Proposing an Artificial Neural Networks-based Model to Simulate Human Visual Cortex

Imagine your brain as the ultimate, super-advanced computer, but instead of silicon chips and wires, it's built from billions of tiny, living cells called neurons. These neurons are like little messengers that talk to each other using electrical sparks and chemical signals. For a long time, scientists have been trying to build artificial computers that think like us, using "Artificial Neural Networks." Think of these as digital clones of our brain's wiring. Usually, we use these digital clones to do cool stuff like recognizing faces in photos or helping self-driving cars see the road. But this research asks a different, more curious question: What if we flipped the script? Instead of just using the computer to do a job, what if we used the computer to understand us? Specifically, how does our brain actually see the world, remember it, and sometimes get overwhelmed by it? By building a digital model of the brain's visual center, scientists hope to peek behind the curtain of human perception, understanding why we remember things the way we do and why certain conditions, like migraines, can make the world feel painfully bright.

---

The Digital Brain: A Story of Seeing, Remembering, and Migraines

In this study, a researcher named Sina Saadati proposes a way to build a "digital twin" of the human visual cortex—the part of your brain that turns light hitting your eyes into the pictures you see. The goal isn't just to make a better camera; it's to simulate how the brain feels and remembers what it sees. The paper suggests that by using specific types of computer networks, we can recreate how our brains process images, how they get messed up during a migraine, and how a song can suddenly bring a whole memory back to life.

Part 1: The Brain's Camera and the "Afterimage" Trick

First, the paper builds a model using something called a Convolutional Neural Network (CNN). You can think of a CNN as a team of tiny, specialized detectives working in layers. The first layer looks at simple things like edges and colors (like a pixel). The next layer looks at those edges and says, "Hey, that looks like a curve!" The layer after that might say, "That curve is part of a face!" This is how the human brain builds a picture from scratch, layer by layer.

The researchers used this digital team to simulate how we store memories. They found that when the model "stares" at an image for a while (which in computer terms means training on it), it changes its internal settings. If you then show it a totally new picture, the old memory still leaks out. It's like staring at a bright light and then closing your eyes; you still see the shape of the light for a moment. The simulation suggests that our brains work the same way: looking at something changes the physical structure of our brain cells (in the model, this is represented by changing the "weights" or importance of connections), which is why we can recall images even when they aren't there. The paper also notes that if you stare at something new for long enough, the old memory fades away, replaced by the new one, just like how our moods and memories shift over time.

Part 2: When the Brain's Alarm Goes Off (Migraines)

Next, the paper asks: What happens when the brain isn't working normally? The researchers decided to simulate a migraine attack. In real life, people with migraines often suffer from photophobia, which is an extreme sensitivity to light. A tiny lamp or a phone screen can feel like a blinding spotlight.

To simulate this, the researchers tweaked their digital brain. They increased the "sensitivity" of the neurons, which in the computer model meant turning up the volume on the connections between the cells. In the real brain, this is caused by a chemical called glutamate being released in higher amounts, making neurons fire too easily.

The result was fascinating. When the model was "normal," a dark room with a small laptop screen looked dark. But when the model was set to "migraine mode," that same tiny screen lit up the entire digital brain as if it were the midday sun. The simulation showed that even a weak signal gets amplified into a painful, overwhelming flash. The author suggests this model is reliable because it matches the real-life reports of people with migraines, who say that even the reflection of a light on a wall can feel like it's scratching their eyes. This digital experiment helps explain why the pain happens: the brain's alarm system is just turned up way too high.

Part 3: The Power of a Song (Audio-Visual Memory)

Finally, the paper explores how we mix different senses to remember things. Have you ever heard a song and suddenly remembered exactly where you were, who you were with, and what the room looked like? The researchers wanted to see if they could simulate this.

They added a second system to their model called a Hopfield Network. Think of this as the brain's "long-term storage" or a giant web where everything is connected to everything else. In this web, a visual memory (a picture) and an audio memory (a song) are tied together.

The experiment worked like this:

1. The model "watched" a scene (like a face or a tree) while "hearing" a specific sequence of musical notes.
2. It stored this combined experience in the Hopfield web.
3. Then, the researchers deleted the picture from the memory and only gave the model the music.
4. The Result: The model successfully "reconstructed" the picture just from the music!

In 37 different experiments, the model was able to recall the visual details with incredible accuracy. On average, it got 98.35% of the picture right. For example, if the model heard the music associated with a picture of a person wearing sunglasses, it could "draw" the sunglasses back into existence just from the tune. The paper suggests this proves that our brains don't store sights and sounds separately; they weave them together into a single, rich tapestry of memory.

What This All Means

This research doesn't claim to have built a real human brain or solved the mystery of consciousness. Instead, it suggests that by using these specific computer models, we can create a reliable way to simulate how our visual cortex works. It shows that:

• Memory is physical: Staring at something changes the brain's structure, which is why we remember it.
• Migraines are an amplification issue: The pain comes from neurons getting too sensitive, turning small lights into blinding flashes.
• Senses are linked: A song can trigger a visual memory because the brain stores them as one connected package.

The author is careful to say that while these simulations match real-world reports and psychological principles, they are still just models. They are a powerful tool for understanding the brain, but they are a map, not the territory itself. Future work might need to add more details, like how touch or pain signals fit into this picture, or how hormones might change the way the brain processes these signals. But for now, this digital journey into the visual cortex gives us a clearer, more vivid picture of how we see, remember, and feel the world around us.