Determination of axonal transport velocities via image cross- and autocorrelation

On their way to the synapse and back, neuronal proteins are carried in cargo vesicles along axons and dendrites. Here, we demonstrate that the key parameters of axonal transport, i.e., particle velocities and pausing times can be read out from CCD-camera images automatically. In the present study, this is achieved via cross- and autocorrelation of kymograph columns. The applicability of the method was measured on simulated kymographs and data from axonal transport timeseries of mRFP-labeled synaptophysin. In comparing outcomes of velocity determinations via a performance parameter that is analogous to the signal-to-noise ratio (SNR) definition, we find that outcomes are dependent on sampling, particle numbers and signal to noise of the kymograph. Autocorrelation of individual columns allows exact determination of pausing time populations. In contrast to manual tracking, correlation does not require experience, a priori assumptions or disentangling of individual particle trajectories and can operate at low SNR.