Multi-modal Dielectric-Holographic Optical Analysis of Single Biological Cells Using Machine Learning

Dielectric spectroscopy and holographic optical image analysis are distinct but important label-free and non-invasive modalities for investigating biological phenomena at the cellular level. The two analysis methods are usually performed independently and provide different, but complementary information on a cell’s state. We describe a microfluidic device that can measure both the dielectric features at radio frequencies and the optical features of individual cells while in flow. We discuss how machine-learning techniques can be used to extract the cell’s dielectric and optical properties from these features and how it can be used to enhance cell phenotype discrimination.