µ = 1.08 and cs = 0.68, which has an adjusted SNR of 1.93 These results might provide interesting hints for the existence of unknown particles – one for massive scalars, another possibly for KK gravitons in the very early Universe The most significant signal of massive particles during inflation in the current CMB data is given by a massive scalar exchange template with µ = 1. p. 385
--- Page 4 ---
The most significant signal of massive particles during inflation in the current CMB data is given by a massive scalar exchange template with µ = 1.85 This may be hinting towards a non-zero scalar bispectrum in the Universe that we live in If this is confirmed with more precise measurements in the future, it would rule out single field inflation and provide a concrete direction for developing our theoretical understanding of the physics of inflation
--- Page 4 ---
The most significant signal of massive particles during inflation in the current CMB data is given by a massive scalar exchange template with µ = 1.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
The AI system can perform systematic parameter scans of collider templates by using orthogonalizing techniques to break degeneracy with single field predictions, as described by
we propose that, given the current state of observations, the most effective strategy to search for PNG signals is through orthogonalizing the collider templates
. -
The improved system can generate analytical templates for massive-exchange correlators using recent developments on the cosmological bootstrap to provide a
systematic classification of all possible predictions based on minimal assumptions
. -
The AI system can identify
the most significant collider signals in the latest data
by comparing constraints from the two analytic templates, Scol and Stilde col, using a pipeline like Modal or CMB-BEST. -
The system can perform constraint generation by utilizing the orthogonalized template Stilde col, which is described as
the primordial template in our analysis that is most favoured by the Planck data
.
Sources
- Primordial Non-Gaussianity
- Inflation: Theory and Observations
- Primordial Non-Gaussianities from Inflation Models
- Quasi-Single Field Inflation and Non-Gaussianities
- Signatures of Supersymmetry from the Early Universe
- Effective field theory approach to quasi-single field inflation and effects of heavy fields
- Cosmological Collider Physics
- The Cosmological Bootstrap: Inflationary Correlators from Symmetries and Singularities
- Snowmass White Paper: The Cosmological Bootstrap
- Boostless Cosmological Collider Bootstrap
- Cosmological Bootstrap in Slow Motion
- Helical Inflation Correlators: Partial Mellin-Barnes and Bootstrap Equations
- Bootstrapping Multi-Field Inflation: non-Gaussianities from light scalars revisited
- The Exact and Approximate Tales of Boost-Breaking Cosmological Correlators
- Planck 2018 results. IX. Constraints on primordial non-Gaussianity
- The shape of primordial non-Gaussianity and the CMB bispectrum
- General CMB and Primordial Bispectrum Estimation I: Mode Expansion, Map-Making and Measures of f_NL
- BOSS Constraints on Massive Particles during Inflation: The Cosmological Collider in Action
- The Effective Field Theory of Inflation
- Non-Gaussianity as a Particle Detector
Related papers
- 21cmEMUv3: a hybrid diffusion-LSTM emulator of 21cmFAST summary observables
- Connecting field-level and n-point function measurements of galaxy bias
- The Impact of Cosmic Variance and Satellites on JWST Clustering Measurements at Redshift around 6
- Universality of Primordial Anisotropies in Gravitational Wave Background
- KiDS-1000 cosmic shear reanalysis using MetaCalibration
- 2Flow2SN: Dusty Supernova Cosmology with Photometric Redshifts and Contamination using Neural Simulation-Based Inference