The boy whose brain could unlock autism

SOMETHING WAS WRONG with Kai Markram. At five days old, he seemed like an unusually alert baby, picking his head up and …

 “There are elements of the intense world theory that better match up with autistic experience than most of the previously discussed theories,” says Ari Ne’eman, president of the Autistic Self Advocacy Network, “The fact that there’s more emphasis on sensory issues is very true to life.”


“We agree that autistic people do have a number of cognitive advantages and it’s valuable to do research on that,” he says. But, he stresses, “People have worth regardless of whether they have special abilities. If society accepts us only because we can do cool things every so often, we’re not exactly accepted.”

Thinking it through: Scientists seek to unlock mysteries of the brain

Understanding the human brain is one of the greatest challenges facing 21st century science. If we can rise to this challenge, we will gain profound insights into what makes us human, develop new treatments for brain diseases, and build revolutionary new computing technologies that will have far reaching effects, not only in neuroscience.

Scientists at the European Human Brain Project—set to announce more than a dozen new research partnerships worth Eur 8.3 million in funding later this month—the Allen Institute for Brain Science, and the US BRAIN Initiative are developing new paradigms for understanding how the human brain works in health and disease. Today, their international and collaborative projects are defined, explored, and compared during “Inventing New Ways to Understand the Human Brain,” at the 2014 AAAS Annual Meeting in Chicago.

Brain Simulation, Big Data, and a New Computing Paradigm

Henry Markram from the Ecole Polytechnique Fédérale de Lausanne (EPFL), in Switzerland, where the Human Brain Project is based, describes how the project will leverage available experimental data and basic principles of brain organization to reconstruct the detailed structure of the brain in computer models. The models will allow the HBP to run super-computer based simulations of the inner working of the brain.

“Brain simulation allows measurements and manipulations impossible in the lab, opening the road to a new kind of in silico experimentation,” Markram says.

The data deluge in neuroscience is resulting in a revolutionary amount of brain data with new initiatives planning to acquire even more. But searching, accessing, and analyzing this data remains a key challenge.

Sean Hill, also of EPFL and a speaker at AAAS, leads The Neuroinformatics Platform of the Human Brain Project (HBP). In this scientific panel, he explains how the platform will provide tools to manage, navigate, and annotate spatially referenced brain atlases, which will form the basis for the HBP’s modeling effort—turning Big Data into deep knowledge.

The Neuroinformatics Platform will bring together many different kinds of data. University of Edinburgh’s Seth Grant, a key member of the HBP, describes how he is deriving new methods to decode the molecular principles underlying the brain’s organization, such as how individual proteins assemble into larger complexes. As Grant explains in Chicago, this has important practical applications as many mutations in schizophrenia and autism converge on these so-called supercomplexes in the brain.

As we understand more and more about the way the brain computes we can apply this knowledge to technology. Karlheinz Meier, of Heidelberg University in Germany and a speaker at AAAS, outlines how he is working to create entirely new computing systems as part of the HBP. These Neuromorphic Computing Systems will merge realistic brain models with new hardware for a completely new paradigm of computing—one that more closely resembles how the brain itself processes information.

“The brain has the ability to efficiently perform computations that are impossible even for the most powerful computers while consuming only 30 Watts of power,” Meier says.

Brain: Get Ready For Your Close-up

At AAAS, Christof Koch lays out another ambitious, 10-year plan from the Allen Institute for Brain Science: to understand the structure and function of the brain by mapping cell types from mice and humans with computer simulations and figuring out how the cells connect, and how they encode, relay, and process information. The project, Koch says, promises massive, multimodal, and open-access datasets and methodology that will be reproducible and scalable.

At Harvard University, George Church is participating in the Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative, which aims to map every neuron in the brain with rapidly advancing technologies. At AAAS, he describes progress on new tools for measurements of brain cell development, connectivity, and functional state dynamics in rodent and human clinical samples.

What do all of these projects have in common? They seek to help find some of the most elusive answers known to man: what makes us human, how does the brain function, what causes neurological and mental illness, and, most importantly, how can we treat or cure these afflictions?

Artificial Brains


External image

At what point will a program make a break from its human origins and become something independent and just plain weird? Intellectual property makes assumptions about human authorship and fixation that are becoming obsolete with new computing models. While self-aware artificially intelligent robots may still be pretty far off, there are some very strange things going on in experimental computing.

As Dorothy might say, “I’ve a feeling we’re not in Kansas anymore.”
  Artificial intelligence researchers have posited different perspectives on what it means to be “creative.” In many ways, it involves the ability to take in input and process it in a way that results in a novel combination of pre-existing ideas and information. Some forms of creative thought are more innovative and human than others. For example, once a given conceptual space is established an entity may process pre-existing ideas in a creative way, evidencing “exploratory creativity.” On the other hand, higher or “transformational creativity” is achieved when the conceptual input is transformed by rejecting or redefining previous conceptual constraints to produce creative output that transforms our understanding, raising it to a new level. Watson (IBM’s AI Jeopardy! champion) showed a high degree of exploratory creativity by being able to process massive amounts of cultural information in a human-like way. It had to zero in on a few essential key words, executed thousands of language analysis algorithms in parallel, then would buzz in when the possibilities had been narrowed down to an answer for which it was sufficiently “confident.” Also interesting is that when Watson’s answer was off, it tended to be hilariously off, which reflects the difficulty of sorting associations in a human-like way.
In 1950 Alan Turing proposed the “Turing test” to test a machine’s ability to appear human. Participants would converse with the machine or a human in a text-only format. They would then indicate if they believed they were communicating with a human or with a machine. The machine would pass the test if it could generate answers that were indistinguishable from a real human. Essentially, a program may be considered somewhat artificially intelligent if it seemed like a human to other humans. Watson may be considered a highly developed example of artificial intelligence according to his mastery of human-like thought and language. Yet while this is an impressive feat of computing, Watson did not in fact learn or rewrite its underlying programming or evidence any degree of transformational creativity. This does show that super-computers are getting powerful enough to process information quickly enough to result in human-like output in response to input.

Neurologists are in the process of reverse-engineering a biologically accurate brain down to the molecular level using a supercomputer. Headed by Henry Markram, the Blue Brain Project is programming elements of the supercomputer to model individual neurons, building to the various brain regions and intends to model the complete human brain within ten years. This is particularly tricky and computation-intensive due to all the interconnection and feedback loops within a human brain, in direct contrast to input-output models which are the industry standard.

I could go into neocortical columns and interconnected influence of one brain region on another, but more interestingly is the question: if a simulated brain DID behave just like a human one, would it be self-aware? (this is a legal mess which I may deliberate upon some other day) Computers are being built which are able to act and literally think like humans. Some very interesting developments in programming are behind these developments. In addition, some computing methods which do not mimic human action at all behave in a way that calls into question legal assumptions about creativity and human authorship.   

In conclusion - some more from the Wizard of Oz:

“Scarecrow: I haven’t got a brain… only straw.

Wizard of Oz: Why, anybody can have a brain. That’s a very mediocre commodity. Every pusillanimous creature that crawls on the Earth or slinks through slimy seas has a brain. Back where I come from, we have universities, seats of great learning, where men go to become great thinkers. And when they come out, they think deep thoughts and with no more brains than you have. But they have one thing you haven’t got: a diploma.”

Neuroscience: Where is the brain in the Human Brain Project?

Launched in October 2013, the Human Brain Project (HBP) was sold by charismatic neurobiologist Henry Markram as a bold new path towards understanding the brain, treating neurological diseases and building information technology. It is one of two ‘flagship’ proposals funded by the European Commission’s Future and Emerging Technologies programme (see Selected after a multiyear competition, the project seemed like an exciting opportunity to bring together neuroscience and IT to generate practical applications for health and medicine (see

Contrary to public assumptions that the HBP would generate knowledge about how the brain works, the project is turning into an expensive database-management project with a hunt for new computing architectures. In recent months, the HBP executive board revealed plans to drastically reduce its experimental and cognitive neuroscience arm, provoking wrath in the European neuroscience community.

The crisis culminated with an open letter from neuroscientists (including one of us, G.L.) to the European Commission on 7 July 2014 (see, which has now gathered more than 750 signatures. Many signatories are scientists in experimental and theoretical fields, and the list includes former HBP participants. The letter incorporates a pledge of non-participation in a planned call for 'partnering projects’ that must raise about half of the HBP’s total funding. This pledge could seriously lower the quality of the project’s final output and leave the planned databases empty.

With the initial funding, or 'ramp-up’, phase now in full swing, the European Commission is currently evaluating the HBP directors’ plan for the larger second part of the project. This offers an opportunity to introduce reforms and reconciliation. Here, we offer our analysis of how the HBP project strayed off course and how it might be steered back.

Read more

Building a Digital Piece of Brain

If you want to learn how something works, one strategy is to take it apart and put it back together again. For 10 years, a global initiative called the Blue Brain Project–hosted at the Ecole Polytechnique Federale de Lausanne (EPFL)–has been attempting to do this digitally with a section of juvenile rat brain. The project presents a first draft of this reconstruction, which contains over 31,000 neurons, 55 layers of cells, and 207 different neuron subtypes.

The research is in Cell. (full access paywall)

Research: “Reconstruction and Simulation of Neocortical Microcircuitry” by Henry Markram et al. in Cell doi:10.1016/j.cell.2015.09.029

Image: This is a photo of a virtual brain slice. Credit: Makram et al./Cell 2015.