Fig. 1. Concept and performance of Grad-TPM. (a) Grad-TPM uses a pair of gradient foci with opposite intensity distributions to successively scan the specimen and generates two images. Shallow objects appear brighter in image 1 (Im1) than in image 2 (Im2), whereas the opposite applies for deep objects. Hence, the depth of a specified object can be decoded from the intensity ratio of the two images. (b) Simplified diagram of our Grad-TPM. Phase-only spatial light modulator (SLM) was incorporated into excitation light path of Gauss-TPM to sculpt gradient foci. (c) Left panels, fluorescent beads imaged by Grad-TPM (upper) and standard Gauss-TPM (lower); right panels, colored depth 3D images of fluorescent beads obtained from two systems. Scale bar, 5 μm. (d) Absolute axial localization difference (|ze|) between Grad-TPM and Gauss-TPM of fluorescent beads. (e) Gauss-TPM depth as a function of Grad-TPM depth and statistics of the localization error (mean with standard deviation). Detailed analyses of the localization error were shown in Fig. S2 in Ref. [27]. Units of z and ze, μm.

Fig. 2. Grad-TPM images of various biological structures show depth resolution and intensity contrast resembling those acquired with Gauss-TPM imaging. (a) Axons in brain slice of Thy1-GFP transgenic mouse. (b) Higher magnification views of boxed axon in (a), and depth profiles along central axis of axon. (c) Microglia in brain slice of CX3CR1-GFP transgenic mouse. First row, stitched images of two axially adjacent volumes (0–22 μm depth); second row, separate images of the upper volume (0–12 μm depth); third row, intensity images corresponding to images in the second row. (d) Depths of microglia cell bodies denoted by 1–5 in images in the first row of (c). (e) Higher magnification views of boxed microglia process in the second row of (c), and depth profiles along central axis. (f) Higher magnification views of boxed areas in the third row of (c) and corresponding intensity profiles along dashed lines for lateral resolution demonstration. Scale bars, 20 μm. Units of z, μm.

Fig. 3. Grad-TPM shows observably lower photobleaching compared with traditional Gauss-TPM in (a) live cell imaging and is, therefore, highly suitable for longitudinal tracking of biological events, such as (b)–(d) phagocytosis of macrophages. (a) HEK293 cells transfected with pCAG-EGFP construct. Inset on upper right corner of each image is a magnified view of boxed area. Circled area in the inset is used to calculate average intensity to create the line graph. (b) Cultured macrophages are phagocytizing fluorescent beads. Scale bars, 20 μm. Time is shown at corner as h:min. Units of z, μm. (c) Representative trajectories of active beads (upper) and immobile beads (lower). Units of coordinate axes, μm. (d) Statistical motility parameters of immobile beads and active beads. Mean velocities indicate no significant difference (ns, two-tailed unpaired t test) of bead velocity between these two categories. Mean displacement curves suggest constrained motility in immobile beads and directed migration in active beads.