Cosmic Imposters: How “Medium” Vision is Changing Our Map of the Early Universe


How far back into the “Once Upon a Time” of our universe can we actually see? For decades, astronomers have been chasing the first glimmers of light that broke the cosmic dark ages. Today, a team led by Yoshihisa Asada (University of Toronto), alongside experts like Chris J. Willott from NRC Herzberg, has released a study that doesn’t just find new galaxies—it teaches us that our eyes might have been playing tricks on us.

By using the James Webb Space Telescope (JWST) with a specialized set of “Medium-band” filters, the team identified eight incredibly rare galaxies from the dawn of time, including one robust candidate existing just 250 million years after the Big Bang. More importantly, they discovered that without these specific filters, nearly half of the “earliest galaxies” we thought we found in other studies might actually be “cosmic imposters”—dusty, younger galaxies pretending to be ancient.

The Mystery of the Cosmic Speedometer

To understand this discovery, we first have to talk about redshift (the way light stretches as it travels through the expanding universe, turning bluer light into longer, redder wavelengths). Think of it like the Doppler effect of a siren passing you by, but for light. The further away a galaxy is, the more its light is “redshifted.”

For years, astronomers have looked for the Lyman Break. This is a sharp “cliff” in a galaxy’s light spectrum. Because hydrogen gas in space soaks up specific ultraviolet light, a very distant galaxy will appear bright in redder filters but completely vanish in bluer ones.

The Analogy: Imagine you are looking for a specific person in a crowded stadium using only three colored spotlights: Red, Green, and Blue. If you see a “red” person, are they actually wearing a red shirt (a young, dusty galaxy), or are they a “white-shirt” person standing so far away that the stadium’s atmosphere has turned their light red (an ancient galaxy)? With just three big lights, it’s hard to tell.

The “Eureka” Moment: Seeing in Technicolor

Most JWST surveys use Broad-band filters, which look at wide chunks of the light spectrum. While efficient, these wide views can be fooled. A younger, dusty galaxy can mimic the “cliff” of an ancient one, leading to “False Positives”.

The CANUCS and Technicolor (TEC) surveys took a different approach. They used 22 different filters, including “Medium-band” filters that act like a finer-toothed comb for light. This allowed Asada’s team to “see” the cliff with incredible precision.

Using this method, they found CANUCS-3214552, a galaxy at a redshift of roughly 15.4. It is faint, compact, and surprisingly red. Most importantly, because they had the “Medium” filters, they could prove it wasn’t a closer interloper. Their data suggests that at these extreme distances, galaxies are smaller and rarer than some previous, broad-filter studies suggested.

Mapping the Future

This discovery is a “reality check” for cosmology. It suggests that the “normalization” (the total number of galaxies we expect to find) might be four times lower than previously thought because so many “candidates” in other surveys are likely closer, dusty galaxies.

What’s next?

  • Spectroscopic Follow-up: The team notes it is “absolutely necessary” to use JWST’s NIRSpec instrument to catch the chemical fingerprint of these galaxies and confirm their distances once and for all.
  • The MINERVA Project: In the upcoming “Cycle 4” of JWST observations, a massive new survey called MINERVA will scan an area 20 times larger than this study, using these same precision filters to find the brightest, rarest “first lights” in the dark.

A Final Thought

We are currently living through the “Golden Age” of discovery. For the first time in human history, we aren’t just guessing how the universe began; we are developed the “glasses” necessary to see it happen. As we refine our vision, the universe is proving to be even more complex—and more exclusive—than we ever imagined.


Publicly available preprint of paper