Localization and analysis of neural generators of auditory deviance detection using simultaneous EEG and fMRI
Abstract
The goal of this study was to use EEG and fMRI simultaneously to determine the major areas of the brain involved in inattentive and attentive auditory deviance detection, with the focus on determining the major areas of the brain that are involved in generating the auditory P300. A first hypothesis was that improved source localization could be achieved by implementing an improved experimental paradigm for eliciting the auditory P300 event-related potential component, which is associated broadly with target detection. Specifically, that areas of activation strictly associated with attentive auditory deviance detection could be differentiated from other areas associated with different types of neural processes. A second hypothesis was that spatial constraints (location-weighting) on the inverse solution would provide a higher level of sensitivity and specificity in terms of localizing the neural generators of the P300. Specifically, that incorporating mathematical, biophysical, anatomical, and physiological constraints using a composite-weighting scheme would provide improved localization results. FMRI activation and ERP source reconstruction results from this study suggest that the major generators of the P300 are distributed in areas of the temporo-parietal junction and parietal cortex. These results are in good agreement with previous studies, i.e., studies showing that lesions in these regions are associated with reduced P300 amplitudes and with deficits in attentional and memory process. The present, well-controlled paradigm allowed for identifying and differentiating the major areas involved in attentive auditory deviant detection from other regions also showing fMRI activation, thus confirming the first hypothesis. Results of EEG source reconstructions employing different location-weighting schemes and realistically-simulated ERP data suggest that 90-99% relative weighting using the composite-weighting scheme was significantly more effective, in terms of the sensitivity and specificity of their localization capabilities, than all other location-weighting schemes tested. Therefore, incorporating mathematical, biophysical, anatomical, and physiological constraints using a composite-weighting scheme did prove to be significantly more effective for localizing the neural generators of the P300, thus confirming the second hypothesis. In conclusion, with a more accurate determination of the major areas of the brain involved in its generation, the P300 component has the potential to become a valuable clinical tool.
This paper has been withdrawn.