Mining neuronal morphology databases using graph transformer

Neuronal morphology plays a crucial role in determining neuronal cell classification, mapping brain connectivity, and ultimately understanding the mechanisms of brain function. Currently, Neuronal morphology analyzing tasks mostly rely on traditional methods based on prior knowledge, greatly limiting the accuracy of the analysis results. To further enhance the effectiveness of neuronal morphology analysis tasks, we propose a novel graph transformer network for analyzing neuronal morphology. We tested our method on a dataset collected from Neuromorpho, and achieved better results than other methods.