A multi-threaded program architecture for an asynchronous and highly responsive GUI for automatic neuronal survival quantification

The goal of this paper is to present a multi-threaded program architecture to monitor and control, through the use of a highly responsive GUI, the results of automatically generated spot detections in neurobiology images. Spot detection is used in this context to investigate through the analysis of hundreds of images the spatiotemporal representation of the decrease of neuronal survival depending on the age of newborn neurons in the olfactive bulb stained with BrdU. Existing solutions to this task are plagued with time-consuming image processing steps, tedious interactions with the program and extensive user inputs and outputs. To increase the productivity of the whole process, we have focused our development on the optimisation of a dedicated GUI, which seperates the interactive process from the image processing background task, so that the user has real-time feedback on settings and results. It results in a dedicated solution that integrates in a single and user friendly application all the processes needed to count and classify spots according to biologists' requirements. Section 1 presents the biological context and computational requirements while Section 2 presents the biological protocol. Section 3 gives a short review of existing solutions and methods. Section 4 describes the suggested solution. Section 5 illustrates the application of quantitative data extracted from image analysis.