Poster F109, Tuesday, March 27, 8:00-10:00 am, Exhibit Hall C
High and low-frequency activity in intracranial electroencephalography reflect the difficulty of mental arithmetic operations
Michael J. Randazzo1, Youssef Ezzyat1, Michael J. Kahana1; 1University of Pennsylvania
Mental arithmetic is fundamental to many everyday behaviors and is hypothesized by several models to be accomplished through either fact retrieval or direct computation, determined primarily based on task difficulty. Recent functional magnetic resonance imaging (fMRI) literature has provided evidence for the existence of both a fronto-parietal network mediating computation and a temporal-hippocampal network mediating recall. In this study, we sought to characterize these two pathways by describing the spectral correlates underlying easy and difficult arithmetic using intracranial electroencephalography. Three hundred thirty-five patients with intractable epilepsy undergoing seizure monitoring completed self-paced calculations requiring addition of three, single-digit numerals. To estimate problem difficulty, a model was fit to response times for all subjects, and trials were subsequently defined as easy or difficult compared to predicted times. Spectral power differences between conditions were computed using a Morlet wavelet decomposition for frequencies between 3-180 Hz, and electrodes were anatomically classified. A significant, overall pattern emerged across regions throughout the frontal and parietal cortices, whereby difficult trials comparably exhibited reduced low frequency power (3–17 Hz) and enhanced high frequency power (56–180 Hz). Furthermore, this pattern was drastically attenuated or reversed in medial temporal lobe and hippocampus. Our results suggest that difficult arithmetic elicits local activation along with a global desynchronization in regions associated with working memory and numerical manipulation, while similar activity in areas implicated in memory recall is predominantly observed during easy arithmetic. These findings support prior fMRI hypotheses with whole-brain spectral evidence of competing pathways for performing mental calculation.
Topic Area: THINKING: Problem solving