Quantitative Analysis of Three-photon Imaging with Green and Red-shifted Calcium Indicators in the Mouse Brain

Abstract

Three-photon microscopy, with combination of genetically encoded calcium indicators (GECIs) has enabled large-scale neural activity recordings deep within the mouse brain. Application of three-photon microscopy for in vivo imaging requires a careful balance between recording fidelity and heating perturbation induced by infrared laser. In this dissertation, we established a quantitative framework for the optimization of 1320-nm three-photon imaging using green calcium indicator, GCaMP6s, and 1650-nm three-photon imaging using red-shifted calcium indicator, jRGECO1a. We calculated and experimentally verified the excitation pulse energy on the surface to achieve the minimum photon count required for the detection of calcium transients for both 1320- and 1650-nm three-photon imaging, with a side-by-side comparison with 920- and 1064-nm two-photon imaging, respectively. We quantified the cross-over depth beyond which three-photon microscopy outperforms two-photon microscopy in recording fidelity by considering the combined effects of in-focus signal attenuation and out-of-focus background generation. We also quantified the maximum allowable average power on the mouse brain surface by continuous three-photon imaging using Monte Carlo simulation and immunohistochemistry. Our results show that three-photon excitation at the wavelength of 1320-nm is more power-efficient in signal generation, has orders of magnitude higher signal-to-background ratio, and preserves the higher calcium imaging sensitivity and discriminability, than 920-nm two-photon excitation of GCaMP6s-expressing neurons deep within the mouse brain. Two-photon excitation has advantage at shallower depth, the required pulse energy on the brain surface to achieve the 0.1 photon per pulse, which is typical multiphoton signal strength, is much smaller than three-photon excitation. To generate the same amount of signal in GCaMP6s-expressing neurons located at the brain surface, three-photon excitation requires ~8 times the pulse energy as two-photon excitation. To maintain the same signal strength, the pulse energy required for two-photon excitation has to increase more rapidly than three-photon excitation, which eventually becomes comparable to that of two-photon excitation as the imaging depth increases to ~700 µm. The same trend was also observed in fluorescein-labeled blood vessels imaging, three-photon excitation requires ~15 times the pulse energy of two-photon excitation when imaging at the brain surface, while at imaging depth of 750 μm, the pulse energy on the brain surface required for three-photon excitation becomes the same as two-photon excitation. The depth where the transition happens is defined as cross-over depth. The maximum allowable average power on the brain surface was investigated for 1320-nm three-photon excitation, which is 100 mW. There is no obvious induced immunoreactivity with 100 mW excitation power at the wavelength of 1320-nm according to the immunohistochemistry results. The induced increasing immunoreactivity was observed with 150 mW excitation power at 0.8-, 1.0- and 1.2-mm imaging depths. The Monte Carlo simulation suggests that the maximum temperature rise is almost linearly with input average power, and it is imaging wavelength, depth, and field-of-view dependent. Similarly, our results indicate that 1650-nm three-photon excitation is more power-efficient in signal generation, has orders of magnitude higher signal-to-background ratio, and preserves the higher calcium imaging sensitivity and discriminability, than 1064-nm two-photon excitation of jRGECO1a-expressing neurons deep within the mouse brain. On the brain surface, three-photon excitation requires ~4 times the pulse energy as two-photon excitation of jRGECO1a-expressing neurons, while the pulse energy required for three-photon excitation becomes comparable to that of two-photon excitation at the brain surface as imaging depth increases (~775 µm). The same trend can also be observed in Texas Red dextran labeled blood vessels imaging, three-photon excitation requires ~10 times more pulse energy of two-photon excitation, to maintain the same signal strength as imaging depth increases, the two-photon pulse energy delivered to the brain surface has to increase more rapidly than that for two-photon excitation. Our results explained that how three-photon excitation achieves better calcium imaging fidelity than two-photon excitation with both green and red-shifted calcium indicators in the deep brain and quantified the imaging depth where three-photon microscopy should be applied. The maximum allowable average power on the brain surface was investigated for 1650-nm three-photon excitation, which is 50 mW. There is no obvious induced immunoreactivity with 50 mW excitation power at the wavelength of 1650-nm, according to the immunohistochemistry results. The induced increasing immunoreactivity was observed with 75 mW excitation power at the wavelength of 1650-nm, at 0.8-, 1.0- and 1.2-mm imaging depths. Our analysis presents a translatable model for the optimization of three-photon calcium imaging based on experimentally tractable parameters. We also demonstrated the possibility and simplicity of long wavelength reflectance confocal microscopy (LWRCM) for label-free imaging in the mouse brain. By using long illumination wavelength at 1700nm, LWRCM can achieve an imaging depth of ~1.3 mm with µm spatial resolution in adult mouse brain (comparable with three-photon microscopy), which is 3-4 times deeper than that of conventional confocal microscopy using visual wavelength. We also show that the LWRCM can be added to any laser-scanning microscopy with simple and low-cost long wavelength laser sources and detectors, such as continuous-wave diode lasers and InGaAs photodiodes. The advantage of label-free low-power requirements provides new opportunities for biomedical research and clinical applications.