Abstract
This study reports a novel cell classification method based on the observation of trajectories that cells inside a fluidic chamber follow under an externally applied acoustic field. Proposed method is significant both as a cell classification method and as a method for characterising the motion of various cell lines under different surface acoustic wave patterns. The difference is mainly due to the characteristic differences of cells such as mass, surface adhesiveness and cellular volume. We discuss the mechanisms that affect the interaction between cells and surface waves. Classification performance is tested using using support vector machine (SVM), max-likelihood and multilayer perceptron (MLP) methods and accuracy, sensitivity and specificity values are reported for each. The results indicate that the method can be used as a powerful classifier particularly for cells that are hard to distinguish visually. It is observed that for a given frequency, the motion characteristics of different cell lines differ due to the difference between dominant adhesion mechanism for that particular cell line. This observation can be utilized for the development of a frequency based cell manipulation method that is able to target specific cells using their characteristic frequencies. We discuss the potential of the proposed acoustic stimulation method as a cell manipulation technique particularly for uncoupling the motion of different cell lines.