Mainak Jas
Mainak Jas
What do you think should be the intended behavior? Say you did ```py ar = Autoreject(picks=picks_train) ar.fit(epochs_train) ar.get_reject_log(epochs_test, picks_test) ``` Use `picks_train` i.e., `ar.picks_` if `picks_test` is `None` ?
`_Autoreject` is for a single `consensus` or `rho` and the `Autoreject` class loops over them and computes the scores. Maybe @dengemann can help us figure out what's going on with...
can you provide a full script which we can run by copy-pasting? and also two seeds -- one where few epochs are dropped and one where too many epochs are...
hi @eioe and @SebastianSpeer I'm sorry I haven't gotten around to this yet. I'll have to sit down with @dengemann or @agramfort and investigate. I doubt it is a bug...
> I found you set the consensus_perc = 0.6, which means only a majority of channel marked as bad, the trial will be drop. Is it too high? It sounds...
I would look at the data.
I'm sorry I haven't found time to get around to this issue. Simple solution for now is fix the seed and inspect what autoreject gives you
what do you get when you do: ```py >>> print(picks) ``` ?
That seems like a bug. A PR to fix it would be welcome
FEM seems like a complex undertaking, particularly because MNE is built around BEM-based forward modeling. It's something I have on the books after submitting the grant as it's proposed in...