Replace z-drift metric function with Jinho's newest drift calcuation

[jkim0731]

It would be also good to plot the correlation graph, or at least the peak value of the correlation to QC the algorithm. It should be higher than ~ 0.6.

In the future, we might be able to set a threshold on the peak correlation value to semi-automatically QC them. My guess is that this value may depend on the expression pattern and SNR - something to watch out for later.

[jkim0731]

A test data (where manual z-drift calculation overrode automatic calculation): http://mouse-seeks/qc/ophys/experiment/1306747798#zdrift_visual_inspection

[samiamseid]

connected to the following issues:

180

38

35