Data from laboratory studies provide evidence of the improvement of scores on cognitive tests. There were a number of different training elements associated with this success. But these studies tended to concentrate on improvements in test scores in a single domain such as memory or attention or on a single test. Studies we describe in this chapter assess the learning of a variety of different tasks or component skills and incorporate tests of generalisation to other tasks not trained and to other general areas of functioning, such as symptoms. Again there were two main functions of these studies, to test cognitive training paradigms and to test models of the relationship between cognitive improvements and functioning outcomes. The key idea is not to show that it is possible to change performance on a single test, but to try to investigate whether the associations apparent between cognitive data and functioning outcomes continue after remediation; that when cognition is improved then functioning will improve. These studies are time intensive, if only because the effects on functioning are unlikely to appear just a few days after the training has been completed. Also, unlike the laboratory studies, they assume that cognition can be changed, rather than being designed to support a zeitgeist of stable and immutable cognitive deficits. Clinical investigations have run in parallel with little cross-connection, so that several training protocols now shown to have little effect on task performance were incorporated into these more general programmes.