In 1949, W. S. Alexander described a boy who died at 16 months of age with a history of megalencephaly, hydrocephalus, and psychomotor retardation (1). A striking feature of the pathology in this case was the accumulation of Rosenthal fibers within astrocytes and an associated degeneration or failure of myelination. During the ensuing 15 yr, several similar patients were reported and given such descriptive diagnoses as fibrinoid degeneration of astrocytes, dysmyelinogenic leukodystrophy, fibrinoid leukodystrophy, and megalencephaly with hyaline panneuropathy. When Friede described the sixth case in 1964, he coined the eponym Alexander disease to refer to this fascinating but mysterious childhood disease (2). Subsequently, prominent accumulations of Rosenthal fibers were found in patients who differed in the age of onset, clinical presentation, and distribution of lesions, so that currently 3 forms of Alexander disease are recognized: the infantile, the juvenile, and the adult (3). However, it has been uncertain whether these are all fundamentally the same disease (4). Recent genetic studies implicating mutations in GFAP, the major intermediate filament protein in astrocytes, as a cause of Alexander disease have transformed our view of this disorder and opened new directions for investigation and clinical practice. How these mutations lead to the pathological and clinical manifestations of the disease, and whether this knowledge will lead to a treatment, are now the central questions before us. In this review we survey the current state of understanding of Alexander disease, including classification and diagnostic criteria, and discuss the genetic studies that point toward a common molecular basis for most cases of the disorder.