Functional Specialisation and Effective Connectivity During Self-paced Unimanual and Bimanual Tapping of Hand Fingers: An Extended Analysis Using Dynamic Causal Modeling and Bayesian Model Selection for Group Studies

ABSTRACT

Introduction: This multiple-subject fMRI study continue to further investigate brain activation within and effective connectivity between the significantly (p<0.001) activated primary motor area (M1), supplementary motor area (SMA) with the inclusion of BA44 during unimanual (UNIright and UNIleft) and bimanual (BIM) self-paced tapping of hand fingers. Methods: The activation extent (spatial and height) and effective connectivity were analysed using statistical parametric mapping (SPM), dynamic causal modeling (DCM) and the novel method of Bayesian model selection (BMS) for group studies. Results: Group results for UNIright and UNIleft showed contra-lateral and ipsi-lateral involvement of M1 and SMA. The results for BIM showed bilateral activation in M1, SMA and BA44. A larger activation area but with lower percentage of signal change (PSC) are observed in the left M1 due to the control on UNIright as compared to the right M1 due to the control on UNIleft. This is discussed as due to the influence of the tapping rate effects that is greater than what would be produced by the average effects of the dominant and sub-dominant hand. However, the higher PSC observed in the right M1 is due to a higher control demand used by the brain in coordinating the tapping of the sub-dominant hand fingers. Connectivity analysis indicated M1 as the intrinsic input for UNIright and UNIleft while for BIM, the inputs were both M1s. During unilateral finger tapping, the contra-lateral M1 acts as the input center which in turn triggers the propagation of signal unidirectionally to other regions of interest. The results obtained for BIM (BIMleft and BIMright) however yield a model with less number of significant connection. M1-M1 connection is unidirectional for UNIleft and UNIright originating from contra-lateral M1, and is inhibited during BIM. Conclusion: By taking into consideration the presence of outliers that could have arisen in any subject under study, BMS for group study has successfully chosen a model that has the best balance between accuracy (fit) and complexity.