Vectorview channel groups

olafhauk · January 14, 2021, 3:13pm

I’m sure I’ve been able to pick MEG sensor groups from Vectorview data (‘left-temporal’, ‘right-frontal’ etc) before, but I cannot find it any more. Any quick tips?

Thanks,

Olaf

richard · January 14, 2021, 4:01pm

You mean in raw.plot()? I believe raw.plot(group_by='selection') should do what you want.

olafhauk · January 14, 2021, 4:24pm

I was too brief. I would like to pick channels (names or indices) for MEG sensor groups (basically like labels in source space).

richard · January 14, 2021, 4:33pm

I don’t know of any public function to do so. There’s a private function that should do what you want though:

github.com

mne-tools/mne-python/blob/826ddd2f669fb35e574698042378c83e9b1217ca/mne/selection.py#L124


            raise ValueError('No match for selection name "%s" found' % n)
    # make the selection a sorted list with unique elements
    sel = list(set(sel))
    sel.sort()
    if spacing == 'new':  # "new" or "old" by now, "old" is default
        sel = [s.replace('MEG ', 'MEG') for s in sel]
    return sel
def _divide_to_regions(info, add_stim=True):
    """Divide channels to regions by positions."""
    from scipy.stats import zscore
    picks = _pick_data_channels(info, exclude=[])
    chs_in_lobe = len(picks) // 4
    pos = np.array([ch['loc'][:3] for ch in info['chs']])
    x, y, z = pos.T
    frontal = picks[np.argsort(y[picks])[-chs_in_lobe:]]
    picks = np.setdiff1d(picks, frontal)

@agramfort we should have a public API for this…

olafhauk · January 14, 2021, 5:43pm

Thanks. I think mne.selection.read_selection is what I remembered

richard · January 14, 2021, 7:09pm

Oh, right that’s perfect!
btw you can use it directly from the mne namespace, i.e., mne.read_selection().

agramfort · January 14, 2021, 8:11pm

you also have a selection parameter in mne.pick_types

Alex

richard · January 14, 2021, 9:09pm

But that would serve a different purpose, unless I’m mistaken? @olafhauk wanted to select channels based on location, whereas mne.pick_types's selection parameter simply restricts the picked channels to a provided list of channel names – but doesn’t help with finding out where on the head these channels are located. Am I missing something?

agramfort · January 15, 2021, 8:08am

indeed. You are right @richard

my bad

Alex