Abstract
Optical microscopy is essential for direct observation of dynamic phenomena in living cells. According to the classic optical theories, the images obtained through light microscopes are blurred for about half the wavelength of light, and therefore small structures below this "diffraction limit" were thought unresolvable by conventional optical microscopy. In reality, accurately obtained optical images contain complete information about the observed objects. Temporal resolution is also important for the observation of dynamic phenomena. A challenge exists here to overcome the trade-off between the time required for measurement and the accuracy of the measurement. The present paper describes a concrete methodology for reconstructing the structure of an observed object, based on the information contained in the image obtained by optical microscopy. It is realized by accurate single photon counting, complete noise elimination, and a novel restoration algorithm based on probability calculation. This method has been implemented in the Super-resolution Confocal Live Imaging Microscopy (SCLIM) we developed. The new system named SCLIM2M achieves unprecedented high spatiotemporal resolution. We have succeeded in capturing sub-diffraction-limit structures with millisecond-level dynamics of organelles and vesicles in living cells, which were never observed by conventional optical microscopy. Actual examples of the high-speed and high-resolution 4D observation of living cells are presented.