For decades, scientists have tried to peer deep inside the human brain using beams of harmless near-infrared light. The technique, called functional near-infrared spectroscopy, or fNIRS, has become a workhorse for neuroscientists. It’s affordable, noninvasive, and portable—qualities that bulky MRI machines lack. But fNIRS has always faced a stubborn wall: it simply can’t see more than a few centimeters below the surface of the skull.

Now, researchers from Glasgow have done what was once considered impossible. They’ve detected photons that traveled from one side of an adult human head to the other—an optical journey spanning over 15 centimeters and requiring passage through skull, brain, and bone. The feat could eventually transform the way we look into the brain, offering a new path to image its most mysterious and hard-to-reach regions.

“[This work] explores the limits of photon transport in the brain,” said Daniele Faccio, a physicist at the University of Glasgow and senior author of the new study, published in Neurophotonics.

A Light That Doesn’t Bounce Back

Traditional fNIRS systems measure how light bounces back from the brain’s surface. Light is shone onto the scalp, where some of it penetrates and scatters. A nearby detector picks up the faint reflections. These signals reveal how oxygen-rich or oxygen-poor the blood is in outer brain regions, offering clues to brain activity.

But that strategy can only go so far because of how the skull is made. “The highly scattering nature of near-infrared light in human tissue makes it challenging to collect photons using source-detector separations larger than several centimeters,” the study explains.

In other words, light gets scrambled by bone and tissue. The deeper it tries to go, the more it scatters, fades, or is absorbed. Most researchers assumed that trying to collect light on the far side of the head—what’s known as diametrical transmission—was futile.

The math sure seemed to back this up. Calculations based on the brain’s optical properties predicted astronomical levels of attenuation. One estimate suggested light would be diminished by a factor of 10⁵³.

But Faccio and his team weren’t convinced.

A Path Through the Fog

To test the limits, the researchers started by firing a powerful, ultrafast laser into the side of a volunteer’s head. The beam, centered above the ear, was expanded into a broad circle to safely increase the light dosage without damaging skin.

On the opposite side of the head, they placed a photon detector inside a light-tight enclosure, designed to block out all ambient light and even reflections bouncing within the skin. The detector—more than twice as sensitive as those used in typical fNIRS systems—could count individual photons and record their arrival times with picosecond precision.

Then they waited.

Over thirty minutes of exposure, the system detected photons that had passed entirely through the head. These photons had not just skimmed the surface or scattered back—they had made the full, brain-spanning journey.

“The measured experimental attenuation was found to be of the order 10¹⁸,” the researchers reported. That’s about one photon detected per second, out of more than ten trillion sent. But it’s still something, and we can send trillions of photons into the skull.

To confirm what they were seeing, the team ran advanced Monte Carlo simulations of light traveling through a digital model of the human head. The simulations matched the data. They also revealed something surprising: light was not taking a random path through the brain. It was following specific routes—like water flowing through underground channels.

The brain’s anatomy itself guides these routes. Regions filled with cerebrospinal fluid—a clear, low-scattering liquid surrounding the brain—acted as light corridors. The light followed these low-resistance paths through and around the brain.

Is Deep Brain Imaging Closer Than We Thought?

The experiment was an extreme proof of concept. It only worked on one participant: a bald man with fair skin. Tests on seven others failed to detect a signal. And the setup—featuring a high-power laser, a sensitive photomultiplier tube, and 30 minutes of stillness—is far from practical for everyday brain scans.

But it is, with all these caveats, a proof of concept.

The implications are profound. If photons can make it across the adult human head, they may also be detectable in less extreme scenarios. Closer source-detector spacing, more efficient sensors, or longer wavelengths could help paint that picture more clearly.

In time, this approach might help develop better methods for detecting and tracking neurological conditions such as strokes, injuries, and tumors. In places without access to MRI or CT machines—such as remote clinics or battlefields—portable optical brain imaging could be a game-changer.

Even more tantalizing is the possibility of imaging deeper brain regions that fNIRS has long ignored: the midbrain, the sulci hidden in cortical folds, and the cerebellum. The simulations show that moving the light source to different positions on the head can preferentially illuminate different parts of the brain’s interior.

Future systems might take advantage of this by combining measurements from multiple angles to reconstruct a tomographic image—something like an optical CT scan.

Challenging the Boundaries of Brain Imaging

The study builds on a lineage of optical brain research that dates back nearly 50 years. In 1977, scientist Frans Jobsis first described how near-infrared light could noninvasively monitor oxygenation in living tissue. His early work hinted that light might travel through the head during hyperventilation, but the data were incomplete and never replicated in adults.

Since then, fNIRS has become widely used in neonatal care, cognitive neuroscience, and even psychology experiments. But its limitation to the brain’s outer shell has always been a barrier.

For now, transmitting light through a full adult head remains an experimental feat. But like the first photographs from the depths of the ocean or the early echoes of radar scans, it reveals that more is possible.