The Brain According to Henry Markram (Video)

12 21 Loading
Henry Markram's computer model of a brain is built one neuron at a time.

Henry Markram's model of a brain is built one neuron at a time.

Once you start building a brain in a box you get two things: admission into the Mad Scientists’ Club, and a chance to speak at TED. Henry Markram is the director of the Blue Brain Project, a collaboration between European scientists and IBM that aims to construct a life-like simulation of a brain using a supercomputer. Earlier this year Markram spoke at TED Global discussing how most of human perception is based on decision making within the brain. BBP hopes it will only take another decade to create a fully functioning computer model of a human mind, and Markram will guide you through some of the most important concepts behind the simulated brain today. Check out the TED Global 2009 video after the break.

This is neither the first Blue Brain Project story, nor even the first video of Markram that we’ve discussed here at Singularity Hub. What keeps us coming back to the topic? The amazing possibilities created by the BBP research and the uncertainty as to whether it will ever succeed. While not aimed at creating an artificial intelligence, a fully functioning simulated brain would be hard to define as anything else. Even ignoring the AI applications, BBP could help neuroscientists learn what causes certain mental illnesses, or how to repair trauma to the brain, or how to improve the way the brain processes information. Right now, BBP is exploring different theories on how the mind works. Yet it’s still much too early to tell if Markram will ever be able to exploit those theories to create a full fledged simulation.


The BBP is focused on the neocortex and specifically the neocortical columns that provide the raw computational power that shapes our minds. These columns are where our higher reasoning and decision making abilities reside. Research into the neocortex not only informs Markram and his team in how to program their supercomputer, it also highlights the fundamental ways in which our minds interact with the universe. As Markram describes in the video, 99% of our perception is based on the decision making processes which analyze incoming sensory information. Our world is cognitively based on our choices. That’s powerful stuff.

Neuroscientists have spent the better part of the last 15 years learning how to build digital models of the branching and connectivity between neurons. In Markram’s discussion he mentions that there is a huge amount of neural diversity, that each cell in our brain is different in its orientation and connection from every other cell, and different from every other neuron in the brains of others. Yet despite this diversity, there are consistent patterns in how neocortical circuitry is wired. Hearing that, I wonder if the efforts to build devices to “read people’s minds” will be enhanced by BBP’s discoveries. Once we understand common patterns in neurons, couldn’t we look for those patterns and translate them as thoughts or emotions? What if those patterns are too complex, or if the diversity is too strong to create a general model of the brain? BBP has a long way to go in understanding the nature of neuron interactions.

Markram’s goals for modeling the electrical activity in the brain may lead to a better understanding of how real world objects are mapped into neuron connections and signals. Such “electrical objects” in the brain show us its perception of the outside world, and could also give us the key to programming the brain. Creating artificial electrical objects in the brain (however one might do that) could allow us to experience things without ever being near them. There are many possible applications of such a technique for virtual reality and accelerated learning. Those technologies would require a tremendous understanding of the brain, however, and it’s unclear how quickly Markram’s team will be able to define the way neurons map the real world.

It’s hard to know if the Blue Brain Project will achieve its goal of creating a life like simulation of a brain in just 10 years. The task is monumental, but if you listen to Markram you sort of believe that they are most of the way there already. Hopefully that confidence will be supported by many years of fruitful work in the near future, because his research is far from over. There are likely some functions of the brain than can ever be revealed by simulation. But as far as mad scientist projects go, this one makes me as hopeful as I am skeptical.

[photo credit: blue brain project]

Discussion — 21 Responses

  • Wolfgang November 13, 2009 on 4:24 pm

    What is this break you’re always talking about?

    • Aaron Saenz Wolfgang November 13, 2009 on 4:37 pm

      On the homepage, stories are show in abbreviated form with a “Read More” link to the rest of the article. The part of the story that appears after the “Read More” is considered to be “after the break”.
      Of course, when you are sent a link to the story, the break does not appear, so the phrase doesn’t make sense.

  • Wolfgang November 13, 2009 on 12:24 pm

    What is this break you’re always talking about?

    • Aaron Saenz Wolfgang November 13, 2009 on 12:37 pm

      On the homepage, stories are show in abbreviated form with a “Read More” link to the rest of the article. The part of the story that appears after the “Read More” is considered to be “after the break”.
      Of course, when you are sent a link to the story, the break does not appear, so the phrase doesn’t make sense.

  • Nick November 14, 2009 on 8:13 am

    This guy is going somewhere.

    Today’s supercomputers are going to be on everyone’s desktop in 10 years if Moore’s Law holds true – all this guy has to do is wait and the processing power will come.

    • popay Nick November 14, 2009 on 3:28 pm

      Except that Moor’s law has been slowing down for the last 10 years and even if you have the raw processing power of a human brain if you don’t have the algorithms it’s useless.

  • Nick November 14, 2009 on 4:13 am

    This guy is going somewhere.

    Today’s supercomputers are going to be on everyone’s desktop in 10 years if Moore’s Law holds true – all this guy has to do is wait and the processing power will come.

    • popay Nick November 14, 2009 on 11:28 am

      Except that Moor’s law has been slowing down for the last 10 years and even if you have the raw processing power of a human brain if you don’t have the algorithms it’s useless.

  • The Avenger November 14, 2009 on 4:25 pm

    “Except that Moor’s law has been slowing down for the last 10 years ”

    What? Can you back that up with some data?

  • The Avenger November 14, 2009 on 12:25 pm

    “Except that Moor’s law has been slowing down for the last 10 years ”

    What? Can you back that up with some data?

  • buy backlinks November 16, 2009 on 3:20 pm

    Only problem – Moores Law isn’t absolute. It has to slow down eventually. The only question is – when ?

  • buy backlinks November 16, 2009 on 11:20 am

    Only problem – Moores Law isn’t absolute. It has to slow down eventually. The only question is – when ?

  • Jacques Hughes November 16, 2009 on 11:14 pm

    After watching the presentation, my first thoughts were surprise at the visualisation of the synapses, just the sheer number and density, and I’m assuming that was just a single column.

    My next thought was that the discovery of which cells store which memories is too simplistic. I base this on the OGOD theory, one gene, one disease. It’s obviously a little more complicated than that. we may find that for a single memory, audio is one location, colour another, movement, yet another. I also imagine that there will be subsets of audio, colour and movement storage.

    The one thing this research should address, and perhaps the most relevant here, is the question of consciousness. If a replica of a human brain can be virtually created, and stimuli added, will we end up with a thinking entity?

    If the answer is yes, then a lot of philosophy goes out of the window, the old dualist/monist argument for example. Also, religion and gods are finished.

    If not, why not? Is consciousness a product only of biochemistry?

    I’ve got 10 years, I’ll be very interested to see how this turns out. I have my suspicions, they say god is dead.

  • Jacques Hughes November 16, 2009 on 7:14 pm

    After watching the presentation, my first thoughts were surprise at the visualisation of the synapses, just the sheer number and density, and I’m assuming that was just a single column.

    My next thought was that the discovery of which cells store which memories is too simplistic. I base this on the OGOD theory, one gene, one disease. It’s obviously a little more complicated than that. we may find that for a single memory, audio is one location, colour another, movement, yet another. I also imagine that there will be subsets of audio, colour and movement storage.

    The one thing this research should address, and perhaps the most relevant here, is the question of consciousness. If a replica of a human brain can be virtually created, and stimuli added, will we end up with a thinking entity?

    If the answer is yes, then a lot of philosophy goes out of the window, the old dualist/monist argument for example. Also, religion and gods are finished.

    If not, why not? Is consciousness a product only of biochemistry?

    I’ve got 10 years, I’ll be very interested to see how this turns out. I have my suspicions, they say god is dead.

  • Smart Brain February 5, 2010 on 12:40 pm
  • Smart Brain February 5, 2010 on 12:50 pm

    Correction to the first link -

    http://neuroinformatics2008.org/congress-movies

  • Smart Brain February 5, 2010 on 12:55 pm

    The link doesn't seem to work from the message, click on my name to go to the lecture.

  • Smriti March 22, 2010 on 11:38 am

    Can anyone tell me that what are the mathematical forms or equations that are primarily fed to the computers? I mean if i excite the brain by showing a flower then what are the forms of equations or mathematical equations that the computers are fed after which the supercomputer is carrying the program?? please help me with this.

  • Smriti March 22, 2010 on 7:38 am

    Can anyone tell me that what are the mathematical forms or equations that are primarily fed to the computers? I mean if i excite the brain by showing a flower then what are the forms of equations or mathematical equations that the computers are fed after which the supercomputer is carrying the program?? please help me with this.

  • shrdlu April 24, 2010 on 3:07 pm

    Interesting simulation, I like the idea.

    I had to wonder how Markram got through a 15 minute lecture on the neocortex without talking about memories, learning, hierarchy, temporal/sequential processing, microcolumns, layers, (uncontroversial structural/theoretical constraints) – frankly his theory of the underlying algorithmic information processing operations of the brain seemed underdeveloped. I’d like to read a paper on his theory if he has one.

    Still, the mission to eliminate animal testing is noble indeed. Go Henry!

  • shrdlu April 24, 2010 on 11:07 am

    Interesting simulation, I like the idea.

    I had to wonder how Markram got through a 15 minute lecture on the neocortex without talking about memories, learning, hierarchy, temporal/sequential processing, microcolumns, layers, (uncontroversial structural/theoretical constraints) – frankly his theory of the underlying algorithmic information processing operations of the brain seemed underdeveloped. I’d like to read a paper on his theory if he has one.

    Still, the mission to eliminate animal testing is noble indeed. Go Henry!