Abstracts

Rationale: Abnormalities in resting neocortical oscillations can be indicative of associated pathology. For example, focal attenuation of fast frequencies and focal slowing may be indicative of a lesion such as gliosis or a tumor. The ability to identify such abnormalities requires the characterization normal patterns of oscillatory activity. However, comprehensive, quantitative standards for such patterns have not been established. Towards creating such standards, we have quantified the types of oscillations that are normally observed in different healthy cortical areas when an individual is at rest. Methods: Approximately 4 minute periods of intracranial electroencephalogram (iEEG) data were acquired from 18 individuals undergoing evaluation for surgical treatment of epilepsy while the individuals were awake but resting. Data from electrodes exhibiting epileptiform activity or near anatomical abnormalities were discarded from analysis. The cortical area under each electrode was estimated by co-registering postimplant CT scans to a preimplant MRI, and parsing the cortex into a set of 35 areas using the FreeSurfer software package and the Desikan-Killiany atlas. iEEG was whitened to highlight power spectrum peaks and divided into one second epochs with 0.5 second overlap. The power spectrum density (PSD) in each epoch was estimated using the discrete Fourier transform with two Slepian tapers and averaged across epochs. The 5% trimmed mean PSD for each channel was normalized to unit power to control for differences in overall power across electrodes and individuals. No attempt was made to control for possible effects of analgesic or antiepileptic medications. Frequencies corresponding to line noise (60 Hz and harmonics) were ignored from analysis. Results: Across all cortical areas, the most commonly found peaks in PSD were ~7 Hz, on the border of conventional definitions of theta and alpha. Theta peaks were most clearly present in several temporal lobe gyri, and beta peaks were evident in paracentral and inferior frontal gyri. The only region to show reliable gamma peaks was entorhinal cortex. Conclusions: Contrary to what might be inferred from scalp EEG, which is dominated by ~10 Hz activity, the most prominent oscillatory activity seen across much of the brain is ~7 Hz. Slower activity appears in the inferior temporal areas and is accompanied by beta activity in the pre- and post-central gyri. Surprisingly, beta activity also extends into inferior frontal areas, and gamma activity can often be quite prominent over entorhinal cortex. The lack of prominent alpha activity in these data is surely at least partially due to limited coverage of occipital cortex in these patients. Future work will investigate how these oscillatory patterns differ when these cortical regions exhibit epileptiform activity, and how they are affected by antiepileptic and analgesic medications.