Developing And Assessing Methods For Calcium Imaging Data

Calcium imaging is an exciting cutting-edge technology used to quantify neuronal activity across a set of several cells in a local field. In this thesis, we analyzed and developed methods for both the processing and analysis of \textit{in vitro} calcium imaging from neuronal cells. The calcium imaging measures activity in a number of cells from patients with Pitt-Hopkins syndrome (PTHS) and control subjects. We proposed a cell segmentation procedure to determine the locations of the cells in the image and processing steps to extract traces from the time series of calcium imaging under different conditions, including those with introduction pharmacological drugs to change activity patterns. We then utilized previously-published methods for time-series segmentation, employed with a variety of parameters, to determine times of activity and compare it to manual segmentation. We recommend a semi-automated processing stream for cell segmentation based on thresholding a median-centered, standard deviation-scaled image. We recommend using a motif-based correlation method for temporal peak detection, as models using results from this method most often agree with the statistical inferences drawn from models fit on a fully manual pipeline, but note some caveats and discrepancies compared to manual methods