A noninvasive detection system for cerebrovascular disease such as aneurysms, stenosis, and occlusions has been developed by using a specially designed PVDF-film sensor and signal analysis based on FFT and the Wigner distribution. About 90 patients with cerebrovascular diseases and 30 normal persons were measured at the Department of Neurosurgery, Ube Industries Central Hospital. FFT spectral differences between patients and normal persons exist over the frequency range from 0.5 to 1.2 kHz. In particular, the differences in spectral components between 750 Hz and 1.5 kHz could be the possible index for diagnosis of the existence of cerebrovascular disease. The FFT spectral pattern showed some differences between stenosis and aneurysms. In stenosis, it seemed that there existed a flat level from 0.4 to 2.2 kHz, while in aneurysms the power decreases smoothly from the peak around 0.7 kHz. Time variance of the blood flow noise according to the cardiac cycle could be seen in the stenosis from 30% to 50%, but not in the cases from 40% to 90%.