22nd June 2013

BigBrain: An ultra-high resolution 3-D roadmap of the human brain

International neuroscientists have produced a fully 3D map of a human brain – scanning and digitising thousands of ultrathin slices to determine its structure at extremely high resolution.

The map is being made freely available to researchers worldwide. It has a spatial resolution of just 20 micrometres (µm), far exceeding the typical 1 mm (1000 µm) from MRI studies. For comparison, a red blood cell is 8 µm wide.

In recent years, major efforts have been getting underway to probe and map the brain, in the hope of conquering physical and mental illnesses, while better understanding the nature of consciousness. In January, the European Commission awarded €1 billion (US$1.3 bn) to the Human Brain Project, intended to create the world's largest experimental facility for brain mapping. In February, Barack Obama announced the Brain Activity Map Project – a decade-long effort to map the activity of every neuron in the human brain. In March, the Human Connectome Project released a major dataset, revealing the complexities of the brain's structure and giving a clearer picture of its role in neural disorders. Yet another major initiative is the Blue Brain Project, founded in Switzerland in 2005, which aims to create a synthetic brain by reverse-engineering the mammalian brain down to the molecular level.

The BigBrain project, seen in the video below, was reported yesterday in the journal Science. Its fine-grained resolution will allow scientists to gain new insights into the basis of cognition, language and emotions. The researchers also stated that they plan to extract measurements of cortical thickness in order to study aging and neurodegenerative disorders; create cortical thickness maps to compare data from in vivo imaging; integrate gene expression data; and generate an even better model with a resolution of 1 micron to capture details of single cell morphology.

The resolution, bandwidth and image reconstruction times of brain scanning technologies have been improving at an exponential rate since the 1970s. With ongoing advances in computer power, this pace of progress is likely to continue in the future. A single neuron model was developed in 2005; a neocortical column with 10,000 neurons was created in 2008; an entire cortical mesocircuit featuring 100 neocortical columns was simulated in 2011. A rat brain with 100 mesocircuits is expected in 2014. If this trend continues, it is reasonable to assume that a fully working, real-time simulation of an entire human brain is possible by the mid-2020s. This would have profound implications – not only for the treatment and understanding of illnesses, but also for the growth of artificial intelligence.

Comments »