OBJECTIVE: The present study used complex approximate and exact arithmetic computation problems to dissociate the brain networks for strategy-based approximate computation and procedure-based exact computation. METHOD: Twenty-eight college students were scanned with MRI while they were solving complex approximate and exact computation problems, including addition, subtraction, multiplication, and division of multidigit integers and fractions. The neuroimaging data were analyzed using whole brain, region of interest, and functional connectivity approaches. RESULTS: Results showed that approximate computation relative to exact computation elicited greater activation typically in the bilateral inferior frontal gyrus (orbital), middle temporal gyrus, angular gyrus, and dorsomedial prefrontal cortex. The brain regions overlapped with the general semantic network that also supports mathematical problem solving. In contrast, exact computation elicited greater activations in the left rolandic operculum and bilateral hippocampus. Functional connectivity analysis based on the psychophysiological interaction approach showed that approximate computation had stronger connectivity from the left intraparietal sulcus to the semantic areas. In contrast, exact computation had stronger connectivity from the left intraparietal sulcus to the phonological areas. CONCLUSION: The results suggest that the semantic network supports complex approximate computation and the phonological network supports complex exact computation. (PsycINFO Database Record (c) 2019 APA, all rights reserved).
OBJECTIVE: The present study used complex approximate and exact arithmetic computation problems to dissociate the brain networks for strategy-based approximate computation and procedure-based exact computation. METHOD: Twenty-eight college students were scanned with MRI while they were solving complex approximate and exact computation problems, including addition, subtraction, multiplication, and division of multidigit integers and fractions. The neuroimaging data were analyzed using whole brain, region of interest, and functional connectivity approaches. RESULTS: Results showed that approximate computation relative to exact computation elicited greater activation typically in the bilateral inferior frontal gyrus (orbital), middle temporal gyrus, angular gyrus, and dorsomedial prefrontal cortex. The brain regions overlapped with the general semantic network that also supports mathematical problem solving. In contrast, exact computation elicited greater activations in the left rolandic operculum and bilateral hippocampus. Functional connectivity analysis based on the psychophysiological interaction approach showed that approximate computation had stronger connectivity from the left intraparietal sulcus to the semantic areas. In contrast, exact computation had stronger connectivity from the left intraparietal sulcus to the phonological areas. CONCLUSION: The results suggest that the semantic network supports complex approximate computation and the phonological network supports complex exact computation. (PsycINFO Database Record (c) 2019 APA, all rights reserved).