We propose to develop an open-source software system that will manage acquisition, storage and retrieval of very large time-lapse microscopy datasets. High-resolution wide-field images will be constructed from multiple images tiles, and the images will be processed in a pipeline manner, allowing a dataset to be reviewed while image acquisition is ongoing. The image data will be stored in a multi-resolution format, enabling fast review irrespective of the data size. The images can be two-or three-dimensional and have one or more channels. The software will also have an interface to the microscope system in order to be able to control its operation remotely. The ability to review datasets as they are being built and provide feedback to the acquisition system will enable long-term time-lapse studies that do not require the researcher to be at the microscope. Kitware will develop this proposed software system. As part of Kitware's innovative open-source business model, we will form a community of researchers that will use and contribute to the open system. We anticipate this software will find widespread use in microscopy. There is increasing demand for software to automate handling of large image datasets;however, current tools are geared mostly toward processing and displaying images that fit in available RAM. Although the proposed software will be designed for time-lapse microscopy, it will also be useful for slide imaging. If tiled image stacks are acquired on a microscope with a motorized stage, the software will be able to align and stitch the image stacks, creating a montage that can be many hundreds of gigabytes in size. The software will also make it practical to view and analyze such a large dataset. Currently there are no alternative solutions available for this application. We propose to use this software for time-lapse studies of neuroplasticity in well defined neural circuits. We are able to create these circuits by using multiphoton laser irradiation to pattern the substrate creating regions that are inimical to growth. Absorption of multiphoton irradiation does not cause damage to nearby neurons. Thus we are able to pattern the substrate around select neurons, isolating them from the rest of the culture and inducing them to interact to form the well defined circuits. In Phase II we will propose to automate this process of substrate patterning, making it possible to create and monitor multiple circuits in parallel. For the validation study in Phase I, we will create circuits by manually controlling the microscope. Synaptic rearrangements will then be monitored by time-lapse imaging of the circuits over many days.
The project will create an open-source toolkit for acquiring time-lapse microscopy. The software will be used to study neurons growing in culture. The research may eventually be applied to spinal-cord injuries.