Functional neuroanatomy during language processing: correspondence of cortical stimulation mapping, fMRI, PEPSI, and ERP during a visual object naming task

Abstract

This study examines the correspondence of functional language maps using Cortical Stimulation Mapping (CSM), functional MRI (fMRI), Proton Echo-Planar Spectroscopic Imaging (PEPSI), and source localizations from Event-Related Potentials (ERP) during a visual object naming task in two patients undergoing treatment for intractable epilepsy. The spatial relationship between fMRI/PEPSI peak activations and CSM language sites was quantitatively analyzed, indicating minimal correspondence across these techniques. Qualitatively, areas of activation that were common across imaging techniques corresponded well with sites of naming disruption in the anterior superior temporal gyrus and posterior middle temporal gyrus. One control subject participated in fMRI, PEPSI, and ERP sessions and is also examined. For all subjects, fMRI/PEPSI activations and ERP sources for N100 and N400 components were consistent with neuroanatomical areas described for visual object naming in the fMRI and PET literature. Areas most consistently activated across fMRI/PEPSI/ERP techniques were the middle temporal gyrus posteriorly and the superior frontal gyrus and insula anteriorly. ERP scalp distributions for the N100 component across subjects were consistently prominent over anterior sites, while prominent amplitudes for the N400 component varied across subjects in posterior, centro-temporal, and anterior sites. Sources for the N100 component were also consistent across subjects in posterior areas, likely reflecting visual processing, while sources for the N400 component varied across subjects in frontal, temporal, and parietal areas, possibly reflecting not only semantic, but phonological processing as well. Concurrent anterior and posterior fMRI activations appear to be functionally distinguishable based on the ERP source data, again with posterior activations reflecting lower-level visual processing and anterior activations reflecting semantic or phonological processing. Implications for the functional role of specific anatomical structures (e.g. middle temporal gyrus) based on these correspondences are discussed. Theoretical issues pertaining to the physiological mechanisms of CSM disruption, fMRI-BOLD activations, lactate metabolism, and ERP source localizations are also examined to account for matches and mismatches across techniques.