Interpretation of the neural findings of deception without considering the ecological validity of the experimental tasks could lead to biased conclusions. In this study we classified the experimental tasks according to their inclusion of three essential components required for ecological validity: intention to lie, social interaction and motivation. First, we carried out a systematic review to categorize fMRI deception tasks and to weigh the degree of ecological validity of each one. Second, we performed a meta-analysis to identify if each type of task involves a different neural substrate and to distinguish the neurocognitive contribution of each component of ecological validity essential to deception. We detected six categories of deception tasks. Intention to lie was the component least frequently included, followed by social interaction. Monetary reward was the most frequent motivator. The results of the meta-analysis, including 59 contrasts, revealed that intention to lie is associated with activation in the left lateral occipital cortex (superior division) whereas the left angular gyrus and right inferior frontal gyrus (IFG) are engaged during lying under instructions. Additionally, the right IFG appears to participate in the social aspect of lying including simulated and real interactions. We found no effect of monetary reward in our analysis. Finally, tasks with high ecological validity recruited fewer brain areas (right insular cortex and bilateral anterior cingulate cortex (ACC)) compared to less ecological tasks, perhaps because they are more natural and realistic, and engage a wide network of brain mechanisms, as opposed to specific tasks that demand more centralized processes.