Neuronal response properties in the brain are not static over time. They can change during development, after deprivation, and following learning. We study such functional plasticity with two-photon calcium imaging, using orientation selectivity in the mouse visual cortex as a model. One way to alter orientation tuning in the visual cortex is stripe rearing, where animals are exposed to contours of only one orientation for a certain period. Earlier studies have shown that stripe rearing causes a relative overrepresentation of neurons in visual cortex tuned to the experienced orientation. It is not clear, however, whether these changes are merely due to a permissive effect, causing cells tuned to the non-experienced orientations to lose responsiveness, or whether the experienced orientation acts in an instructive fashion, such that some cells actively change their tuning. The main reason for this uncertainty is that with conventional methods it is difficult to assess the proportion of unresponsive cells. This problem can be overcome by two-photon calcium imaging, where all neurons are labeled, thereby allowing for an unbiased determination of the fraction of unresponsive cells. We have raised juvenile mice for three weeks with cylinder lens goggles limiting visual experience to only one orientation. Following this period, orientation preference in the visual cortex was determined with two-photon calcium imaging. Stripe rearing changed the distribution of preferred orientations such that more cells responded to the experienced orientation than to the orthogonal orientation. The fraction of responsive neurons was lowered, but this effect could not fully account for the changes observed in the distribution of preferred orientations. The magnitude of the stripe rearing effect increased with cortical depth: the distributions of preferred orientations changed only modestly in upper layer 2/3, but we noted a pronounced drop in the fraction of responsive cells. In contrast, neurons deeper in layer 2/3 did not change their overall responsiveness, but we found a clear shift towards the experienced orientation. Thus, diverse mechanisms contribute to the changes in preferred orientation following stripe rearing, but the effect is at least partially mediated by an instructive process, by which individual neurons change their orientation preference. We next asked the question whether orientation tuning in the visual cortex also shows plasticity under behaviorally relevant conditions. During active vision, the visual cortex is subject to extensive feedback signaling and top-down modulations. However, in which way and to what extent the visual cortex is involved in visual perceptual learning remains highly controversial. We used repeated two-photon calcium imaging in anaesthetized mice over twelve days with the genetically encoded calcium indicator GCaMP3. Applying this technique, we quantified changes in orientation tuning in individual neurons in the visual cortex before, during and after orientation discrimination learning. While overall orientation preference was not affected, tuning width and response amplitude changed during learning in neurons with specific differential preferred orientations (with regard to the rewarded orientation). Strikingly, these changes were correlated with task performance. Moreover, we found a pronounced gain in the number of orientation-selective neurons in mice that performed well in the orientation discrimination task. These neurons were mostly tuned to the rewarded or the orthogonal orientation, pointing to an enhanced neuronal responsiveness to these orientations. These data support, in line with previously proposed theories, a reweighting of visual information during visual perceptual learning.