With some trepidation, because equations and their ilk usually make me feel distinctly unwell, I ventured (along with several hundred others) to the featured lecture yesterday evening on neural networks, and what computers can offer neuroscience. And with relief, I’m happy to report it was really rather inspiring.
Maybe it was the palpable and infectious enthusiasm of the speaker, MIT’s Sebastian Seung. Accustomed to hearing about how computer scientists are being inspired to create ever faster machines based on how the brain crunches data, the nub of his own talk, as he phrased it in the Kennedy-esque finale – was: “ask not what the brain can do for computers – ask what computers can do for the brain.”
The ‘star of the show’ was a video that one of his students had put together, and when he first saw it, he said, it blew him away. The vid uses loads of data from a high-res scanning technique that gives 2D images of brain tissue, and then pieces them all together into a 3D reconstruction. You could see the pleasure on his face as he pressed play, and we were treated to a geeky rollercoaster ride, following the path of one neuron as it snaked through a piece of brain tissue – all the while accompanied by a dramatic movie soundtrack (which Nature Neuroscience editor Charvy Narain reliably informs me was the music from Blade Runner). The audience responded with enthusiastic applause.
Being able to do this kind of ‘connectomics’ – mapping each and every connection between each neuron – for the entire brain is Seung’s lofty ambition. But who knows, perhaps his call for support from others in the same line of work, not to mention his dramatic movie soundtracks, will have enthused the assembled scientists (as it did your equation-phobic correspondent) to help him achieve it.
Leave a Reply