The present work is an extensive investigation of the contribution of orienting behavior to preference decision making in humans. The central claim of this thesis is that gaze assists cognition in choosing preference, integrating phenomena previously demonstrated in the literature, the mere exposure effect and preferential looking, in a positive feedback loop leading to the conscious decision. In other words, the more we look at something, the more we like it, but also the more we like it, the more we look at it. This leads to the effect we observed when tracking subjects' eye-movements while they were choosing the preferred stimulus in two-alternative forced-choice tasks: the likelihood of gazing towards choice continually increased as the decision moment was approaching. We called this pattern the "gaze cascade effect" to illustrate its reinforcing nature, and we showed that it is an indispensable part of any preference decision, no matter the circumstances, as long as subjects' gaze is natural and unrestricted. We obtained the cascade effect even when the stimuli were no longer present on the screen, and the decision was based entirely on internal reconstruction. Moreover, we influenced observers' preference by manipulating their gaze, an effect that could not be accounted for by mere exposure. These results demonstrate that preference formation starts early in the visual processing stream, sharing fast reciprocal connections with simpler, somatic-based behaviors, such as orienting. As general implications, our work contradicts views of the brain as a collection of sequential modules, starting with sensation, through perceptual integration and cognitive association to emotional valence, decision making, motor preparation and motor response. Instead, it supports more recent views, which assume a multitude of parallel modules heavily interconnected and exercising influences on each other from very early on in information processing.