There’s a lot of concern about robots replacing humans, but it might be the machines that we should be worried about.
Swedish scientists have developed the world’s first “living computer” made from human brain tissue.
It consists of 16 lab-grown organoids, or clumps of brain cells, that send information to each other.
These work in much the same way as traditional computer chips, sending and receiving signals through neurons that act like circuits.
But what makes these neurons special is that biological neurons use a million times less energy than the digital processors currently in use, making these biological machines less energy intensive.
Compared to the world’s best computers, such as the Hewlett Packard Enterprise Frontier, scientists found that to achieve the same speed and 1,000 times the memory, the computer uses only 21 megawatts, compared to 10 to 20 watts for the human brain.
One megawatt is equal to one million watts.
Click here to resize this module
This living machine was developed by scientists at FinalSparks, who focus on creating solutions using biological neural networks.
Dr Fred Jordan, co-CEO of FinalSpark, a startup focused on developing solutions using biological neural networks, told DailyMail.com: This idea is common in science fiction, but has rarely been explored in practice.”
Organoids are small, self-organizing, three-dimensional tissue cultures created from stem cells.
Such cultures can be created to recapitulate much of the complexity of an organ or to represent selected aspects of an organ, such as generating only certain cell types.
Scientists take the stem cells and culture them for about a month until they develop neuron-like characteristics.
FinalSparks’ mini-brain was constructed from an estimated 10,000 living neurons, each about 0.5mm in diameter.
The organoids are trained with dopamine and are rewarded with large amounts of the chemical when they perform a task correctly.
Scientists administer dopamine by exposing specific areas of the brain organoids to light, similar to how dopamine is released when certain areas of the human brain are activated.
The mini-brain is surrounded by eight electrodes that measure the organoids’ activity, allowing researchers to send electrical currents through the electrodes to affect neurons.
These electrodes serve the dual role of stimulating the organoids and recording data for processing by the organoids.
The organoids are housed within a microfluidic incubator that acts as a mini plumbing system delivering tiny amounts of fluid to nourish the cells, providing them with the nutrients they need to survive.
The incubator keeps the organoids at body temperature, automates the flow and maintenance of cell media, and provides a stable, bacteria- and virus-free environment.
The cells of this “living computer” aggregate into 3D organoid structures and undergo repeated cycles of life and death within 100 days.
But they resemble those of real human brains and have similar electrical activity.
“The lifespan of neurons in the brain is about 80 years. When we die, we have the same neurons as when we were born. Humans cannot keep neurons alive as long as nature does, so the lifespan of a neuron is 100 days.”
Scientists then simply grow new organoids to replace the dead ones.
The team recently launched the Brain Computer, an online platform that allows researchers around the world to remotely conduct in-vitro experiments on living neurons.
More than 30 universities have already expressed interest in using the platform.
Jordan described his “living computer” as “wetware” because, like a real human brain, it sits somewhere between computer hardware (i.e. the chips that process information) and software (the programs that run on the hardware).
“We call it ‘wetware’ — I don’t know who coined that term — but the brain is somewhere between the software and the hardware,” Jordan said.
“In computers, there is a clear distinction, and we run different software on the same hardware.
“But in our brains, to learn something we need to physically change the hardware that forms the synaptic connections, so we need a new word, and the word ‘wetware’ makes sense because cells need a moist environment to survive.”
The world is in the midst of an energy crisis due to fuel shortages, supply chain disruptions, geopolitical tensions and a transition to renewable energy.
Not to mention, the rise of AI is predicted to consume 29.3 terawatt-hours of electricity per year (1 terawatt is equal to 1 trillion watts).
What’s the next step?
For now, the company’s focus is on cloud computing, particularly the energy-hungry data centers that power AI.
Labs and universities around the world are already connecting to FinalSpark and testing their hardware.
Later this month, Jordan will be attending the world’s first “biocomputing” conference in Vienna this month, with participants from Australia and the U.S. “I don’t think the organisers realised how small the turnout would be,” Jordan laughed.
He said his phone has been ringing off the hook since he made his findings public, adding: “Lots of people are calling me offering me money.”
Another potential breakthrough is a new understanding of the human brain itself, which could lead to treatments for disease.
The technology to create organoids is relatively new, and the idea of turning human neurons into computers has been little explored until now.