Numerical cognition is an essential everyday life skill and encompasses a host of number-related cognitive functions. As with most mental abilities, numerical cognition changes over time. However, while the literature has paid much attention to the early development of numerical cognition, studies on elderly remain scarce. The current evidence on aging is not unequivocal. Moreover, current literature on aging in numerical cognition is rather descriptive and not explicitly framed in theories of cognitive aging. The triple code model and the HAROLD hypothesis respectively postulate predictions on the laterality of processes underlying numerical processing and compensational changes in laterality over age. Combining those two theories results in some unanswered questions, as current literature provides no clarity on how numerical task performance and underlying laterality are related to one another over age. Therefore, this project aims to investigate underlying brain laterality using Functional Transcranial Doppler Sonography (fTCDS), as this may prove a promising avenue to address this issue and to come to a better understanding of aging effects in mathematical cognition. In case aging impairs numerical task performance, whether or not accompanied by maladaptive underlying laterality, it would be favorable to counteract this impairment. Therefore, this project aims to investigate how training affects both numerical task performance and underlying laterality in elderly.