Argument picks of method AutoReject.get_reject_log(self, epochs, picks=None) is not used.

[mscheltienne]

As spotted in PR #225, line 1042 to 1087 in autoreject.py:

def get_reject_log(self, epochs, picks=None):
      """Get rejection logs of epochs.

      .. note::
         If multiple channel types are present, reject_log['bad_epochs_idx']
         reflects the union of bad trials across channel types.

      Parameters
      ----------
      epochs : instance of mne.Epochs
          The epoched data for which the reject log is computed.
      picks : str | list | slice | None
          Channels to include. Slices and lists of integers will be
          interpreted as channel indices. In lists, channel *type* strings
          (e.g., ``['meg', 'eeg']``) will pick channels of those types,
          channel *name* strings (e.g., ``['MEG0111', 'MEG2623']`` will pick
          the given channels. Can also be the string values ``'all'`` to pick
          all channels, or ``'data'`` to pick data channels. None (default)
          will use the .picks attribute. Note that channels in
          ``info['bads']`` *will be included* if their names or indices are
          explicitly provided.

      Returns
      -------
      reject_log : instance of autoreject.RejectLog
          The reject log.
      """
      # XXX gut feeling that there is a bad condition that we miss
      ch_names = [cc for cc in epochs.ch_names]
      labels = np.ones((len(epochs), len(ch_names)))
      labels.fill(np.nan)
      reject_log = RejectLog(
          labels=labels,
          bad_epochs=np.zeros(len(epochs), dtype=np.bool),
          ch_names=ch_names)

      picks_by_type = _get_picks_by_type(info=epochs.info, picks=self.picks_)
      for ch_type, this_picks in picks_by_type:
          this_reject_log = self.local_reject_[ch_type].get_reject_log(
              epochs, threshes=self.threshes_, picks=this_picks)
          reject_log.labels[:, this_picks] = \
              this_reject_log.labels[:, this_picks]
          reject_log.bad_epochs = np.logical_or(
              reject_log.bad_epochs, this_reject_log.bad_epochs)
          reject_log.ch_names = this_reject_log.ch_names
      return reject_log

picks is not used; instead self.picks_ is used.

[jasmainak]

What do you think should be the intended behavior?

Say you did

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 ?

[mscheltienne]

@jasmainak I have no idea. I never used it and I don't know what it's supposed to do. I just noticed that the argument of the function is not used at all; which is definitely not normal. I will probably give it a try in the coming weeks, but for now, as I never used it, I can't help you figure out what was the intended behavior.

I did notice that in the class _AutoReject (not sure what it is, I would have expected AutoReject to inherit from _AutoReject) the method _AutoReject.get_reject_log does properly use its picks argument.

[jasmainak]

_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 the picks but there might be something to fix

[dengemann]

I have to wrap my head around this ... I think we wrote this ~4 years ago. The general purpose of picking is that you can apply autoreject sequentially on MEG/EEG channels present in one recording but also exclude channels. The .picks_ attribute is good because it does some smart logic at fit time. But that picks does not override the attribute sounds like a bug. It sounds like a good test / example is needed here in which overriding the picks computed at fit time is performed.

[mscheltienne]

Any way you put it, the fact you are asking an argument to the user; to then throw it into the trash looks very suspicious ;)