Cerebral blood flow (CBF) is an important parameter for the study of brain function. The present paper examines to what extent CBF in a resting state reflects a stable latent trait and to what extent it reflects phasic situational effects. In 38 healthy subjects resting CBF was measured with continuous arterial spin labeling. Data analyses were performed within the methodological framework of latent state-trait theory, which allows the decomposition of the measured variables into temporally stable differences, occasion-specific fluctuations, and measurement errors. For most of the regions of interest, about 70% of the observed variance in resting CBF was determined by individual differences on a latent physiological trait whereas about 20% of the variance was due to occasion-specific influences. Thus, an aggregation across measurement occasions is not necessary in order to yield a stable physiological trait.