What if the key difference between human intelligence and AI isn’t computing power, but the chemistry of memory? New research proposes that the brain may not simply transmit information electrically, it may also “write” and “read” memories chemically, through interactions among neurons, astrocytes and their surrounding molecular environment. The theory connects memory with emotion and thought, raising a timely question as AI grows more sophisticated: Can computation alone ever produce the kind of consciousness generated by a living brain?

A familiar song can bring back a memory from decades ago. A particular smell can recall a person or place. And a frightening memory can trigger a simultaneous emotional and physical response almost instantly.

Computers, too, can store and retrieve information. But is biological memory fundamentally different?

A new study by Dr. Gerard Marx of MX Biotech Ltd. and Prof. Chaim Gilon of the Hebrew University of Jerusalem, Dept. Chemistry, approaches this question by examining the physical and chemical processes that may connect memory, emotion and thought.

In Consciousness, Memory and Intelligence: A Mechanistic Perspective, the researchers propose that understanding memory requires looking beyond neurons and electrical signaling to include astrocytes and the complex chemical environment surrounding brain cells.

The study builds on Marx and Gilon’s earlier work on the molecular basis of memory and extends it to a larger question: Could the mechanisms that allow the brain to create and retrieve memories also help us understand thought, intelligence and consciousness?

A chemical code for memory

At the center of their research is what they call the Tripartite Mechanism of Memory.

The model proposes that memory arises through interactions among three elements: brain cells, particularly astrocytes interacting with neurons; the molecular material surrounding those cells; and chemicals including metal ions, neurotransmitters and gliotransmitters.

In simpler terms, the researchers suggest that the brain may do more than transmit information through electrical signals between neurons. It may also write and read information chemically.
Astrocytes, abundant brain cells that interact closely with neurons, have a particularly important place in the model. Marx and Gilon propose that clusters of astrocytes participate in processing cognitive information and interact with both neurons and the extracellular environment surrounding brain cells.

Their latest paper takes this proposed memory mechanism a step further. If experience can be physically encoded as memory, they ask, could retrieving and recognizing that information, and giving it meaning, contribute to what we experience as thought?

Memory comes with emotion

A key element of the researchers’ argument is that biological memories are not simply pieces of stored logical information.

A childhood smell can evoke happiness. A familiar voice can bring comfort. A frightening memory can cause an immediate physical reaction.

Marx and Gilon argue that any physical explanation of memory must therefore take this emotional context into account. Their model includes chemical messengers, such as neurotransmitters and other molecules, as part of the proposed mechanism by which cognitive information is encoded.

This connection between memory and emotion leads the researchers toward the broader questions of thought, intelligence and consciousness.

What does this mean for AI?

The comparison with artificial intelligence provides a timely test of these ideas.

Both computers and brains can store and retrieve information. In a computer, information ultimately has a physical aspect encoded in its hardware. Marx and Gilon ask whether biological memory also has a physical aspect, but one based on the cells and chemistry of living brain tissue rather than on electronic components.

That distinction becomes important as AI systems become increasingly capable of performing tasks associated with human logical intelligence. Their paper challenges the assumption that increasing computational power and complexity alone would necessarily produce biological-like consciousness. The researchers argue that electronic systems lack the biological infrastructure and emotional processes characteristic of living organisms.

Their argument is not that AI cannot perform increasingly sophisticated intellectual tasks. Rather, it raises a more fundamental question: Is displaying intelligent behavior the same as possessing the biological processes that give rise to mentality manifest as memory, emotions and conscious experience?

From memory to thought
Marx and Gilon do not claim to have solved the mystery of consciousness. Their paper explicitly acknowledges that conscious life remains an enigma.

Instead, they propose a biochemical process connecting brain cells and chemistry to memory. Their tripartite mechanism may offer a new route for considering how mental processes are actuated.

Their earlier work asked how memory might physically be encoded in the brain. Their new study broadens the question:

If we can understand how a living brain turns experience into memory, that could that bring us closer to understanding how memory becomes thought—and what ultimately distinguishes biological intelligence from artificial intelligence.