Infusion of basic fibroblast growth factor (bFGF) prevents the loss of cholinergic neurons in the septum/diagonal band of broca following fimbria-fornix transection. However, an in vivo test of whether bFGF will also rescue injured non-cholinergic or cortical neurons has not been carried out. Previous studies have shown that the majority of layer II stellate neurons utilize an excitatory amino acid as their neurotransmitter. In order to determine if bFGF acts on non-cholinergic cortical neurons, a paradigm was developed to examine whether or not bFGF could spare layer II entorhinal stellate cells from axotomy induced death or atrophy. Axotomy of the medial entorhinal cortex fibers projecting to the dentate gyrus of the hippocampal formation via the perforant path lead to retrograde cell loss in entorhinal cortex. Fourteen or thirty days after a unilateral knife cut axotomy of the perforant path, layer II of medical entorhinal cortex showed a 28% decrease in large stellate neurons as well as many weakly stained, hollow cells compared to the non-lesioned side or naive controls. Layer IV neurons, however, which do not project via the perforant path, showed little detectable change in the number of cells ipsilateral to the knife-cut as compared to the contralateral side. Intraventricular infusion of bFGF over a period of 14 days reduced the 28% cell loss to less than 6%. Thus, bFGF is capable of preventing cortical neuronal loss and/or atrophy associated with retrograde degeneration of non-cholinergic neurons following axotomy.