Fig 1.
ETL installed in the excitation path and controlled via software GUIs.
(A) An ETL in the excitation pathway shapes the incoming beam for remote focusing. Beam shape (red/green) is controlled by curvature of the ETL. L2 and L3 lenses are placed to conjugate the ETL to the back aperture of the objective lens. (B) ETL voltages are correlated to linear stage movements using an automated alignment routine. (C) A GUI controls ETL shape. Voltage values are translated to Z position, and the relative shift of the imaging plane is previewed in a 3D graphic. Z values corresponding to voltages are set either by a linear constant or polynomial curve. (D) Tilted imaging is previewed by adjusting voltage range, as the ETL voltage is altered in phase with the slow-scanning galvanometer.
Fig 2.
Custom autofocus and drift correction parameters track spines over time.
(A) Sample result of the autofocus algorithm selection tool based on 30 Z stacks with 6 slices each. All algorithms adapted from Pertuz, et al. [29]. Abbreviations expanded in Table 1. Percent relative accuracy (gray bars) indicates standardized mean distance of Z position selected by algorithm vs target. 100% = no distance, 0% = maximum distance, 50% = distance if position is picked at random. Blue lines indicate average time to calculate focus value for each slice. (B) Parameters for focus and drift correction are controlled through a GUI. Users identify an algorithm, Z range and amount of steps, and whether extra images are collected for autofocus. Drift correction can be enabled to use galvanometers (scan shift) or motor repositioning. Users also have the option to enable or disable the ETL. (C) Reference-based drift correction speed is correlated with pixel size. Image resolution indicates pixel count for one dimension of a square image. (D) Live updates inform users of the selected focus position (red box) and spine ROI (white box) used to determine relative focus value.
Fig 3.
Non-motorized, automated, multi-position ROI selection, imaging, and photoactivation is controlled through a user-friendly interface.
(A) GUI showing all motor positions that are translated to galvanometer scanning coordinates within a single field of view (FOV, large square). (B) A master GUI to keep track of and move between all imaging positions. Settings and coordinates can be saved and loaded. Z depth is set for each position to automatically modulate uncaging laser power to amounts of tissue interference in brain slices. Experimental Notes are automatically saved with each imaging cycle and can be altered to reflect experimental parameters. (C) A custom timeline interface allows users to design and preview imaging and (blue, green) and uncaging (red) cycles at each position. (D) Sample reference images used for drift correction. A zoomed-out reference image (right) is used for initial alignment. Threshold intensity values are set so each uncaging ROI (E, red) is shifted appropriately relative to the cell dendrite perimeter (E, blue).
Fig 4.
Plasticity in dendritic spines induced using automated focus, drift correction, and glutamate uncaging.
(Top) CA1 dendrite pre- and post- uncaging. Arrow indicates photoactivation ROI. Scale = 1μm. Middle: Average volume change in spines following glutamate uncaging at t = 0. Uncaging lasts 60s. Bottom: Quantification of transient (1–3 min) and sustained (26–30 min) change in spine volume. **** = p<0.0001, * = p<0.05. n = 24 Stimulated spines, 7 neurons.
Table 1.
Abbreviations for autofocus operators.