MagnetoEncephaloGraphy (MEG) is a non-invasive technique for measuring magnetic fields induced by the electrical activity of neurons in the brain. The synchronous activation of a neuronal column – a portion of cortex (grey matter) implying more than 100000 neurons – produces a dipole of current, at the origin of the electric and magnetic signal recording by EEG and MEG respectively.
These magnetic fields are of the order of few femtoteslas (10 -15 Tesla), that is to say 10 billion times weaker than the Earth magnetic field. Very sensitive magnetometers, named SQUIDs, allow the measurements of such weak fields. More specifically, a dense array of sensors (248 magnetometers in our case), is placed around the head participant, inside a magnetically shielded room.
All MEG projects belonging to Brain and Language Institute (BLRI) are conducted in the MEG Center of Marseille (Timone Hospital, France). In that center, a dense array of 248 sensors (248 magnetometers), is placed around the head participant, inside a magnetically shielded room.
As for EEG, the issue is to link the magnetic activity that is recording at the surface, to the underlying neuronal activity. The very high temporal precision of the measurements (ms-order) permits a fine study of the brain dynamics during the task.
An experiment usually consists in displaying to the subject visual stimuli (words or images at the screen) and/or auditive stimuli (sentences, sounds), and in recording at the same time the surrounding magnetic field. Depending of the study, participant must react or not in response to the stimulation (by pressing an answer button for example). The measure signal being extremely weak, the general approach is to repeat several ten times the experimental task on a group of participants.
All MEG projects conducted as part of the Brain and Language Research Institute (BLRI), are carried out in the MEG laboratory of Marseille (France), which is a component of the “Institut de Neurosciences des Systèmes“.
Specific software products and toolboxes are available to process MEG data. At CREx team, we mainly use the common Matlab toolbox Brainstorm and FieldTrip. Around these tools, we have developed our own scripts to facilitate the processing of the large amount of data, which we propose to share with the community. The MEG laboratory has highly contributed to the methodological choices made at the different processing stages.
Here are the links to the data processing description:
As mentioned earlier, the signal which is recording to explore brain activity is extremely weak to detect. To ensure reliable results, all the steps of the experimental study, from the experimental design development and the recording, to the data processing and analysis, have to be done with an extreme rigour. This is one of the priorities of our team.