Image-based identification of single neurons for noninvasive imaging purification

A single-cell-based screening assay requires strict identification and isolation of particular target cells from a mixture of various kinds of cells. We have developed a visual-image-based on-chip microfluidic cell sorting method for the collection of neurons. One of the advantages of our method of purifying neurons is the direct monitoring and reorganization of neurons with specific image indexes, such as the cell size, shape, internal complexity, and spatial distribution of a fluorescent dye of a specific antibody marker by phase-contrast/fluorescence microscopy and image processing, which has not been realized using conventional diffraction-based cell sorting systems. First, we compared the differences of microscopic images (shapes) of neurons and glia cells, and found that only neurons have neurites extending from the cell body. We also found that the smooth surface shape indicates neurons, and the rough surface shape indicates glia cells. After picking the neuron cells manually chosen by observing their shapes as described above, we confirmed that the purified neurons can be cultivated and can keep their electrophysiological functions on the chip even after the purification procedure. The results indicate the potential of a nonlabel, noninvasive on-chip cell sorting procedure for neurons using micrograph images for an on-chip ultrahigh-speed camera-based imaging cell sorter.