Eye-tracking measures can be useful in the study of cognition by serving as indicators of processes such as physiological arousal, attention and mental effort. These measures have been used extensively to study cognition in adults but to a much lesser degree in children. The goal of the present study is to use measures of gaze patterns to gain insights into cognitive processes used by neurotypical middle school children performing on-line math lessons. To achieve this goal, we are developing a novel system that combines on-line learning tools with web camera-base eye tracking. Here we present promising preliminary that demonstrates the potential effectiveness of this system. We intend to finish developing this system and the use it to relate gaze patterns to lesson performance and probe questions to identify the specific conditions under which students do and do not learn effectively from these lessons and use that information to help design more effective lessons.