Dynamic reassembly of the cytoskeleton and structural changes represented by dendritic spines, cargo transport, and synapse formation are closely related to memory. However, the visualization of the nanoscale topography is challenging because of the diffraction limit of optical microscopy. Scanning ion conductance microscopy (SICM) is an effective tool for visualizing the nanoscale topography changes of the cell surface without labeling. The temporal resolution of SICM is a critical issue of live-cell time-lapse imaging. Here, we developed a new scanning method, automation region of interest (AR)-mode SICM, to select the next imaging region by predicting the location of a cell, thus improving the scanning speed of time-lapse imaging. The newly developed algorithm reduced the scanning time by half. The time-lapse images provided not only novel information about nanoscale structural changes but also quantitative information on the dendritic spine and synaptic bouton volume changes and formation process of the neural network that are closely related to memory. Furthermore, translocation of plasmalemmal precursor vesicles (ppvs), for which fluorescent labeling has not been established, were also visualized along with the rearrangement of the cytoskeleton at the growth cone.