Background: A simulated functional magnetic resonance imaging (fMRI)
environment, or "fMRI simulator", is helpful for prototyping behavioural tasks and
acclimatizing individuals prior to actual fMRI of brain activity. When a position
tracking system is integrated, such simulators can potentially be used to train individuals
to perform behavioural tasks while suppressing their head motion. This is an
important endeavor, as fMRI remains easily contaminated by unacceptable levels of
motion artifact and no global strategy exists to eliminate the problem.
New Method: In the present work, simulator-training procedures were developed that
included visual feedback of head motion parameters. In two experiments, simulator
training was applied to patients recovering from stroke, as well as cohorts of healthy
young and elderly adults, who performed hand motor tasks.
Results: Simulator training lead to statistically significant reductions in task-correlated motion for a
subset of young healthy adults who had elevated motion levels at the outset (p<0.05). All stroke patients
also showed large intra-individual reductions. Task-correlated motion was not reduced in healthy elderly
adults, who exhibited motion with much slower trends.
Comparison with existing methods: For young healthy adults with elevated task-correlated motion,
simulator training reduced head motion to levels seen in a motion-screened, high compliant group of individuals
with low head motion. Similarly, substantial reductions in task-correlated motion were observed
in stroke patients after simulator training.
Conclusion: These results support use of simulator training in cases where task-correlated head motion
is expected to be problematic during fMRI experiments.