Unexplained "filaments" of radio-wave emission close to our galaxy's centre may hold proof of the existence of dark matter, researchers have said.
Dark matter is believed to make up most of the mass of our Universe, but it has yet to be definitively spotted.
A report now suggests the filaments' emission arises from dark matter particles crashing into each other.
However, the work, posted to the Arxiv repository, requires extensive further experiments to support or refute it.
The filaments have been something of a mystery to astronomers since they were first discovered in the 1980s.
They are known to be regions of high magnetic fields, and they emit radio waves of high frequency - some of them with striking intensity.
"There's a long literature about these objects, and there have been some ideas as to what might generate their emission - but frankly no one really knows," said Dan Hooper, an astrophysicist at the Fermi National Accelerator Laboratory (Fermilab) in the US and co-author of the paper, which is still under review by academics.
One explanation for this emission would be what is called synchrotron radiation, which arises when charged particles are accelerated in a magnetic field. There are several ideas that could account for the emission which do not invoke dark matter - so called "astrophysical" mechanisms.
Now, Dan Hooper and his colleagues suggest that electrons - created when high-energy dark matter particles smash into each other - could be the what gives rise to the synchrotron radiation detected here on Earth.
He credits co-author Tim Linden for coming up with the idea, which he said "can explain a lot of the different features that are observed" in the filaments' emission - something he said more prosaic "astrophysical" explanations could not claim.
"One thing it explains that the astrophysical possibilities don't is that the filaments that are closer to the galactic centre are brighter than those that are farther away," Dr Hooper told BBC News. "We would say that's because there's more dark matter as you come closer to the galactic centre - it provides a natural explanation for that."
In the model that the team has developed, the electrons in all the filaments that were studied should have a high energy - between five and 10 billion electron volts (5-10GeV).
Dr Hooper said:"The question is: why would all of these filaments which are different astrophysically, contain different stuff, located in different places - all sorts of different properties - all have electrons with that much energy?
"In the dark matter explanation, that's easy - dark matter is the same everywhere."
Dr Hooper has also published papers recently suggesting that dark matter particles of the same energies fit with recent results from the Fermi space telescope (in an article in Physics Letters B) and with efforts to detect dark matter on Earth in so-called "direct detection" experiments (in an as-yet unpublished paper on Arxiv).
"That's definitely one of the strengths of this model; the results seem promising," said Sukanya Chakrabarti, an astrophysicist from the Florida Atlantic University.
However, theoretical models of a substance that has never been detected necessarily require a number of educated guesses and estimates - guesses that could radically affect whether or not a given theory stands up.
"When you do these kind of 'indirect detection' experiments, there are many parameters that go into your model," Professor Chakrabarti told BBC News. "All that stuff that's not known - it's hard to do a study of all these and convince yourself of all mechanisms [that lead to the emission]."
Troy Porter, an astrophysicist from Stanford University, said that dark matter particles of energies as high as 10 GeV are "already in severe conflict with the recent [preliminary and as-yet unpublished] results reported by the Fermi-LAT collaboration at the Rome Fermi symposium for an analysis of nearby dwarf spheroidal galaxies".
The results from detections in underground experiments on Earth are also not widely agreed to point to a dark matter explanation, but Dr Hooper said forthcoming results from the Cresst experiment in Italy will lend further credence to his team's theory.
What will resolve these issues in the case of the filaments are simply more observations using more radio telescopes.
"Many of these filaments have only limited data available about them," said Dr Hooper. "I hope this paper inspires radio astronomers to look more carefully at these objects."