The extremely slow, steady pulsations of light from many red giant stars may finally have an explanation.

According to a new analysis, these mysterious fluctuations in brightness are not caused by internal processes after all, but by binary companions obscured in clouds of dust siphoned off the dying giants.

When stars of intermediate mass below around eight times the mass of the Sun reach the twilight of their lives, they go through some pretty dramatic changes.

When they have fused all the hydrogen in their cores to helium, the nuclear fusion within ceases, and the core starts to contract. This brings more hydrogen into the region immediately around the core, forming a hydrogen shell; then, fusion starts up again, dumping helium into the core. This is called hydrogen shell burning.

During this time, the outer layers of the star expand - by a lot. When this eventually happens to the Sun, for example, it will expand out past the orbit of Earth. This is the red giant branch of stellar evolution.

Red giant stars often fluctuate in brightness a little, over regular periods. The red giant star Betelgeuse is a perfect example of this. It has several brightness cycles, including one that occurs over around 425 days, and another over around 185 days. These are caused by acoustic waves bouncing around inside the star as it expands, contracts, and expands again.

The longest of its cycles is more mysterious. It's what we call a "long secondary period", and it's 5.9 years long. Not all giant branch stars have long secondary periods, but a lot of them do - scientists have detected a long secondary period in around a third of all known giant branch stars - and these periods cannot be explained in the same way.

A few explanations have been put forward for these mysterious thrums in the light of dying stars, including a different kind of internal oscillation, magnetic activity, or the presence of a binary companion.

To try and get to the bottom of the mystery, a team of astronomers led by Igor SoszyƄski of the University of Warsaw in Poland conducted a close study of red giant stars with long secondary periods. From available survey data, they collated optical and mid-infrared observations of 16,000 of these stars, from which they extracted around 700 stars with a well-defined infrared light curve - a plot of the way the light changes over time - for a closer analysis.

When comparing the optical and infrared light curves for these 700 stars, something curious emerged. In both light curves for all the stars, there was a large dip, as expected, corresponding to the stars' dimmer periods. But for around half of the stars, there was a second, shallower dip only in the infrared light curve, exactly opposite the primary dip.

This, the team said, is an important clue. Mid-infrared light is often produced by dust - it absorbs starlight, and re-emits it at longer wavelengths.

This can neatly explain what's happening around the red giant stars. If the star is being orbited by a smaller companion that has siphoned off material from the star and is therefore trailing a long dust cloud, this companion will produce a long, strong dip in starlight at all wavelengths when it passes between us and the star.

Then, as this dusty object moves around to the side of the star, we will be able to see mid-infrared light as the starlight is absorbed and re-emitted. This mid-infrared light will dip when the binary companion moves behind the star, only to glow again when the companion re-emerges out the other side.

According to the team's analysis, the amplitudes of the light curves suggest that the companion is either a very low mass star, or a brown dwarf. But brown dwarfs - stars that didn't grow large enough to be stars, but grew too large to be planets - are relatively rare.

If the companions are brown dwarfs, the team said, they could have started their lives as smaller exoplanets, and siphoned material off the red giant stars' outer envelopes. This suggests that most red giants with long secondary periods are orbited by objects that used to be exoplanets.

In turn, the researchers said, this finding could allow long secondary period giant branch stars to be used as tracers for studying the planetary population of the Milky Way.

The team's research has been published in The Astrophysical Journal Letters.