The neural substrates of graphomotor sequence learning.
Contreras-Vidal, Jose L
Braun, Allen R
MetadataShow full item record
Performing sequences of movements easily and automatically is an integral part of our everyday lives. This dissertation examines how a set of individual movements are assembled into a movement sequence, focusing on the neural regions involved, and the timing of their participation. A second, related question is whether the order of encoding of the individual movements can be detected with kinematic and neuroimaging methods. Understanding how sequences are learned is important for expanding our knowledge of how the brain performs neural computations within healthy persons, and of the alterations of these processes in persons with neurological disorders. To examine these questions, we combined behavioral, kinematic and neuroimaging methods to examine motor sequence learning in healthy adults. The behavioral task involved subjects learning to copy a novel sequence of line-pairs (a graphomotor trajectory sequence learning paradigm) while blood oxygen level dependent (BOLD) functional magnetic resonance imaging (fMRI) data were simultaneously acquired. Sequence learning, measured by normalized jerk, was best characterized by a curve with a double exponential fit, implying that an early, fast learning process (time-bins 1-2) merged with a slower learning process (bins 3-5). The early portion of the sequence learning process was characterized by dorsal and ventral visual stream activation; dorsal lateral premotor cortical activity; and deactivation of the anterior putamen, head of the caudate nucleus and posterior vermis. The deactivation of portions of the basal ganglia and cerebellum during the early phase of sequence learning may indicate that these regions were being reset, due to the high number of errors being produced. The pattern of neural activity in the the second, slower phase of sequence learning suggests that emphasis in the sequence learning process had shifted from visuomotor mapping to improving the kinematic and dynamic motor plans for the new sequences (sequence encoding). Taken together, a model of graphomotor sequence learning emerges, including patterns of neural activation and functional connectivity that correspond to changes in subject performance. This model adds to our current understanding of the neural substrates of graphomotor sequence learning, and may be important in explaining the alterations to these networks in persons with neurodegenerative disorders.