|
Figure 1. Comprehensive imaging. (A) The random-access, two-photon laser-scanning system used to analyze neuron morphology and functional activity. (B) Schematic of Xenopus tadpole brain visual circuit showing tectal neurons targets (green) for comprehensive imaging in vivo. Neurons are labeled with a space-filling fluorophore to capture morphology and a calcium indicator to monitor activity. Neuronal morphology is first determined using a stack of 2D planes encompassing the entire neuron. (C) Next, the user traces the 3D dendritic arbor of the neuron, from which points-of-interest (POI) along the entire dendritic arbor and cell body are converted to a 3D tree-structure relationship, and (D) automatically interpolated at 2 μm intervals. A “rapid-scan” executes a routine employing AOD-based random-access imaging to simultaneously sample the interpolated POIs to record the AP and all synaptic activity across the neuron.
|
|
Figure 2. Optical train and the stimulation module. (A) The pre-processing optical train for the AOD TPLSM undergoes 4X beam expansion and is passed through a prism used to spatially compensate for the dispersion created in the acousto-optic deflectors deflecting the beam into the back aperture of a 60X objective. (B) The beam enters the optical train of the AOD-TPLSM, passes through (C) two AODs orientated 90 degrees with respect to each other and provides X-Y scanning. Two lenses collimate the laser at the pupil of the back aperture of the microscope objective. The fluorescence light returns through the objective and (D) enters the post-processing optics by reflecting off a dichroic mirror. The emission path on the AOD-TPLSM is split between two channels using a short-pass dichroic mirror. Channel 1, filtered green emissions, and Channel 2 captures red emissions. Calcium transients were observed using jGCaMP7s on the Channel 1. The fluorescence light is detected by a PMT dedicated to each channel.
|
|
Figure 3. POI distribution and Z-actuator schedule and trajectory. (A) A typical distribution of POIs in a tectal neuron and the scanning schedule of POIs synchronizing the activity of the Z-actuator and the AODs. (B) An illustration of the neuron structure representing the POIs scanned in A, and an example of three generic POIs on the neuron in 3D space and how they relate to a possible Z-stage trajectory planned for scanning and synchronization with the AODs (inset). The “short” line scan of each POI is created in image space using 11 pixels. The center pixel exists at the location of the POI and each pixel is created from 5 summed samples of the PMT output.
|
|
Figure 4. Timing diagram for system synchronization. Synchronization between devices is coordinated using the sample clock signal, TCLK. Initially, the start signal (START) is activated indicating the beginning of the measurement sequence. All PMT channels initiate recording and AOD “chirps,” which are responsible for deflecting the path of the laser for line scans for F3D or FS measurement sequences. The Z-stage command signal, CMD-Z, begins after START, and feedback is recorded providing an absolute comparison between the desired position and actual position measured.
|
|
Figure 5. Visual-evoked calcium responses in a brain neuron using GCaMP6m. (A) A line scan of the neuron's soma was acquired at 1,000 Hz showing the evoked response following a 50 ms light OFF stimulus. The maximum intensity peak was 0.404 s after the OFF pulse. The rise time parameter (at intensity 1-1/e of peak) was measured 0.237 s after the OFF pulse, and the decay time parameter (intensity 1/e of peak) occurred 2.907 s after the peak of the signal rising transient. (B) Frequency response of (A) after applying the Fourier transform. Most information is in the very low frequency range (under 2 Hz), with diminishing power until 7 Hz, and noise thereafter.
|
|
Figure 6. System user interfaces. (A) The measurement user interface (i.e., Msmt. UI) provides user access for hardware selection, and experimental design. (B) The cell trace and analysis toolkit user interface (CTAK UI) serves as an interface for drawing the initial 3D structure of the neuron, correcting minor positional errors following scans, and for post-analysis following either the F3DS or FSS routines.
|
|
Figure 7. Scan class hierarchy and scan types. (A) The focus scan (FS) allows the operator to adjust the position of the neuron in 3D-space, and defines the boundaries during scanning. (B) The Full 3D scan (F3DS) images the volume defined in the FS at an interval along the Z-axis, ZINT, which is defined by the operator. (C) After the F3DS is acquired, the operator traces the neuron on each X-Y image planes forming a (D) skeletonized frame of the neuron in 3D space. (E) The registration points, drawn by the operator, are then linearly interpolated over the entire arbor using a spatial separation distance of dL. (F) The POIs are scheduled for scanning and the volume rapid scan (V-RS) is executed scanning all interpolated POIs at calcium imaging rates. (G) Following the RS, the user has an opportunity to make adjustments to the position of the neuron in 3D space to account for minor position shifts.
|
|
Figure 8. The experiment protocol used to validate the microscope. An initial stimulus, consisting of alternating screens “ON” and “OFF” are displayed during the initial F3DS, after which the user draws the initial morphology in the CTAK UI. The RS executes immediately after the initial drawing of the neuron structure, and stimuli are displayed while recording the 3D, evoked activity in the dendritic arbor AP firing. In this stimulation routine, 9 stimuli are presented in sequence-−4 OFF pulses on a bright background a transition shift from the bright background to the dark background and 4 ON pulses on a dark background. Each stimulus is displayed within an 8 s interval (TP) and has a buffer time, TB, before and after the stimulus.
|
|
Figure 9. The visual stimulation chamber was designed to allow simultaneous imaging and visual stimulation by projecting red and black images to screen (diffuser film) using a laser projector (inset). The head of a tadpole is stabilized inside a small cavity that has an opening allowing the tadpoles eye to view the screen.
|
|
Figure 10. Optical train and piezoactuator performance validation. (A) The 0.176 Kg objective assembly mounted on the Aerotech, piezo actuator, and (B) the Bode plot output following a stability test proved the actuator stable up to ~125 Hz. (C) The Z-stage, piezo actuator's response to a step command over 100 and 10 μm, respectively. The Piezo achieved a maximum velocity of 0.043 m/s with a following error less than ± 5 × 10−6 m, and a settling time of <5.3 × 10−3 s.
|
|
Figure 11. Visually evoked tectal neural activity can be recorded across the complete dendritic arbor and soma using random access TPLSM. (A,B) A standard deviation projection image in the green (A) and red channels (B) of a Xenopus tectal neuron expressing jGCaMP7s and mCyRFP created from a F3DS stack of images is shown. (C) The user-drawn POIs tracing out the structure of the neuron. A total of 609 POIs are drawn located at 2 μm distances from their nearest neighbor. (Ci) ΔF/F0 output from the resulting F3DS of the 609 POIs, collected at 3 Hz over a 73 second interval. Responses can be seen for four 50 ms “OFF” stimuli, a “transition shift” stimulus and four 50 ms “ON” stimuli at spaced at pseudorandom time points. White asterisks indicate evoked somatic events. (D–G) The user-drawn POIs tracing out the structure of the neuron with the colored POIs sampled using a series of planar RS
(Di-Gi) collected between 204 and 246 Hz with a similar experimental protocol as used previously (C-Ci). (H–I) A series of POIs that were recorded in both a 3 Hz F3DS scan and 232 Hz planar RS. (J) The traces for the POIs recorded at 3 Hz over 73.00 s. (K) The traces for the POIs recorded at 232 Hz over 80.08 s.
|
|
Figure 12. Visually evoked synaptic calcium transients across the neuron can be recorded with increased temporal resolution using Segmented Scanning. (A,B) A standard deviation projection image in the green (A) and red channels (B) of a Xenopus tectal neuron expressing jGCaMP7s and mCyRFP created from a F3DS stack of images is shown. (C) The user-drawn POIs tracing out the structure of the neuron with the POIs that are sampled in each sub-scan of the Segmented scan are colored and the subsequent visually evoked calcium activity is shown. A total of 324 POIs located at the soma, branch points, filopodia bases and tips are collected at a rate of 6 Hz. (D,E) The maximum of the average ΔF/F0 evoked responses for both a series of four pseudorandom 50 ms “OFF” (D) and four 50 ms “ON” stimuli (E). (F) A series of POIs recorded corresponding to filopodia tips and bases and (G) the individual calcium traces of those POIs.
|
|
Figure 13. Visually evoked synaptic-localized glutamate transients across the neuron can be detected using Segmented Scanning. (A) A standard deviation projection image in the green channel of a Xenopus tectal neuron expressing shown sfGFP-iGluSnFR-A184S. (B) Locations of a total of 259 POIs sampled across the soma, dendritic branch points, filopodia bases and tips. (C) Stimulus-evoked iGluSnFr ΔF/F0 response for 3 s post-stimulus at each POI, collected at a rate of 6 Hz and averaged over 9 50 ms “OFF” stimuli. (D) Transient events detected using a matched filter algorithm (Sakaki et al., 2018), identifying evoked increases that match Super-folder-GFP-iGluSnFR-A184S dynamics (τR¯ = 5 ms, τF¯ = 150 ms) and a window size of 8. (E) Locations and strengths of the responses located spatially across the arbor. (F) A series of POIs recorded corresponding to filopodia tips and bases and (G) the individual iGluSnFr traces of those POIs.
|
