Closed ktavabi closed 6 years ago
yes CIs is broken when you pass gradiometers and they are combined for plotting but not to estimated the confidence intervals :(
I'll try to look over the next few days unless someone beats me to it...
@dengemann
I think I got the CI estimation for grads headed in the right direction with
...
if ch_type == 'grad':
data = [(_merge_grad_data(evoked_.data[picks, :])).
mean(0) for evoked_ in evokeds[condition]]
data = np.asarray(data)
else:
data = np.asarray([evoked_.data[picks, :].mean(0)
for evoked_ in evokeds[condition]])
ci_array[condition] = _ci_fun(data) * scaling
...
But I am still unable to plot them correctly, almost like CI's are scaled differently, and it's got me stumped.
Not sure if this is an inconvenience to anyone WRT #4526 convo, but I am curious to know if I am headed in the right direction.
Still not clear to me if there is a bug in how confidence intervals for gradiometers in
plot_compare_evokeds
are computed? Using simulation example to illustrate...