| Summary: | 博士 === 國立臺灣大學 === 物理學研究所 === 106 === Since the day of Cajal, neuroscientists have accumulated significant amount of knowledge of single neuron or few-neuron circuits. However, to understand the emergent properties of brain, which composed of three-dimensional (3D) networks from thousands to millions of micron-sized neurons with millisecond to second temporal dynamics, suitable tools should be adopted to explore the functional dynamics throughout whole living brain with single neuron spatiotemporal resolution, i.e., functional connectome.
To study functional connectome, electrophysiology has been successfully applied to single neuron measurements with millisecond resolution in an intact brain, and functional magnetic resonance imaging (fMRI) has been widely used to study whole human brain functional properties. However, electrophysiology is invasive, and the number of simultaneously monitored neurons is limited, while fMRI provides only indirect results of brain activities and nonsufficient spatiotemporal resolution to distinguish single neuron. On the other hand, optical methods provide noninvasive measurements, high spatiotemporal resolution to distinguish single neuron, and whole-brain observation when applied to small animal brains, is the optimal tool. In this dissertation, Drosophila is selected as our research target due to its nearly-complete anatomical connectome.
When using optical method to study the brain, confocal/two-photon microscope (2PM) is widely adopted due to their sectioning capability, which is suitable for tissue inspection. However, their 3D acquisition speeds are limited due to sectioning. In this dissertation, we enhance 3D acquisition speed by integrating an ultrasound lens (UL) with a commercial 2PM, providing hundreds of kHz to one MHz axial scan rate with more than 100 μm axial extent. Combined with a commercial scanner that allows arbitrary curve scan on lateral plane, a novel ribbon scan imaging modality is developed. It is demonstrated to monitor millisecond temporal dynamics of 3D neurons of interest without motion artifacts, which is best suited for densely-packed Drosophila brain.
During Drosophila brain functional studies, it is unexpectedly discovered that 2PM cannot penetrated the whole ~ 200 μm living brain. The reason is the extraordinary strong aberration/scattering from the tracheae structures. To improve imaging depth, a 1300-nm laser combined with three-photon excitation (3PE) is adopted to achieve whole-brain observation with subcellular resolution for the first time. The long excitation wavelength simultaneously reduces scattering, and aberration caused by phase error. In addition, 3PE process renders exceptional optical section capability. To explore the mechanism that limit two-photon imaging depth in Drosophila brains, the brain optical properties at various excitation wavelengths are quantitatively characterized for the first time. Surprisingly, at short wavelength, scattering dominates; while aberration exceeds it at long wavelengths and becomes the main impeding factor of whole-brain observation in a living Drosophila.
Through the validations of the 3D high-speed and deep-tissue optical imaging techniques, together with comprehensive understanding of light interaction in Drosophila brains, it paves the way toward constructing the first whole-Drosophila-brain functional connectome.
|
|---|
