Commands from the behavioral control computer synchronized the robotic movements of the objective stage with voluntary head restraint to move the objective into imaging position with each insertion. Using this approach, the objective could be held at a safe position (typically 1 mm) above the imaging position, selleck chemicals llc then lowered to the imaging position during restraint, and finally retracted to the safe position prior to release of the kinematic clamp. A necessary

criterion for successful in vivo imaging is that brain motion artifacts are small enough so that they are addressable through software (Dombeck et al., 2007). To quantify the performance of this aspect of the combined microscope and head-restraint apparatus, we measured across-trial

registration and within-trial brain motion during voluntary head restraint in eight Selleck Verteporfin trained rats by imaging GCaMP-labeled neurons in AGm (six rats) or V1 (four rats) through an implanted optical window (Figure 4C, see Experimental Procedures). Prior to implantation of the window, the dura was removed and AAV-GCaMP3 (AGm and V1) or AAV-GCaMP6s (V1) was injected into layer II/III of the exposed cortical region. One to four weeks after implantation of the optical window, GCaMP fluorescence was observed in the perinuclear somata and processes of neurons (Figure 4D). For analysis of brain motion, images of GCaMP-labeled neurons were acquired at a rate of 10 Hz over a 6–8 s head restraint period (Figure 5). Motion correlated with the activation of the kinematic clamp limited visibility during the first few hundred milliseconds of the behavioral trial, delaying the start of the effective imaging period until approximately 600 ms after the initiation of head restraint (Figure 5A).

In addition, when an immersion fluid objective was employed, optical distortions caused by the removal of immersion fluid prevented image acquisition in the last 500 ms of the head-restraint trial. In vivo trial-to-trial displacement (4.7 μm in x, 8.4 μm in y, 3.5 μm in z; Figures 5B, 5C, and S3) was slightly larger than that measured by manual insertion of an isolated headplate. In most cases these registration errors could be corrected by offline image registration Liothyronine Sodium algorithms (see Experimental Procedures). However, on a subset of trials in which the immersion objective was used (10.0% ± 11.5%, n = 13/130 trials), no visible image was produced. This problem was caused by loss of imaging fluid, by the formation of bubbles in the imaging fluid, or by movements of the rat’s head after it had triggered a behavioral trial but before the kinematic clamp was fully engaged. Nevertheless, because of the large number of trials performed per day, the loss of 10% of imaging did not significantly impact the utility of the immersion objective.