Intrinsic connections between fronto-parietal network during mental arithmetic tasks

Abstract

Arithmetic knowledge are composed of several components, including procedural skill which are simple mathematics computation (MC) and complex mathematics computation, (MM), conceptual understanding and factual knowledge (FR). These components have been the basic references when conducting the neuroimaging study of cognitive arithmetic. Previous studies showed evidence of cortical activation of fronto-parietal regions with different intensity while performing various types of mental arithmetic problems. In addition, some studies have suggested that problem-size or difficulty can affect the activation of fronto-parietal regions. However, not much research has been done on effects of problem-size and arithmetic components in terms of their brain connectivity. Therefore this study was carried out to determine both empirical effects on the aspect of Functional Specialization (FS) and Effective Connectivity (EC) of fronto-parietal networks. In this study, functional Magnetic Resonance Imaging (fMRI) scans were performed on twenty-two (n = 22) healthy male participants. Each participant performed a series of arithmetic problems which presented in pseudorandom order. For each task, an arithmetic problem of true and false answers were presented in block stimulus paradigm within 30s followed by 30s rest. Participants made a judgement by pressing the handgrip button to indicate the wrong answer. Statistical Parametric mapping (SPM8) and Dynamic Causal Modelling (DCM10) were used to determine brain Functional Specialization (FS) and Effective Connectivity (EC). For each participant, a number of 56 EC models for PART I (Effects of Problem-Size) and 77 for PART II (Effects of arithmetic components) were constructed for each hemisphere to test the existence of couplings between Superior Parietal Lobule (SPL), insular, Inferior Occipital Gyrus (IOG) and Dorsolateral Prefrontal Cortex (DLPFC). These regions have been proved to be the key role in mental arithmetic and representing the fronto-parietal networks. The EC results show the existence of interactions between the regions of interest, but with different pattern of connectivity between hemispheres. In PART I, the SPL ? DLPFC connection during single-digit MM task was the only one that was significant (Posterior probability, P < 0.90) on the left hemisphere compared to other tasks. Meanwhile, the same connection, but on the right hemisphere was only significant during a MM task in PART II. This is due to both regions play important roles in fronto-parietal network, in which SPL is associated with number processing, while DLPFC is generally associated with working memory and task difficulty. Comparison study on MC and FR tasks where the answer can be directly retrieved from the memory, MM tasks were more complex, thus forcing the subject to retrieve, memorize and compute the answers at the same time. These findings indicated that both left and right hemispheres are involved in arithmetic processing. The results obtained from this study revealed that problem-size and arithmetic components do affect the EC between regions of interest despite their insignificantly difference in cortical activation.