OSCILLATORY DYNAMICS · Examining Oscillatory Dynamics with Magnetoencephalography and Intracranial Electroencephalography

Examining Oscillatory Dynamics with Magnetoencephalography and Intracranial Electroencephalography

Dr Dalal was an incoming Marie Curie fellow from the United States (US). The primary project goal was to resolve the neural activity of deep brain structures, using noninvasive MEG and invasive intracranial EEG (iEEG) and advanced reconstruction techniques. A number of actions were implemented as part of the project, in particular. - Simultaneous MEG-iEEG data was acquired from epilepsy patients. The relationship between the two modalities was quantified. - All algorithmic development from this project has been made publicly available as part of the open-source NUTMEG software http://nutmeg.berkeley.edu. Realistic head models for MEG/EEG based on boundary element modelling were integrated into NUTMEG by interfacing with the OpenMEEG toolbox. - Unaveraged auditory brainstem responses data were acquired and analysed. - Simultaneous recordings ultimately could not be performed with tinnitus patients due to logistical difficulties. Instead, stronger cooperation was formed with the tinnitus group at the University of Konstanz. Researchers at Konstanz were trained in the use of NUTMEG and FieldTrip, and applied time-frequency analysis to various experiments related to the perception of tinnitus. - Finally, an auditory discrimination experiment was conducted using MEG with healthy subjects and iEEG with epilepsy patients. Strong alpha / beta suppressions were observed in both MEG and iEEG. Some high gamma band activity found with iEEG could also be observed with MEG. These results were published in NeuroImage. Meanwhile relatively high correlations were found between hippocampus electrodes and patches of MEG sensors across all patients. However, these patterns differed significantly from the prediction of multisphere forward models; therefore, it was concluded that successful MEG reconstruction of deep sources such as hippocampus may require more realistic forward models. A workflow was therefore created to segment MRI images using BrainVisa and generate realistic head models using OpenMEEG for use with source localisation in NUTMEG. Initial results indicate better correspondence to hippocampal correlation patterns. These results are expected to be published in 2011. To quantify the sensitivity of iEEG electrodes, electrocortical stimulation data were analysed. Observed electrode gains decayed as expected with distance from the stimulating electrodes, and also reflected a relationship to the subtended angle as predicted by Gloor (1987). Furthermore, near skull boundaries and the Sylvian fissure, the gain profile appeared more complex than the classical 'half-space' dipole model. These results inform source localisation efforts from intracranial EEG, currently in progress. Finally, an auditory discrimination experiment was performed with MEG. Time-frequency beamformer analysis revealed that auditory cortex exhibited power decreases in the alpha and beta bands, along with power increases in the gamma band. This study has been presented at a conference, and the manuscript is currently in preparation. Dr Jean-Philippe Lachaux provided valuable training with respect to use of frequency and phase analysis techniques, including the Hilbert transform and phase locking value. Other U821 personnel and collaborators provided training on technical aspects of depth EEG acquisition and analysis. Dr Olivier Bertrand provided insight into auditory responses of the human brain, including the auditory brainstem response and auditory discrimination. These results inspire the acquisition of further linked MEG-iEEG experiments using tasks that specifically modulate the hippocampus, in conjunction with realistic head models. This would improve the confidence and success of MEG (and possibly EEG) as a tool for examining both normal and pathological hippocampus activity. Additionally, iEEG source localisation can reduce the invasiveness of electrode implantation in epilepsy patients, by reducing the number of electrodes needed to acquire clinically relevant information.

Data: CORDIS, © European Union