Key to the E-M1's promise of bringing together the Four Thirds and Micro Four Thirds systems is its ability to perform phase-detection autofocus, using dedicated pixels on the imaging sensor itself.
The key difference between contrast-detection autofocus (as generally used in compacts and mirrorless cameras), and phase detection (as traditionally used in DSLRs) is that phase detection is able to assess how out-of-focus the image is, and determine directly how far and in what direction the lens needs to move its focus group to achieve a sharp image. Contrast detection has to scan through at least part of its focus range to find the point of optimal focus.
This difference totally changes to the way lenses need to be designed - those optimised for phase detection need to be able to race to a specified location very quickly, whereas contrast detection lenses need to be able to scan back and forth very quickly. Traditionally, very few lenses designed for phase detection have coped very well with the subtle, scanning motion required for contrast detection. Those designed for Four Thirds SLRs could autofocus on previous Micro Four Thirds cameras, but only slowly and hesitantly.
The E-M1 devotes every 16th pixel to autofocus duties, arranged in offset stripes of left and right-looking pixels. The focusing pixels sit behind clear sections of the color filter array to increase their capability in low light and do not contribute to the final image. The result is a 37-point AF array that is used when a Four Thirds lens is mounted.