MEG_rec MagnetoEncephaloGraphy (MEG) MEG_Vienna

MEG records one or more components of the magnetic field that originates from electric activity inside the brain. In contrast to the electric potential measured by EEG, the magnetic field is a physical quantity that has a unique value at each point in space.  Moreover, unlike the electric field it does not interact with the surrounding tissue and reaches the sensors unperturbed. Typical MEG signals are around 100 fT (femto=10-15 Tesla), which is about a billion times weaker than the magnetic field of the earth. Sophisticated sensors, so-called SQuIDs (Superconducting Quantum Interference Devices), and magnetically shielded rooms are necessary to record such tiny signals. The sensors are sitting inside a dewar in liquid helium at a temperature of 4K (-269oC=-452oF). The requirements of permanent cooling and magnetic shielding using µ-metal render MEG a quite expensive technology.

MEG_patThe MEG signal that is recorded with up to several hundred sensors surrounding the head originates mainly from intra-cellular currents in the cortical gray matter. Depending on the MEG instrument manufacturer either two perpendicular tangential components of the magnetic field are picked up, or, as in the image on the left, the field component (more precisely its gradient) that is perpendicular to the head surface. Magnetic field lines that are exiting through the scalp are colored in red/yellow, whereas blue regions indicate areas where the field lines are entering the head. Applying the right-hand rule to both hemispheres allows to determine the underlying current flow as right-downward in the right and left-downward in the left hemisphere. The image shows the activity in the two auditory cortices about 100ms after the subject was exposed to an auditory stimulus. In EEG this response is known and N100 and typically seen as a single minimum of the electric potential over the vertex as in the insert on the left. MEG resolves the activity from the two hemisphere as two dipolar patterns because the magnetic field penetrates the tissue between the source currents and the sensors without interaction. In contrast, the electric field gets scattered mainly at boundaries where the conductivity of the tissue changes (like the cerebrospinal fluid and the skull) and as a result the signals are smeared out before they can be picked up at the scalp. On the other hand, MEG is most sensitive to current sources that are tangentially to the head, i.e. located in the sulcus walls of the folded cortical surface, and virtually blind to currents in the radial direction in the gyri. The reason for this insensitivity is simply physics: A radial current in a conducting sphere induces secondary currents such that their magnetic fields cancel outside the sphere.

In our research we used the MEG systems at CTF in Vancouver, the General Hospital in Vienna and the Hospital for Sick Children in Toronto in an effort to gain a better understanding of the interplay between brain and behavior. Most of this work involves experiments where MEG is recorded while subjects perform tasks of sensori-motor coordination like syncopation or synchronization with an auditory metronome. In addition, we systematically investigated rhythmic auditory stimuli and rhythmic finger movements separately, where the main focus was on the dependence of the MEG signal on the stimulus presentation and movement rate. To ensure that subjects performed rhythmic finger movement at rates that could be purposefully manipulated, we employed a continuation paradigm where during the first few cycles the movements were paced by a metronome, which was then switched off and the subjects continued moving at the pre-set rate. For rhythmic auditory stimuli, in addition to the well-known decrease in amplitude with increasing rate, we found a form of resonance around a presentation rate of 2Hz.

The image on the right shows the time courses of MEG signals at the sensor locations in topological plots together with the color-coded amplitudes of the radial field component taken at the time of maximal power during a syncopation-synchronization experiment. The top part corresponds to syncopation at a stimulus/movement rate of 1.25Hz whereas the bottom display shows the activity recorded at 2.5Hz. At the slower rate the topological patters is very similar to auditory evoked responses, for the faster rate the pattern resembles a motor field for a right-handed finger movement. This change takes place at the rate where subjects cannot syncopate with the metronome anymore and switch spontaneously to synchronization in order to maintain a one-to-one ratio with the stimulus.