Good progress as Human Brain Project celebrates first birthday
The EU’s €1bn project to simulate the human brain is making good progress as it reaches its first anniversary, organisers say.
Breakthroughs in high-performance computing, brain imaging and medical informatics were outlined at the opening of the second annual summit of the Human Brain Project (HBP) at Heidelberg University, a year after the project’s launch.
It was selected as one of the European Commission’s two 10-year Future and Emerging Technology flagship projects last January by a group of 25 world-renowned experts, alongside another to commercialise the use of wonder material graphene.
The project’s aims – to create a neuron for neuron digital replica of the human brain on supercomputers to aid neuroscience research – have been criticised from some quarters for being overly ambitious, but with the venture having met 90 per cent of its targets to date director Henry Markram is confident of their progress.
“The plan we set out on day one is being executed remarkably well,” Markram, also a professor at the École Polytechnique Fédérale de Lausanne (EPFL), told journalists at a press conference yesterday evening. “Some (sub-projects) are even seeing steps we didn’t think would happen in the first two to three years.”
One of the major success stories has been the creation of a biologically realistic, three-dimensional representative model of a rat cerebellum microcircuit – one of the most challenging brain regions to mode as it represents 50 per cent of the brain’s neurons despite only accounting for 10 per cent of the volume.
The HBP is already routinely running simulations of a rodent neocortex with three million morphologically detailed neurons on supercomputers at the Swiss National Supercomputing Centre (CSCS), as well as simulations of up to one billion more simple point neurons at the Forschungszentrum Jülich research centre in Germany.
The memory footprint of these simulations has been reduced by a factor of six since the start of the project and real-time streaming between CSCS and EPFL has been achieved to allow ‘analysis on the fly’ of data created by the simulations, a system the project leaders have labelled interactive supercomputing.
“It’s a completely new form of supercomputing that I think has been pioneered by this project,” said Karlheinz Meier, the HBP’s co-director in charge of computing and a professor at Heidelberg.
Despite the progress, the venture attracted less positive coverage in July after several hundred neuroscientists signed an open letter criticising the project for its narrow focus, overambitious aims and poor governance, and the topic came up repeatedly on the first day of the summit.
“Debate and differences of opinion is a sign of change,” Markram told attendees. “It’s very important so we welcome it in a way. We need to embrace it and work closely together with our colleagues who have different views.”
But while a mediator has been brought in to smooth over disagreements and commitments were made to open dialogue with key detractors and create community forums, the phrase “manage expectations” was repeated several times.
“It’s a strategy and it’s a vital strategy, but it’s not a science-fiction strategy of a human brain that can sit having a conversation,” said Markram.
“You have to realise we are spending €7bn a year on the brain already today and we are not making significant progress in terms of treatments and diseases. We do need to do a phase shift if we are going to make a difference.”
The open letter suggested that EU officials reviewing the plans and funding for the project’s operational phase, due to start in 2016, would find “substantial failures”, but the European Commission’s director of ‘Excellence in Science’ Thierry Van der Pyl told delegates that the HBP’s proposal had been “favourably evaluated”.
“Indeed it’s risky, but how often has Europe been accused of not being ambitious,” he said. “I am aware that some described the goals as over ambitious, but the flagships are not business as usual.”
But better communication is needed, Van der Pyl told delegates, and making the governance of the project more inclusive is a major challenge that needs to be addressed by the board.
The project’s third co-director Professor Richard Frackowiak pointed out that, like all aspects of the project in its ramp-up stage prior to full operation, governance is dynamic and the board were happy to make changes.
“To be told that the governance needs looking at for the next phase is perfectly natural,” he said. “There’s this general feeling of crisis that I don’t feel. I feel we are seeing the natural evolution of a really big, really ambitious project.”