The item-specific proportion congruent (ISPC) effect in a Stroop task – the observation of reduced interference for color words mostly presented in an incongruent color – has attracted growing interest since the original study by Jacoby, Lindsay, and Hessels [(2003) Psychonomic Bulletin & Review, 10(3), 638–644]. Two mechanisms have been proposed to explain the effect: associative learning of contingencies and item-specific control through word reading modulation. Both interpretations have received empirical support from behavioral data. Therefore, the aim of this study was to investigate the responsible mechanisms of the ISPC effect with the classic two-item sets design using fMRI. Results showed that the ISPC effect is associated with increased activity in the anterior cingulate (ACC), dorsolateral prefrontal (DLPFC), and inferior and superior parietal cortex. Importantly, behavioral and fMRI analyses specifically addressing the respective contribution of associative learning and item-specific control mechanisms brought support for the contingency learning account of the ISPC effect. Results are discussed in reference to task and procedure characteristics that may influence the extent to which item-specific control and/or contingency learning contribute to the ISPC effect.