Reliability of fMRI motor tasks in structures of the corticostriatal circuitry: implications for future studies and circuit function.

The corticostriatal circuits are important information processing networks. There is evidence that these circuits may be dysfunctional in a variety of neuropsychiatric conditions ranging from Parkinson's disease to bipolar disorder. Cross-sectional fMRI studies may clarify normal circuit function, and longitudinal studies may provide information on changes related to age in control subjects, as well as illness progression and treatment response in patient groups. In this paper, we report a comprehensive analysis of the utility of several motor tasks as cross-sectional and longitudinal probes of corticostriatal function in terms of their activation strength and reliability. Our findings suggest that the motor tasks studied can be useful probes of corticostriatal function for studies utilizing group comparisons. However, longitudinal clinical studies in which individual results are important will need to take into account wide variation in individual activation and reliability. For example, measures of activation strength and reliability based on percent signal change display a dichotomy between simple motor tasks, which have high reliability and low activation, and complex tasks, which have lower reliability and higher activation. Size and overlap ratios calculated from activation maps produced a different view of reliability than intraclass correlation coefficients (ICC) based on percent signal change. Finally, these results suggest that the corticostriatal circuitry exhibit individualized responses to motor adaptation.