Error position electrodes when doing ICA

💬 Support & Discussions
, ,
1

Hi everyone,

I’m preprocessing my data but I have an issue when I do the ICA analysis. Before I do the ica, I plot the montage using this code:

raw.plot_sensors(kind='topomap', show_names=True, show=False,sphere='auto')

I have created a montage with the code:

mne.channels.make_dig_montage(ch_pos=channel_positions, nasion=nas, lpa=lpa, rpa=rpa)

When I plot this montage I get what you see in the first figure

image
image440×423 21.4 KB

However, after I do that plot, I ran the ICA and I get the IC with the positions of the electrodes incorrect. Why are the electrodes in the IC positioned differently that in the raw.plot_sensors()? And how can I make them be in the same positions as in the sensor plot? Thank you

image
image305×357 62.9 KB

2

Which code do you use for plotting either figure?

3

For plotting the topoplot I used raw.plot_sensors(kind=‘topomap’, show_names=True, show=False,sphere=‘auto’) and for plotting the ICs from the ICA I used ICA.plot_components()

4

You are passing your own channel positions – did you make sure that they are in units of millimeters meters?

5

I just saw that I should also put sphere= ‘auto’ in the ICA.plot_components().

So never mind :slight_smile:

6

Yes, I made sure the units are correct.

7

I would say it still suggests that something is not quite right about those coordinates…

1 Like
9

sphere="auto" automatically fits the sphere to your digitization points; which by-passes any scaling (unit) issue. But the fact that the topographic plot on a sphere of radius 0.095 m (default of MNE) yields the ICA plot you see shows that there is a scaling issue somewhere.

Mathieu

2 Likes