To examine astrocytic function, we have developed model mice harboring astrocyte-specific disease causal gene and tried to examine astrocytic function in vivo. Alexander disease, megalencephalic leukodystrophy with subcortical cysts (MLC), and autistic spectrum disorder with neuroligin 3/4 mutations are known to be astrocyte-specific disease so far. First of all, we have established Alexander disease model mouse. Alexander disease is caused by coding mutation in glial fibrillary acidic protein (GFAP) and mutant GFAP forms inclusion bodies, called Rosenthal fibers, in astrocytes. Its pathophysiology is still unknown. We generated transgenic mice that express human GFAP R239H mutant under the control of mouse GFAP promoter. Lines with single copy exhibited weak human GFAP expression in astrocytes that did not produce aggregates despite the existence of mutation, whereas lines with multi copies exhibited strong expression and the formation of aggregates, starting at P14. The line with aggregates showed higher sensitivity to kainate than the line without them, whose sensitivity was not different from the wild type mouse, suggesting that the presence of GFAP aggregates but not the presence of mutant GFAP altered the sensitivity. Changes in several electrophysiological parameters, including facilitation of LTP, were also observed in this model mouse. We believe that this transgenic line is a useful tool to study astrocytic function in vivo.