Articles‎ > ‎

Songs For MRIs

posted Apr 9, 2013, 10:33 AM by Ellen Pearlman


Song of songs “Better Man Than He” as MRI

Sivu is the name of a young British singer originally born in St. Ives outside of Cambridgeshire, but now residing in London. Along with director Adam Powell and Black Dog Films he made his first video “Better Man Than He” trilling while lying down inside a MRI chamber for over three hours.  He had the cooperation of Barts Hospital and the London Charity, as well as the charityCLEFT, and of Drs. Marc E. Miguel and Andrew David Scott. The idea came from director Powell’s contemplation of children with cleft palates and lips, and the kinds of techniques that were being developed to help them, including viewing their disabilities inside of MRIs. Powell was simultaneously wrestling with the problem of how to capture images without traditional cameras and lens. Discovering these new MRIs techniques showed him non-traditional methods of image capture were possible. This also fits in with James Elkin’s notion of “visual culture” that includes medical imagery as entering our daily lives in an acceptable manner. 


Sivu as double take inside the machine

The imagery employs a new technique called real time MRI” or “flash MRI” (fast low angle shot ) that uses algorithms that can record up to 50 frames per second.


From the brain to the stars

Previously images could only record one image per second. The flash MRI was first used to analyze speech in this eerie but compelling video.


First live time MRI of speech

The technique was first developed at the Max Plank Institute in Gottingen, Germany. It basically takes the noise out of the image as show in the photo below of a human heart taken at 30 frames per second. A minute of images takes a half hour to process, though that is expected to improve in the future. It cleans up the resolution using different filters and variables.


 Direct reconstruction of image, and new cleaned up  ”iterative” reconstruction

Enormous processing power is needed to render these images. To produce a minute in real time looks at between 2000 and 3000 images that take up two gigabytes of processing power. The computer had four graphical processing units (Tesla, C1060, Nvidia). At this point one needs the cooperation or large research or medical facilities. But the future remains to be seen.