We have over 50 user accounts representing numerous NIMH, NINDS and NIDCD protocols. There is continual interest from additional groups that are working on protocols and planning studies. The MEG Core staff has been working interactively with these users in terms of study design, task programming development, acquisition protocols, and signal processing and data analysis. Procedures have been setup for data security, transfer and storage. A substantial Policy and Procedures Manual has been established. We have also worked with the Scientific and Statistical Computing Core to enable transfer of CTF MEG files to AFNI and developed tools for group statistical analysis. This work has been extended to utilize the instant correlation feature of AFNI to investigate connectivty pattern in MEG data. Technical and scientific results have been excellent. Signal analysis development includes event-related SAM (synthetic aperture magnetometry) and 275 channel ICA (independent component analysis). Development of time-frequency analysis methods has included Stockwell and wavelet transforms as well as multi-taper techniques. Of particular interest is coherence analysis of virtual channels as a method to investigate interacting brain regions. Staff are working with other MEG groups to integrate several signal processing packages including FieldTrip, NUTMEG and BrainStorm. The goal is to have a unified tool package with a user-friendly interface available to the user community. The SAM software is being successfully run on the Biowulf Cluster (utilizing the open source OCTAVE software installed by MEG Core staff)allowing for tremendous increase in computing power. Dr Robinson has introduced new time slice and entropy based source analysis methods to the MEG user community. The ability to localize not only cortial surface sources but deeper structures has been demonstrated. For a working memory task MEG activation patterns for beta band have shown exceptional agreement with fMRI (functional magnetic resonance imaging) results in the same subject group. Beta desynchronization patterns agree highly with the network of bilateral DLPFC (dorsolateral prefrontal cortex) and posterior parietal cortex seen during working memory in fMRI tasks. Altamura et al have shown that there are anticipatory signals seen in the modulation of prefrontal activity that appear to arise from preparation for upcoming task demands. Qian Lou and James Blair in an earlier study examining the neural dynamics of facial threat processing have been able to utilize the fine temporal resolution of MEG to learn that there are brian related responses in the amygdala even earlier than in the visual cortex. This supports the suggestion of a quick processing route in the brain specific to fear expressions. Understanding these brain mechanisms will be important to further study in mood and affective disorders. They have followed with an investigation of the relation of visual awareness and gamma band signals. More recently they have shown that early amygdala response was automatic whereas later activity was subject to attentional load. Newer work by others has found changes in association areas when prosody and emotional expressiond are interwoven. Dr Lou is now Director of the MEG Center in the Dept. of Neurosurgery at St. Louis University. In earlier work Brian Cornwell and colleagues have demonstrated that MEG can reliably discriminate amygdala and hippocampal signals using MEG beamforming techniques. Continuing studies have shown that hippocampal function is impaired in patients with major depression as well as other brain changes when treated with ketamine. These studies are of particular interest to possibly elucidate the mechanism of the anti-depressant action of ketamine infusion. Recent results have shown that increased anterior cingulate activity may be a biomarker that predicts the rapid antidepressant response to ketamine. Salvadore and Cornwell have now further found that functional connectivity during a working memory task can predict the antidepressant response of ketemine. A commentary has suggested that 'psychiatric stress testing'may become a strategy for translational psychphamacology. This work is now continuing in several treatment and pharmacology based studies by Dr Zarate and others. Studying how the brain organizes itself into functional networks is key to understanding normal human cognition as well as when it becomes disordered in mental illness. To this end Bassett and co-workers used the spatial and temporal ability of MEG to study how the brain changes configuration during a motor task compared to when at rest. They found that functional networks were characterized by small-world properties indicating a mix of both local connections and long range connections. They have continued this work to demonstrate that dysfuctional networks can be detected and related to behavioral differences in clinial groups. Rutter et al a have also found differences in resting network patterns in patient groups. The continued interest in 'resting activity'has spurred several addtional MEG studies. Duyn and colleagues (Liu,Fukunga,de Zwart, Duyn, NeuroImage 2010) have shown that the large scale fluctuations seen in fMRI have correlates with that seen in brain electrical activity by MEG. A new NINDS investigator, Dr Biyu He, has begun studies of scale free properties in functional imaging distinquished at rest and during tasks. Reorganization of functional brain networks can also be investigated using these methods as shown in recent work by Dr Weiss and colleagues (see MH002890). Multidisciplinary studies have shown the value of multiple modalities in the finding of a correlation of GABA concentrations from MR Spectroscopy with gamma band power in MEG. Another example is the work by Jabbi et al where a combination of 18FDOPA PET, fMRI, and MEG findings show that midbrain dopamine differentially predict neural response to happy and fearful facial expressions. Studies of resting networks are currently a busy area for a number of investigators;here the temporal dynamics of MEG will allow a fine grained examination. Ghuman and Martin are now extending this work to autism spectrum disorder. Network dynamics and structure and the use of graph theoretic analysis methods have been shown by Bassett and colleagues. Current development work in the Core is aimed at refinement of individual subject analyses with robust statistics and the application to functional connectivity.
Showing the most recent 10 out of 48 publications
