Light Patterns That Can Dispel the Twilight of the Mind

First, let's talk about lasers. Essentially (if you bear with me), a laser is monochromatic, coherent electromagnetic radiation. In scientific terms, it's a beam of light of a very specific color, polarized in a specific way (and if you look at it through polarized glasses, you'll have to turn your head at just the right angle to see it).

One of the main advantages of these characteristics is that a laser beam is highly focused and doesn't lose energy as much over distance as non-laser beams. This is extremely useful in a wide variety of fields, from laser vision correction and fiber-optic communications to industrial manufacturing. But there's one small problem.Suppose you want to use a laser as a kind 

Підписуйтеcь на наш Telegram-канал Lenta.UA - ЄДИНІ незалежні новини про події в Україні та світі

of advanced radar, scanning a specific surface and collecting the reflected beam using precision electronics. Given the extremely short wavelength of a laser beam, any reflective surface will appear rough and cause the beam to be reflected at an angle. This causes the beam to interfere with itself, much like the waves from a stone thrown into a lake interfere with themselves when reflected off a dock. This interference can be constructive (meaning the reflected light is brighter than the original beam) or destructive (meaning the reflected light is dimmer than the original beam). In either case, the resulting image is not as sharp as desired. This phenomenon is known as "speckle."

This creates problems even if the reflective surface is completely static. But what if it vibrates? Speckle patterns also become dynamic—a kind of dancing spot of light. And suddenly, this problem opens up interesting applications. For example, suppose a laser beam is reflected off a window behind which a conversation is taking place in the room. Sound waves vibrate the window, and these vibrations are reflected in dancing spots of light, which are speckle patterns. This isn't a hypothetical application, but a very real one. A very similar system was used by the KGB as early as 1947 to spy on Western embassies in Moscow.

Fast forward to today, and speckle patterns are beginning to open a whole new field of research related to brain imaging. This breakthrough work is being conducted in the photonics lab led by Professor Ze'ev Zalevsky at Bar-Ilan University. We spoke with Natalia Segal, a doctoral student leading this work.

Let's switch gears for a moment to brain research. One of the main methods scientists use to study the human brain is called fMRI, or functional magnetic resonance imaging. This method uses cutting-edge physics to detect minute changes in blood flow within the brain. By knowing that a certain area of ​​the brain receives more blood flow at a particular moment, scientists can determine that a particular stimulus—be it an image, a sound, or even a cognitive process—activates that area. This way, a brain map is gradually created. What's the problem with this approach? MRI machines are very expensive, claustrophobic, and unsuitable for many types of research.

Speckle patterns come to the rescue. When blood rushes to a specific area of ​​the cerebral cortex—the part of the brain located directly under the skull—this flow causes tiny vibrations in the skull. These vibrations, with an amplitude of thousandths of a millimeter, are invisible to the human eye. But shine a laser beam at the skull, capture the reflected light with a high-speed camera, and you'll get speckle patterns that the human eye still can't interpret (it only sees moving spots of light). This is where Natalia's experience and expertise in machine learning come in handy.

Video is just that: video. Using methods developed in recent years for video analysis, as well as several of her own analysis techniques, Natalia Sigal can determine which videos with speckle patterns correspond to specific situations processed by the brain. For example, in one early experiment, subjects were played recordings of either intelligible speech or speech in an unfamiliar language. A machine learning model trained on these videos learned to distinguish between the two. A laser was aimed at a region of the skull behind which lies Wernicke's area, the region of the cerebral cortex responsible for speech recognition. This is both intriguing and surprising, but what real-world applications do the researchers envision for this technology?

A later experiment did the opposite. Instead of focusing on speech recognition, an attempt was made to shed light on speech generation. Subjects thought of a specific word (by silently saying it)—"yes" or "no"—and the model learned to distinguish between the two. This brings us to one of the most exciting areas of research of our time: the brain-computer interface, or BCI for short. This is the ability for humans to interact with computers directly—with the power of thought. Our readers may have heard of Elon Musk's company Neuralink, one of the pioneers in this field. Most approaches to BCI involve invasive procedures—essentially implanting electrodes under a person's skull and attaching them directly to the brain. But not with speckle patterns, which can be read not only non-invasively but also contactlessly (at a distance), and inexpensively.

Of course, there's still a long way to go before this technology reaches maturity and allows people to communicate freely and fully with the world. But when it does, the possibilities will be astounding: from enabling completely paralyzed people to communicate with their relatives to providing brain researchers with an accessible and convenient way to unlock the mysteries of the mind. In the coming years, Natalia Segal and her team plan to build on the foundation they've already established, and we'll continue to follow her exciting research.

Fig. Setup that worked for the language comprehension

Fig. Lightweight deep neural network used for the comprehension task. The upcoming BCI study employs a more advanced architecture - stay tuned for the forthcoming publication.

Fig. Speckle patterns examples

Читайте также: Deciphering Thought: The Silent Revolution of Light