A team of Johns Hopkins University neuroscientists has discovered patterns of brain activity that may underlie our remarkable ability to see and understand the three-dimensional structure of objects.
Computers can beat us at math and chess, but humans are the experts at object vision. (That's why some Web sites use object recognition tasks as part of their authentication of human users.) It seems trivial to us to describe a teapot as having a C-shaped handle on one side, an S-shaped spout on the other and a disk-shaped lid on top. But sifting this three-dimensional information from the constantly changing, two-dimensional images coming in through our eyes is one of the most difficult tasks the brain performs. Even sophisticated computer vision systems have never been able to accomplish the same feat using two-dimensional camera images.
The Johns Hopkins research suggests that higher-level visual regions of the brain represent objects as spatial configurations of surface fragments, something like a structural drawing. Individual neurons are tuned to respond to surface fragment substructures. For instance, one neuron from the study responded to the combination of a forward-facing ridge near the front and an upward-facing concavity near the top. Multiple neurons with different tuning sensitivities could combine like a three-dimensional mosaic to encode the entire object surface. An article describing these findings appears in the November issue of Nature Neuroscience, available online here: http://www.nature.com/neuro/journal/vaop/ncurrent/full/nn.2202.html.
"Human beings are keenly aware of object structure, and that may be due to this clear structural representation in the brain," explains Charles E. Connor, associate professor in the Zanvyl Krieger Mind-Brain Institute at The Johns Hopkins University.
In the study, Connor and a postdoc
|Contact: Lisa DeNike|
Johns Hopkins University