Hi all!
I was wondering if there is any way of clasifying my sensors in ROIs according to a Brainstorm atlas.
I have MEG Yokogawa sensors, but the manufacturer has not provide me with the area (or ROI) for each sensor. However, I want to classify them into regions (for example, using the Desikan-Killiany Atlas).
Is there any way of doing that using Brainstorm?
I want to know, for example, "Post Central area: AG115, AG116, AG117 and AG118 sensors"
I hope I have expressed myself well. If any part is not clear, I can try to explain it in another way.
Thank you in advance!!!
Victor
Hi Victor,
Unfortunately, this is not something that can be done easily.
A MEG sensor captures activity from brain regions that can be away from your sensor, and may not see anything from a source located just underneath it, depending on the orientation of the cortical sources (MEG is more sensitive to orientations than EEG).
Additionally, the sensors are not attached to the head: the head moves within the MEG helmet, and the head size and shape vary a lot between people. Therefore it is not possible to know a priori over which part of the participant's head a given sensor would be. For this reason, the head of the participant is localized in the MEG helmet at the beginning of each scan.
To get more familiar with the concepts of MEG recording and source imaging, you can refer to the background literature listed at the end of the tutorial page:
https://neuroimage.usc.edu/brainstorm/Tutorials#Background_readings
Francois
Hi Francois!
Thank you for your reply.
As a part of my study, I want to assess the influence of the dispersion in MEG sensors (i.e. how well the sensors reflect the activity underneath them). To do that I'm going to correlate the activity in the sensors with the activity in the areas below them.
I have considered the option of making somehow the brain (with a plotted atlas) as big as the sensor layout, and the visually assign each sensor to an area of the atlas. This could be a very slapdash option, but is better than nothing. Do you know any way of doing this? I would need to increase the size of the brain (with the atlas) to make it as big as the sensor layout, and then plot the sensors and the brain with the atlas simultaneously.
Do you have any other idea to avoid doing this mess?
Thank you again Francois!
I'm not sure how meaningful this is...
Why don't you simply compute the sources, and then the correlation between the source signal in a specific scout with all the sensors?
Thank you very much! It could be a solution! However, in that way, we will not be measuring the activity below the sensors, and thus the aim of the study will not be accomplished.
Thinking about it, maybe using the Desikan-Killiany atlas would not be the best idea because it has too many ROIS (68). Neverhteless, a way of classifying sensors in (bigger) areas is still required. Being realistic 12 areas would be enough, for example they could be: frontopolar, frontal, central, temporal, parietal and occipital, all of them left and right. The 10-20 system provides (in the names) a way of classifying the sensors in these areas, I just want to find a way of doing something similar with the Yokogawa layout.
I know that what I ask for is pretty complicated, but finding a way of doing it, will help a lot with my investigation.
Thank you very much for your help Francois!!!
Victor.
If you need, you can create group of sensors:
https://neuroimage.usc.edu/brainstorm/Tutorials/ChannelClusters
The 10-20 system provides (in the names) a way of classifying the sensors in these areas
With EEG, it makes more sense to associate one sensor with one brain region than with MEG: the signal of a given electrode is more directly correlated with the cortical activity just underneath it, and the sensor is attached to the head at a relative position with the brain that is the same for all the participants.