This Amazing Computer Chip Is Made of Live Brain Cells

18,192 6 Loading

A few years ago, researchers from Germany and Japan were able to simulate one percent of human brain activity for a single second. It took the processing power of one of the world’s most powerful supercomputers to make that happen.

Hands down, the human brain is by far the most powerful, energy efficient computer ever created.

So what if we could harness the power of the human brain by using actual brain cells to power the next generation of computers?

As crazy as it sounds, that’s exactly what neuroscientist Osh Agabi is building. Koniku, Agabi’s startup, has developed a prototype 64-neuron silicon chip.

Their first application? A drone that can smell explosives.

A drone with a bee’s sense of smell

A bee can navigate incredibly well because of its powerful ability to detect and interpret smells. Agabi was fascinated by this and wondered, “What if we could extract just the part responsible for that in the bee and put it in a drone? Suddenly we have a drone which has a sense of smell which equals that of bee.”

Such a drone would be able to smell bombs several kilometers away, says Agabi. It could also be used for surveying farmland, refineries, manufacturing plants — anything where health and safety can be measured by an acute sense of smell. There are no silicon devices which are able to give us the level of sensitivity that we find in biology.

How does the Koniku chip actually work? Let’s dig into it.

Baby steps towards biological computing

To create their first working prototype, Agabi and his team had to solve three big problems. They had to learn to:  

  • Structure neurons the same way they are structured in our brains
  • Read and write information in single neurons
  • Create an environment to keep the neurons stable

Using induced pluripotent stem cell technology—a method in which adult cells (from the skin, for example) are genetically reprogrammed into a stem cell—any cell can be turned into a neuron. But what do you do once you have a living neuron? Like electronic components, live cells need a specific environment to be able to operate.

By creating a special shell for each neuron, the Koniku team is able to control the temperature and pH levels and can also send nutrients to the neurons to keep them alive. They are also able to control how the neurons communicate with each other in these shells.

An electrode under the shells enables information to be read and written into the neurons. Agabi describes the process like this, “We coat the electrodes with DNA and enriched protein which encourages the neurons to form an artificial tight junction with the electrodes. That way, we can read the information from the neurons. We can write the information into neurons using the same electrodes or using other means, e.g. light or chemical.”  

The Intel of wetware

Agabi believes that harnessing the power and efficiency of the human brain is the future of computing.“There are no practical limits to how large we can make our devices or how much we can engineer our neurons. I believe as intelligent computation goes—biology is the ultimate frontier.”

He’s not shy about his ambitions either, “We want to be the Intel of wetware...”

What’s the future?

With 64 neurons, a drone can have a powerful sense of smell. What’s next? By Agabi’s approximations, a Koniku chip with….

  • 500 neurons can power a driverless car
  • 10,000 neurons enables real-time image processing, at the level of the human eye
  • 100,000 neurons enables robotics with multiple sensory inputs
  • 1 million neurons will give us a computer that can think for itself.

Now, think about this: there are 100,000 neurons in a piece of brain matter the size of a grain of sand.

It’s not hard to see the potential here, yet there are sure to be significant challenges along the way. There has been much work done in attempting to build a brain-inspired machine but that’s very different than actually using biological matter like live neurons for computing at scale.  

As magnificent as our advances in technology are— biology is a far more advanced technology than anything ever created by man. It is only in striving to replicate biological processes that we understand how much we have to learn from nature’s billions of years of evolution.

Follow me

Sveta McShane

Sveta writes about the intersection of biology and technology (and occasionally other things). She also enjoys long walks on the beach, being underwater and climbing rocks. You can follow her @svm118.
Follow me

Discussion — 6 Responses

  • Manu Herrán March 17, 2016 on 11:31 am

    Sounds like a big risk of spreading sentience and pain, to take into account if we want to reduce suffering in the universe

    • Sveta McShane Manu Herrán March 17, 2016 on 1:20 pm

      Yes, the ethics of something like this are complicated and something I didn’t get to touch on in this article. Thanks for bringing it up, curious what others think too…

  • DSM March 17, 2016 on 10:43 pm

    There is a limit to the (neurological) device size, brains may have synchronisation problems if they grow much larger than they are now. However that still allows for a genius level bio-computer, and there is always the option of a hybrid system with optical interconnects, a meta brain of brains. One piece of the conceptual puzzle is still missing though, a universal ontology, nobody has been able to assemble all of the concepts within all known human languages into a single coherent hierarchy, which would be required for linked brains to operate in concert with maximum speed.

  • Kit March 19, 2016 on 3:02 pm

    I think that this is a step in the right direction and a hugely exciting area but can a 64 neuron system perform anything particularly well? A honeybee has around 1 million neurons and I’m sure many more than 64 of those are dedicated to olfactory processes.

    According to this wikipedia list (https://en.wikipedia.org/wiki/List_of_animals_by_number_of_neurons) a roundworm is the simplest brain and that has 302 neurons.

  • Emanuele Rossi March 20, 2016 on 1:33 am

    Hi, according to the Integrated Information Theory http://integratedinformationtheory.org/ developed by prof. Tononi (Wisconsin-Madison University), every system which can integrate information can generate consciousness. In spite of the level of consciousness generated I think that developing brain-like computing systems could allow also to discover, and/or better understand, consciousness patterns and dynamics. Obviously this could lead to questioning the ethics of this kind of development. I am not worry about research in this direction, actually I truly find it fascinating and useful; I am only worry about developing technologies based on something we do not completely understand. This means to me that we are not able to forecast the consequences of these kind of research. I am well aware that innovation in science are very often borderline but this could lead to profound discussions over own nature as human beings. Not being able to understand how our brain works could lead us to problems we are not yet able to see, nor maybe (culturally) ready to face. Thank you for this article and for reading these few lines.

  • Mike LaMartina March 22, 2016 on 11:31 am

    I can’t help but think how vital this technology could be; especially today in the wake of the bombings in Belgium. Research money should be funneled to labs like Dr. Agabi’s before the insanity of fanatical suicide bombers makes our world unlivable.