When subjects adapt their reaching movements in the setting of a systematic force or visual perturbation, generalization of adaptation can be assessed psychophysically in two ways: by testing untrained locations in the work space at the end of adaptation (slow postadaptation generalization) or by determining the influence of an error on the next trial during adaptation (fast trial-by-trial generalization). These two measures of generalization have been widely used in psychophysical studies, but the reason that they might differ has not been addressed explicitly. Our goal was to develop a computational framework for determining when a two-state model is justified by the data and to explore the implications of these two types of generalization for neural representations of movements. We first investigated, for single-target learning, how well standard statistical model selection procedures can discriminate two-process models from single-process models when learning and retention coefficients were systematically varied. We then built a two-state model for multitarget learning and showed that if an adaptation process is indeed two-rate, then the postadaptation generalization approach primarily probes the slow process, whereas the trial-by-trial generalization approach is most informative about the fast process. The fast process, due to its strong sensitivity to trial error, contributes predominantly to trial-by-trial generalization, whereas the strong retention of the slow system contributes predominantly to postadaptation generalization. Thus, when adaptation can be shown to be two-rate, the two measures of generalization may probe different brain representations of movement direction.

In neuroimaging studies of human cognitive abilities, brain activation patterns that include regions that are strongly interactive in response to experimental task demands are of particular interest. Among the existing network analyses, partial least squares (PLS; McIntosh, 1999; McIntosh, Bookstein, Haxby, & Grady, 1996) has been highly successful, particu-larly in identifying group differences in regional functional connectivity, including differences as diverse as those associated with states of aware-ness and normal aging. However, we address the need for a within-group model that identifies patterns of regional functional connectivity that ex-hibit sustained activity across graduated changes in task parameters. For example, predictions of sustained connectivity are commonplace in stud-ies of cognition that involve a series of tasks over which task difficulty increases (Baddeley, 2003). We designed ordinal trend analysis (OrT) to identify activation patterns that increase monotonically in their expres-sion as the experimental task parameter increases, while the correlative relationships between brain regions remain constant. Of specific interest are patterns that express positive ordinal trends on a subject-by-subject basis. A unique feature of OrT is that it recovers information about func-tional connectivity based solely on experimental design variables. In par-ticular, there is no requirement by OrT to provide either a quantitative model of the uncertain relationship between functional brain circuitry and subject variables (e.g., task performance and IQ) or partial informa-tion about the regions that are functionally connected. In this letter, we provide a step-by-step recipe of the computations performed in the new OrT analysis, including a description of the inferential statistical meth-ods applied. Second, we describe applications of OrT to an event-related fMRI study of verbal working memory and H2 15 O-PET study of visuo-motor learning. In sum, OrT has potential applications to not only studies of young adults and their cognitive abilities, but also studies of normal aging and neurological and psychiatric disease.