The contrast response function (CRF) characterizes fundamental input-output properties of visual neurons. In functional MRI, CRF can be obtained for each individual's retintopically defined early visual areas (Boynton et al, 1999). Here, we measured CRFs of BOLD activity in V1, V2, V3, V3A, and V4v of the human visual cortex to identify the mechanisms of covert attention. Using a rapid event-related design, we measured BOLD responses to a brief (100 ms) sinusoidal grating windowed by a 5-7° annulus. Each fMRI run consisted of four grating contrast conditions in a counter-balanced pseudo-random sequence: 0, 1x, 3x, and 10x of each subject's orientation identification threshold. In separate runs, subjects reported either the orientation of the grating (45±5°) or the identity of a letter ("T" or "L") presented simultaneously in the center of the display. CRFs in the five visual areas responding to the grating annulus were obtained in both the central letter ("unattended") and the peripheral grating ("attended") conditions. The Naka-Rushton equation, R = R_max (Gc + c_spn)² /[c₅₀² + (Gc + c_spn)²], where c_spn and G represent spontaneous activities and contrast gain of attention, was fitted to the pair of CRFs in each visual area. The impact of attention depended critically on stimulus contrast: Attention had a greater effect for low stimulus contrasts; but diminished and becomes absent in high stimulus contrast conditions. In all five cortical areas, attention increased spontaneous activity by a factor of about 4.0 and amplified stimulus contrast by a factor of 2.0 to 3.4. By simultaneously increasing the contrast gain and spontaneous activity of the visual system, attention enhances the ability of the perceptual system to detect weak signals at the expense of sacrificing its dynamic range. Studying attentional modulation of the full CRFs provides a framework to systematically evaluate the impact of attention on the fMRI BOLD response in early visual areas.