Due to tissue scattering, the current two-photon in vivo brain imaging techniques have
limited imaging depth. By inserting miniature invasive probes such as prism and GRIN lens into
the brain tissue, researchers can increase the imaging depth from sub-millimeter-scale to several
millimeter-scale deep brain regions. The major disadvantage of these techniques is the small field
of view limited by the size of miniature invasive probes. In this thesis, we develop the Clear
Optically Matched Panoramic Access Channel Technique (COMPACT) to increase the field of
view of in vivo deep brain imaging. Instead of directly inserting and fixing miniature invasive
probes inside the brain, we insert a quartz capillary to serve as a channel inside the brain. We can
freely spin and move the imaging probe inside this channel and form a large volume image around
this channel. This technique has been applied in this thesis to millimeter-scale structural, functional,
and behavior studies in the deep brain region of mice.
Another effect caused by tissue scattering is optical aberration in deep brain imaging. Typical
techniques, including COMPACT, utilize spatial light modulators (SLM) or deformable mirrors
(DM) to compensate for the aberration. However, these instruments are expensive and susceptible
to ultraviolet wavelength range and high light intensity. To circumvent these problems, we develop
a technique to modulate optical waves with high accuracy and at low cost. The key idea is to
fabricate a phase profile inside a quartz plate by printing a 3D refractive index profile through
multi-photon ionization. This technique has been applied in this thesis to optical waves control in
ultraviolet wavelength range and high light intensity, creating microlens and in situ optical
aberration correction with high accuracy.
Another method to increase the imaging depth of in vivo brain imaging is three-photon
microscopy. Typical three-photon microscopy has a millimeter-scale imaging depth but low
throughput due to the low repetition rate of laser sources. In this thesis, we develop a technique to
increase the throughput of three-photon microscopy. Instead of scanning the whole field of view
in raster order, we excite the fluorophore of each neuron with one laser pulse by random-access
scanning. By efficiently using the laser power, we could increase the throughput with low tissue
damage.