Human learning is a complex process whereby future behavior is altered via the modulation of neural activity. In many cases, such activity displays markers of collective dynamics, leading to non-trivial fluctuations in patterns of functional connectivity. Such fluctuations can display characteristic structure both across time and across subjects. Yet, a fundamental understanding of how neural activity and connectivity track learning processes that are shared across subjects versus those that distinguish one subject from another has remained elusive. Here, we seek to address this challenge in a longitudinal experiment in which healthy adult human participants learned the values of novel objects over the course of 4 days training sessions. To assess the degree to which patterns of functional activity were subject-general versus subject-specific, we calculated the intersubject correlation of fMRI BOLD time series. To perform a complementary assessment of the degree to which patterns of statistical relations between those time series were subject-general versus subject specific, we introduced a measure of intersubject functional connectivity: the Pearson correlation between the functional connectivity matrices of each subject and the functional connectivity matrices of all other subjects. Intersubject correlations in both activity and connectivity were greater than expected in non-parametric permutation tests in the lateral occipital cortex, lingual gyrus, supramarginal area, and sensorimotor cortex. In addition, intersubject correlations in activity were greater than expected in the medial prefrontal cortex while intersubject correlations in connectivity were greater than expected in the superior parietal cortex, posterior cingulate gyrus, frontal pole, superior and middle frontal gyri. Interestingly, over the whole brain intersubject correlations in both activity and connectivity peaked in the early stages of learning, while intersubject correlations in connectivity became steady in the later stages of learning. Finally, individual differences in performance accuracy tracked intersubject correlations in connectivity but not activity. Taken together, our results point to both a conserved and variable brain network substrate for value learning, and begin to distinguish the time scales over which these substrates vary with task performance.