Wavelet analysis of cerebral oxygenation oscillations in the screening of Moyamoya disease

Near-infrared spectroscopy (NIRS) was used to investigate the cerebral oxygenation of Moyamoya and healthy subjects. Continuous recordings of NIRS signals for 20 min (20 min signals) were obtained from 17 healthy subjects (age: 37.4±11.3) and 17 Moyamoya subjects (age: 40.1±11.2). Spectral analysis based on wavelet transformation identified five frequency intervals (I, 0.0095 Hz to 0.02 Hz; II, 0.02 Hz to 0.06 Hz; III, 0.06 Hz to 0.15 Hz; IV, 0.15 Hz to 0.40 Hz; and V, 0.40 Hz to 2.00 Hz) in the 20 min signals and three frequency intervals (III, 0.06 Hz to 0.15 Hz; IV, 0.15 Hz to 0.40 Hz; and V, 0.40 Hz to 2.00 Hz) in the 3 min signals (the first 3 min signals were continuously extracted from the 20 min signals). Significant differences (p < 0.05) were found in frequency intervals III and V. The former exhibited the myogenic activity of smooth muscle inside the blood vessels in both 20 min and 3 min signal analyses; the latter showed hemodynamic oscillation caused by heart pumping. This finding agrees with the vascular pathological changes in Moyamoya disease. As a potential screening method for Moyamoya disease, the simple threshold method exhibited 73.5% accuracy.