Objective: Alternatives to standard methods used to detect feigned memory impairment would have practical value. Eye movements have been used to index memory in past work, and might be sensitive to simulated memory impairment. Method: In this experiment, 48 participants were instructed to perform normally or to simulate memory impairment on a recognition memory task. Participants studied 24 pictures twice each, for 3s and were tested with 48 three-object displays; half contained a studied object and two novel foils (target-present displays), and half contained three novel objects (target-absent displays). Participants indicated whether each display contained a studied object and eye movements were recorded. Finally, during a post-test, participants viewed the same three-object displays and indicated the presence and locations of studied objects. In this phase, simulators were instructed to cease manipulating performance, to determine whether or not past efforts to withhold memory had any effect on subsequent memory expression. Results: As expected, simulators performed more poorly on the explicit recognition test than controls, but eyetracking data indicated that simulators looked disproportionately at studied objects, even when their explicit recognition responses were incorrect. This eye-movement-based memory effect was evident shortly after test display onset for both groups of participants. At post-test, simulators improved significantly on the recognition test, but continued to be outperformed by controls. Conclusion(s): Eye movements belied simulators' behavioral responses and indicated viewing patterns were sensitive to memory for studied items. These results suggest that eye movements show promise as an alternative for detecting malingering and deserve further study.