Organoid Intelligence: Biocomputers Made of Brain Cells

Computers have relied on silicon chips for decades, but traditional hardware is approaching its physical limits. To solve complex problems and reduce energy consumption, scientists are turning to a powerful existing processor: the human brain. This field is called Organoid Intelligence (OI). It involves fusing lab-grown brain tissue with electronic chips to create “biocomputers” that could eventually outperform traditional supercomputers.

What Are Brain Organoids?

Before understanding the computer aspect, it is helpful to understand the biological component. Brain organoids are tiny, three-dimensional cultures of tissue derived from human stem cells. They are often called “mini-brains” in popular media, though they are only about the size of a pen dot.

Researchers at institutions like Johns Hopkins University create these by taking skin samples, reprogramming them into stem cells, and coaxing them to grow into brain cells (neurons). Unlike flat cell cultures in a petri dish, organoids have a 3D structure. This allows them to develop functional connections, or synapses, that mimic the architecture of a real human brain.

Currently, a typical organoid contains roughly 50,000 cells. While this is small compared to the 86 billion neurons in a human brain, it is enough to process basic signals and form networks.

The DishBrain Experiment: Playing Pong

The concept of biological computing moved from theory to reality with a project known as “DishBrain.” In 2022, researchers at Cortical Labs in Australia successfully grew a layer of human neurons on top of a multi-electrode array. They then connected this biological system to a computer simulation of the vintage video game Pong.

The results were startling:

  • Rapid Learning: The brain cells learned how to play the game in roughly five minutes.
  • Feedback Loops: The system used electrical pulses to tell the cells when they missed the ball (unpredictable stimulus) and when they hit it (predictable stimulus). The cells reorganized themselves to minimize the unpredictable feedback.
  • Comparison to AI: While a silicon-based AI can play Pong perfectly, it typically requires massive amounts of data and training time to learn the rules. The biological cells grasped the cause-and-effect relationship almost instantly.

Why Biology Beats Silicon

You might wonder why we need biological computers when our phones and laptops are already so fast. The answer lies in two areas: energy efficiency and learning capability.

The Energy Crisis in Computing

Modern supercomputers are incredibly powerful, but they are energy vampires. For example, the Frontier supercomputer in Kentucky exceeded the computational speed of a single human brain in 2022. However, achieving that speed requires 21 megawatts of power.

In contrast, the human brain operates on roughly 20 watts. That is barely enough energy to power a dim lightbulb. If scientists can harness the efficiency of biological neural networks, they could create AI systems that require a fraction of the energy used by current data centers.

Parallel Processing

Silicon chips process information linearly. They solve problems one step at a time, just very quickly. Brains process information in parallel. They can handle incomplete data, recognize patterns, and adapt to new situations much better than standard code. This makes organoid intelligence ideal for tasks that stump traditional computers, such as:

  • Complex pattern recognition.
  • Decision-making with fuzzy data.
  • Creative problem solving.

How Biocomputers Are Built

Creating a biocomputer requires a specific interface. You cannot simply plug a USB cable into a cell culture. The interface relies on Microelectrode Arrays (MEAs).

These are tiny grids of sensors and stimulators. The organoid sits directly on top of the MEA.

  1. Input: The computer sends electrical patterns through the grid to the organoid. This mimics sensory input (like seeing a ball move in Pong).
  2. Processing: The neurons react to the signal, firing synapses and communicating with each other.
  3. Output: The MEA sensors detect the neural activity and translate it back into digital signals the computer can understand.

The Roadmap: From 50,000 to Millions

Professor Thomas Hartung at Johns Hopkins University is leading the charge to scale this technology. He estimates that for OI to compete with standard computers, organoids need to scale up to around 10 million neurons.

There are significant hurdles to reaching this scale:

  • Blood Supply: In the human body, blood vessels deliver oxygen and nutrients to the brain. In a lab, organoids die from the inside out once they get too big because they lack this vascular system. Researchers are currently developing microfluidic systems to act as artificial blood vessels.
  • Memory: While neurons can react, scientists are still figuring out how to store long-term data in these biological networks.

Ethical Concerns and the Baltimore Declaration

Fusing human brain tissue with machines raises obvious ethical questions. Could these organoids develop consciousness? Could they feel pain?

Currently, the organoids are too primitive to possess consciousness or feelings. However, as they grow larger and more complex, the line blurs. To address this, Hartung and his team proposed the “Baltimore Declaration.” This is a framework calling for embedded ethics in the research process. It ensures that bioethicists are involved at every step of development, rather than waiting until the technology is fully mature to ask difficult questions.

Future Applications

While a commercially available biological laptop is likely decades away, Organoid Intelligence has immediate applications in medicine:

  • Personalized Medicine: Doctors could grow organoids from a patient’s own skin cells to test which drugs work best for their specific genetic makeup.
  • Neurological Research: Instead of testing Alzheimer’s or autism treatments on rats (which have very different brains), researchers can test them on human brain organoids.
  • Toxicology: Companies can test chemicals or pesticides on organoids to see if they damage human cognitive function without risking human subjects.

Frequently Asked Questions

Are biocomputers faster than regular computers? Not yet. Silicon computers are still much faster at simple math and logic. Biocomputers excel at learning, pattern recognition, and energy efficiency. The goal is to create hybrid systems that use silicon for math and biology for complex reasoning.

Where do the brain cells come from? They come from induced pluripotent stem cells. Scientists usually take a skin sample from a donor, reverse-engineer the cells into a stem-cell state, and then reprogram them to grow into neurons. No embryos are used in this specific process.

Is Organoid Intelligence the same as AI? No. Artificial Intelligence (AI) runs on code and silicon chips. Organoid Intelligence (OI) runs on actual biological tissue. However, the two fields may merge to create hybrid AI systems.

Can an organoid feel pain? Current organoids do not have the complex structure or sensory inputs required to feel pain or suffer. They are essentially biological processors. Ethical guidelines monitor this closely as the technology advances.