Automatic reconstruction contest

Managing anatomically complex model cells with the CellBuilder. Importing morphometric data with NEURON's Import3D tool or Robert Cannon's CVAPP. Where to find detailed morphometric data.
Post Reply
Site Admin
Posts: 5796
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine

Automatic reconstruction contest

Post by ted »

A message from G. Ascoli:

Scientists Challenged to Create Better Tools for Image Analysis

The Allen Institute for Brain Science, the Howard Hughes Medical Institute (HHMI), and the Krasnow Institute for Advanced Study at George Mason University are launching an international scientific challenge to speed development of new computational tools that accurately and automatically reconstruct the "shape" of brain cells from available light microscopy data.

The organizers hope the DIADEM Challenge - short for Digital Reconstruction of Axonal and Dendritic Morphology - will lead to innovative solutions to a frustrating problem that has slowed efforts to create a functional atlas of the brain. Neuroscientists agree that a systematic characterization of neurons with their dendrites and axons is essential, since these tree-like structures are highly correlated with the electric activity of, and precise connections between, neurons and are thus linked to the functions of specific brain circuits. But scientists currently spend weeks - and, in some cases, months - tracing the intricate neuronal processes by hand, using data supplied by imaging studies.

"Manual tracing of neurons has created an intolerable bottleneck and is currently limiting the pace of discovery in neural circuit analysis," said Giorgio A. Ascoli of the Krasnow Institute for Advanced Study. "Automating this process will open the exciting path to the comprehensive characterization of neuronal structure and connectivity."

The DIADEM Challenge is open to individuals and teams from the private sector and academic laboratories. The organizers will award a $75,000 cash prize to the winning individual or team whose algorithm is judged to perform the best in tests using real data. Funding for the prize is provided by HHMI and the Allen Institute.

"Solving this computational bottleneck will be key for larger scale studies of brain wiring and to generate an atlas of connections in the brain," said Allan Jones of the Allen Institute. "Sponsoring the DIADEM Challenge fits in well with the Allen Institute's mission of providing broad enabling tools and data to the scientific community."

Competitors will have a year to implement an algorithm for digital reconstruction of neuronal morphology and to test it against manual reconstruction, which is the current "gold standard." Up to five finalists will compete in a tournament at HHMI's Janelia Farm Research Campus in Ashburn, Virginia, in August 2010.

The National Institutes of Health is providing partial support to a scientific conference that is independent of - but held in conjunction with - the final round of the DIADEM Challenge. Yuan Liu, program director for Computational Neuroscience and Neuroinformatics at the National Institute of Neurological Disorders and Stroke, is co-organizing the scientific conference with Ascoli and Karel Svoboda of Janelia Farm.

The idea for the DIADEM Challenge was originally discussed in 2007 at a scientific workshop at Janelia Farm. Scientists at the meeting noted that progress in understanding neural circuits was being slowed by the tedious task of tracing the structure of individual nerve cells by hand.

Even with the advent of computer technology that enables mapping in three dimensions, the full reconstruction of single neurons may take months. The vast majority of axons (the long neuronal projections that transmit information to neighboring cells) and dendrites (the branches on nerve cells that receive information from neighboring cells) must be traced manually. Researchers trace axons and dendrites that have been labeled with markers, such as green fluorescent protein, and imaged using a variety of microscopy techniques.

Participants in the DIADEM Challenge will have the opportunity to test their algorithms on the latest data supplied by neuroscientists. Thus, they will have a chance to assess their solutions in a real-world environment.

Ascoli, Liu, and Svoboda believe the DIADEM Challenge and associated conference could lead to significant scientific and technical advancements.

"It will certainly result in a critical assessment of the remaining obstacles to a complete solution," said Svoboda. "This will be an exciting opportunity to bring computational and experimental scientists together to see if they can solve this problem."

Full details about the DIADEM Challenge - including detailed rules and information for competitors - can be found at
Post Reply