Former MSc student Jelle van Dijk published a study in NeuroImage in which we quantify the test-retest reliability of population receptive field measurements by comparing mapping experiments separated by several weeks. We found that visual field positions are extremely reliable but second-order parameters, such as pRF size and cortical magnification factor, is far less reliable (although still well correlated across sessions). Control experiments further suggested that these findings do not critically depend on the carrier stimuli used for mapping – but that reliability is a lot better if comparing mapping experiments conducted within the same scanning session. This suggests that changes of the scanning environment over time (both in terms of the setup or the scanner itself) are the greatest source of noise limiting reliability.
van Dijk, JA, de Haas, B, Moutsiana, C, & Schwarzkopf, DS (2016). Intersession reliability of population receptive field estimates. NeuroImage 143: 293–303.