Correction of z-motion artefacts to allow population imaging of synaptic activity in awake behaving mice

Functional imaging of head-fixed, awake, behaving mice using two-photon imaging of fluorescent activity reporters has become a powerful tool in the studying the function of the brain. Motion artefacts are an inevitable problem during such experiments and are routinely corrected for in x and y dimensions. However, axial (z) shifts of several microns can also occur, leading to intensity fluctuations in structures such as synapses that are small compared to the axial point-spread function of the microscope. Here we present a simple strategy to correct z-motion artefacts arising over the course of a time-series experiment in a single optical plane. Displacement in z was calculated using dye-filled blood vessels as an anatomical marker, providing high contrast images and accuracy to within ∼0.1 µm. The axial profiles of ROIs corresponding to synapses were described using a Moffat function and this “ROI-spread function” used to correct activity traces on an ROI-by-ROI basis. We demonstrate the accuracy and utility of the procedures in simulation experiments using fluorescent beads and then apply them to correcting measurements of synaptic activity in populations of vasoactive-intestinal peptide (VIP) interneurons expressing the synaptic reporter SyGCaMP6f. Correction of z-motion artefacts had a substantial impact on the apparent correlation between synaptic activity and running speed, demonstrating the importance of correcting for these artefacts for the interpretation of in vivo imaging experiments in awake mice.Summary of Key PointsMotion artefacts associated with motor behaviour are an inevitable problem of multiphoton imaging in awake behaving animals, particularly imaging synapses.Correction of axial motion usually requires volumetric imaging resulting in slower rates of acquisition.We describe a method that is easy to implement to correct z-motion artefacts that allows population imaging of synaptic activity while scanning a single plane in a standard multiphoton microscope.The method uses a reference volume acquired in two colour channels – an activity reporter and an anatomical marker of blood vessels. The procedure estimates the z-displacement in every frame and applies an intensity correction in which the z point-spread function for each synapse is modelled as a Moffat function.We demonstrate that the method allows synaptic calcium activity signals to be collected from populations of synaptic boutons in mouse primary visual cortex during locomotion.