python_embed/Lib/site-packages/mne/channels/channels.py

# Authors: The MNE-Python contributors.
# License: BSD-3-Clause
# Copyright the MNE-Python contributors.


from __future__ import annotations  # only needed for Python ≤ 3.9

import os.path as op
import string
import sys
from collections import OrderedDict
from copy import deepcopy
from dataclasses import dataclass
from functools import partial
from pathlib import Path

import numpy as np
from scipy.io import loadmat
from scipy.sparse import csr_array, lil_array
from scipy.spatial import Delaunay
from scipy.stats import zscore

from .._fiff.constants import FIFF
from .._fiff.meas_info import (  # noqa F401
    Info,
    MontageMixin,
    _merge_info,
    _rename_comps,
    _unit2human,  # TODO: pybv relies on this, should be made public
    create_info,
)
from .._fiff.pick import (
    _check_excludes_includes,
    _pick_data_channels,
    _picks_by_type,
    _picks_to_idx,
    _second_rules,
    channel_indices_by_type,
    channel_type,
    pick_channels,
    pick_info,
    pick_types,
)
from .._fiff.proj import _has_eeg_average_ref_proj, setup_proj
from .._fiff.reference import add_reference_channels, set_eeg_reference
from .._fiff.tag import _rename_list
from ..bem import _check_origin
from ..defaults import HEAD_SIZE_DEFAULT, _handle_default
from ..utils import (
    _check_dict_keys,
    _check_fname,
    _check_option,
    _check_preload,
    _get_stim_channel,
    _on_missing,
    _validate_type,
    fill_doc,
    legacy,
    logger,
    verbose,
    warn,
)


def _get_meg_system(info):
    """Educated guess for the helmet type based on channels."""
    have_helmet = True
    for ch in info["chs"]:
        if ch["kind"] == FIFF.FIFFV_MEG_CH:
            # Only take first 16 bits, as higher bits store CTF grad comp order
            coil_type = ch["coil_type"] & 0xFFFF
            nmag = np.sum([c["kind"] == FIFF.FIFFV_MEG_CH for c in info["chs"]])
            if coil_type == FIFF.FIFFV_COIL_NM_122:
                system = "122m"
                break
            elif coil_type // 1000 == 3:  # All Vectorview coils are 30xx
                system = "306m"
                break
            elif (
                coil_type == FIFF.FIFFV_COIL_MAGNES_MAG
                or coil_type == FIFF.FIFFV_COIL_MAGNES_GRAD
            ):
                system = "Magnes_3600wh" if nmag > 150 else "Magnes_2500wh"
                break
            elif coil_type == FIFF.FIFFV_COIL_CTF_GRAD:
                system = "CTF_275"
                break
            elif coil_type == FIFF.FIFFV_COIL_KIT_GRAD:
                system = "KIT"
                # Our helmet does not match very well, so let's just create it
                have_helmet = False
                break
            elif coil_type == FIFF.FIFFV_COIL_BABY_GRAD:
                system = "BabySQUID"
                break
            elif coil_type == FIFF.FIFFV_COIL_ARTEMIS123_GRAD:
                system = "ARTEMIS123"
                have_helmet = False
                break
            elif coil_type == FIFF.FIFFV_COIL_KERNEL_OPM_MAG_GEN1:
                system = "Kernel_Flux"
                have_helmet = True
                break
    else:
        system = "unknown"
        have_helmet = False
    return system, have_helmet


@verbose
def equalize_channels(instances, copy=True, verbose=None):
    """Equalize channel picks and ordering across multiple MNE-Python objects.

    First, all channels that are not common to each object are dropped. Then,
    using the first object in the list as a template, the channels of each
    object are re-ordered to match the template. The end result is that all
    given objects define the same channels, in the same order.

    Parameters
    ----------
    instances : list
        A list of MNE-Python objects to equalize the channels for. Objects can
        be of type Raw, Epochs, Evoked, Spectrum, AverageTFR, Forward, Covariance,
        CrossSpectralDensity or Info.

        .. versionchanged:: 1.11
            Added support for :class:`mne.time_frequency.Spectrum` objects.
    copy : bool
        When dropping and/or re-ordering channels, an object will be copied
        when this parameter is set to ``True``. When set to ``False`` (the
        default) the dropping and re-ordering of channels happens in-place.

        .. versionadded:: 0.20.0
    %(verbose)s

    Returns
    -------
    equalized_instances : list
        A list of MNE-Python objects that have the same channels defined in the
        same order.

    See Also
    --------
    mne.channels.unify_bad_channels
    mne.channels.rename_channels
    mne.channels.combine_channels
    """
    from ..cov import Covariance
    from ..epochs import BaseEpochs
    from ..evoked import Evoked
    from ..forward import Forward
    from ..io import BaseRaw
    from ..time_frequency import BaseTFR, CrossSpectralDensity
    from ..time_frequency.spectrum import BaseSpectrum

    # Instances need to have a `ch_names` attribute and a `pick_channels`
    # method that supports `ordered=True`.
    allowed_types = (
        BaseRaw,
        BaseEpochs,
        Evoked,
        BaseSpectrum,
        BaseTFR,
        Forward,
        Covariance,
        CrossSpectralDensity,
        Info,
    )
    allowed_types_str = (
        "Raw, Epochs, Evoked, Spectrum, TFR, Forward, Covariance, CrossSpectralDensity "
        "or Info"
    )
    for inst in instances:
        _validate_type(
            inst, allowed_types, "Instances to be modified", allowed_types_str
        )

    chan_template = instances[0].ch_names
    logger.info("Identifying common channels ...")
    channels = [set(inst.ch_names) for inst in instances]
    common_channels = set(chan_template).intersection(*channels)
    all_channels = set(chan_template).union(*channels)
    dropped = list(set(all_channels - common_channels))

    # Preserve the order of chan_template
    order = np.argsort([chan_template.index(ch) for ch in common_channels])
    common_channels = np.array(list(common_channels))[order].tolist()

    # Update all instances to match the common_channels list
    reordered = False
    equalized_instances = []
    for inst in instances:
        # Only perform picking when needed
        if inst.ch_names != common_channels:
            if isinstance(inst, Info):
                sel = pick_channels(
                    inst.ch_names, common_channels, exclude=[], ordered=True
                )
                inst = pick_info(inst, sel, copy=copy, verbose=False)
            else:
                if copy:
                    inst = inst.copy()
                # TODO change to .pick() once CSD, Cov, and Fwd have `.pick()` methods
                inst.pick_channels(common_channels, ordered=True)
            if len(inst.ch_names) == len(common_channels):
                reordered = True
        equalized_instances.append(inst)

    if dropped:
        logger.info(f"Dropped the following channels:\n{dropped}")
    elif reordered:
        logger.info("Channels have been re-ordered.")

    return equalized_instances


def unify_bad_channels(insts):
    """Unify bad channels across a list of instances.

    All instances must be of the same type and have matching channel names and channel
    order. The ``.info["bads"]`` of each instance will be set to the union of
    ``.info["bads"]`` across all instances.

    Parameters
    ----------
    insts : list
        List of instances (:class:`~mne.io.Raw`, :class:`~mne.Epochs`,
        :class:`~mne.Evoked`, :class:`~mne.time_frequency.Spectrum`,
        :class:`~mne.time_frequency.EpochsSpectrum`) across which to unify bad channels.

    Returns
    -------
    insts : list
        List of instances with bad channels unified across instances.

    See Also
    --------
    mne.channels.equalize_channels
    mne.channels.rename_channels
    mne.channels.combine_channels

    Notes
    -----
    This function modifies the instances in-place.

    .. versionadded:: 1.6
    """
    from ..epochs import Epochs
    from ..evoked import Evoked
    from ..io import BaseRaw
    from ..time_frequency.spectrum import BaseSpectrum

    # ensure input is list-like
    _validate_type(insts, (list, tuple), "insts")
    # ensure non-empty
    if len(insts) == 0:
        raise ValueError("insts must not be empty")
    # ensure all insts are MNE objects, and all the same type
    inst_type = type(insts[0])
    valid_types = (BaseRaw, Epochs, Evoked, BaseSpectrum)
    for inst in insts:
        _validate_type(inst, valid_types, "each object in insts")
        if type(inst) is not inst_type:
            raise ValueError("All insts must be the same type")

    # ensure all insts have the same channels and channel order
    ch_names = insts[0].ch_names
    for inst in insts[1:]:
        dif = set(inst.ch_names) ^ set(ch_names)
        if len(dif):
            raise ValueError(
                "Channels do not match across the objects in insts. Consider calling "
                "equalize_channels before calling this function."
            )
        elif inst.ch_names != ch_names:
            raise ValueError(
                "Channel names are sorted differently across instances. Please use "
                "mne.channels.equalize_channels."
            )

    # collect bads as dict keys so that insertion order is preserved, then cast to list
    all_bads = dict()
    for inst in insts:
        all_bads.update(dict.fromkeys(inst.info["bads"]))
    all_bads = list(all_bads)

    # update bads on all instances
    for inst in insts:
        inst.info["bads"] = all_bads

    return insts


class ReferenceMixin(MontageMixin):
    """Mixin class for Raw, Evoked, Epochs."""

    @verbose
    def set_eeg_reference(
        self,
        ref_channels="average",
        projection=False,
        ch_type="auto",
        forward=None,
        *,
        joint=False,
        verbose=None,
    ):
        """Specify which reference to use for EEG data.

        Use this function to explicitly specify the desired reference for EEG.
        This can be either an existing electrode or a new virtual channel.
        This function will re-reference the data according to the desired
        reference.

        Parameters
        ----------
        %(ref_channels_set_eeg_reference)s
        %(projection_set_eeg_reference)s
        %(ch_type_set_eeg_reference)s
        %(forward_set_eeg_reference)s
        %(joint_set_eeg_reference)s
        %(verbose)s

        Returns
        -------
        inst : instance of Raw | Epochs | Evoked
            Data with EEG channels re-referenced. If ``ref_channels='average'``
            and ``projection=True`` a projection will be added instead of
            directly re-referencing the data.
        %(set_eeg_reference_see_also_notes)s
        """
        return set_eeg_reference(
            self,
            ref_channels=ref_channels,
            copy=False,
            projection=projection,
            ch_type=ch_type,
            forward=forward,
            joint=joint,
        )[0]


class UpdateChannelsMixin:
    """Mixin class for Raw, Evoked, Epochs, Spectrum, AverageTFR."""

    @verbose
    @legacy(alt="inst.pick(...)")
    def pick_types(
        self,
        meg=False,
        eeg=False,
        stim=False,
        eog=False,
        ecg=False,
        emg=False,
        ref_meg="auto",
        *,
        misc=False,
        resp=False,
        chpi=False,
        exci=False,
        ias=False,
        syst=False,
        seeg=False,
        dipole=False,
        gof=False,
        bio=False,
        ecog=False,
        fnirs=False,
        csd=False,
        dbs=False,
        temperature=False,
        gsr=False,
        eyetrack=False,
        include=(),
        exclude="bads",
        selection=None,
        verbose=None,
    ):
        """Pick some channels by type and names.

        Parameters
        ----------
        %(pick_types_params)s
        %(verbose)s

        Returns
        -------
        inst : instance of Raw, Epochs, or Evoked
            The modified instance.

        See Also
        --------
        pick_channels

        Notes
        -----
        .. versionadded:: 0.9.0
        """
        idx = pick_types(
            self.info,
            meg=meg,
            eeg=eeg,
            stim=stim,
            eog=eog,
            ecg=ecg,
            emg=emg,
            ref_meg=ref_meg,
            misc=misc,
            resp=resp,
            chpi=chpi,
            exci=exci,
            ias=ias,
            syst=syst,
            seeg=seeg,
            dipole=dipole,
            gof=gof,
            bio=bio,
            ecog=ecog,
            fnirs=fnirs,
            csd=csd,
            dbs=dbs,
            temperature=temperature,
            gsr=gsr,
            eyetrack=eyetrack,
            include=include,
            exclude=exclude,
            selection=selection,
        )

        self._pick_drop_channels(idx)

        # remove dropped channel types from reject and flat
        if getattr(self, "reject", None) is not None:
            # use list(self.reject) to avoid RuntimeError for changing dictionary size
            # during iteration
            for ch_type in list(self.reject):
                if ch_type not in self:
                    del self.reject[ch_type]

        if getattr(self, "flat", None) is not None:
            for ch_type in list(self.flat):
                if ch_type not in self:
                    del self.flat[ch_type]

        return self

    @verbose
    @legacy(alt="inst.pick(...)")
    def pick_channels(self, ch_names, ordered=True, *, verbose=None):
        """Pick some channels.

        Parameters
        ----------
        ch_names : list
            The list of channels to select.
        %(ordered)s
        %(verbose)s

            .. versionadded:: 1.1

        Returns
        -------
        inst : instance of Raw, Epochs, or Evoked
            The modified instance.

        See Also
        --------
        drop_channels
        pick_types
        reorder_channels

        Notes
        -----
        If ``ordered`` is ``False``, the channel names given via ``ch_names`` are
        assumed to be a set, that is, their order does not matter. In that case, the
        original order of the channels in the data is preserved. Apart from using
        ``ordered=True``, you may also use ``reorder_channels`` to set channel order,
        if necessary.

        .. versionadded:: 0.9.0
        """
        picks = pick_channels(self.info["ch_names"], ch_names, ordered=ordered)
        return self._pick_drop_channels(picks)

    @verbose
    def pick(self, picks, exclude=(), *, verbose=None):
        """Pick a subset of channels.

        Parameters
        ----------
        %(picks_all)s
        exclude : list | str
            Set of channels to exclude, only used when picking based on
            types (e.g., exclude="bads" when picks="meg").
        %(verbose)s

            .. versionadded:: 0.24.0

        Returns
        -------
        inst : instance of Raw, Epochs, or Evoked
            The modified instance.
        """
        picks = _picks_to_idx(self.info, picks, "all", exclude, allow_empty=False)
        self._pick_drop_channels(picks)

        # remove dropped channel types from reject and flat
        if getattr(self, "reject", None) is not None:
            # use list(self.reject) to avoid RuntimeError for changing dictionary size
            # during iteration
            for ch_type in list(self.reject):
                if ch_type not in self:
                    del self.reject[ch_type]

        if getattr(self, "flat", None) is not None:
            for ch_type in list(self.flat):
                if ch_type not in self:
                    del self.flat[ch_type]

        return self

    def reorder_channels(self, ch_names):
        """Reorder channels.

        Parameters
        ----------
        ch_names : list
            The desired channel order.

        Returns
        -------
        inst : instance of Raw, Epochs, or Evoked
            The modified instance.

        See Also
        --------
        drop_channels
        pick_types
        pick_channels

        Notes
        -----
        Channel names must be unique. Channels that are not in ``ch_names``
        are dropped.

        .. versionadded:: 0.16.0
        """
        _check_excludes_includes(ch_names)
        idx = list()
        for ch_name in ch_names:
            ii = self.ch_names.index(ch_name)
            if ii in idx:
                raise ValueError(f"Channel name repeated: {ch_name}")
            idx.append(ii)
        return self._pick_drop_channels(idx)

    @fill_doc
    def drop_channels(self, ch_names, on_missing="raise"):
        """Drop channel(s).

        Parameters
        ----------
        ch_names : iterable or str
            Iterable (e.g. list) of channel name(s) or channel name to remove.
        %(on_missing_ch_names)s

        Returns
        -------
        inst : instance of Raw, Epochs, or Evoked
            The modified instance.

        See Also
        --------
        reorder_channels
        pick_channels
        pick_types

        Notes
        -----
        .. versionadded:: 0.9.0
        """
        if isinstance(ch_names, str):
            ch_names = [ch_names]

        try:
            all_str = all([isinstance(ch, str) for ch in ch_names])
        except TypeError:
            raise ValueError(
                f"'ch_names' must be iterable, got type {type(ch_names)} ({ch_names})."
            )

        if not all_str:
            raise ValueError(
                "Each element in 'ch_names' must be str, got "
                f"{[type(ch) for ch in ch_names]}."
            )

        missing = [ch for ch in ch_names if ch not in self.ch_names]
        if len(missing) > 0:
            msg = "Channel(s) {0} not found, nothing dropped."
            _on_missing(on_missing, msg.format(", ".join(missing)))

        bad_idx = [self.ch_names.index(ch) for ch in ch_names if ch in self.ch_names]
        idx = np.setdiff1d(np.arange(len(self.ch_names)), bad_idx)
        if len(idx) == 0:
            raise ValueError("All channels would be dropped.")
        return self._pick_drop_channels(idx)

    @verbose
    def _pick_drop_channels(self, idx, *, verbose=None):
        # avoid circular imports
        from ..io import BaseRaw

        msg = "adding, dropping, or reordering channels"
        if isinstance(self, BaseRaw):
            if self._projector is not None:
                _check_preload(self, f"{msg} after calling .apply_proj()")
        else:
            _check_preload(self, msg)

        if getattr(self, "picks", None) is not None:
            self.picks = self.picks[idx]

        if getattr(self, "_read_picks", None) is not None:
            self._read_picks = [r[idx] for r in self._read_picks]

        if hasattr(self, "_cals"):
            self._cals = self._cals[idx]

        pick_info(self.info, idx, copy=False)

        for key in ("_comp", "_projector"):
            mat = getattr(self, key, None)
            if mat is not None:
                setattr(self, key, mat[idx][:, idx])

        if hasattr(self, "_dims"):  # Spectrum and "new-style" TFRs
            axis = self._dims.index("channel")
        else:  # All others (Evoked, Epochs, Raw) have chs axis=-2
            axis = -2
        if hasattr(self, "_data"):  # skip non-preloaded Raw
            self._data = self._data.take(idx, axis=axis)
        else:
            assert isinstance(self, BaseRaw) and not self.preload

        if isinstance(self, BaseRaw):
            self.annotations._prune_ch_names(self.info, on_missing="ignore")
            self._orig_units = {
                k: v for k, v in self._orig_units.items() if k in self.ch_names
            }

        self._pick_projs()
        return self

    def _pick_projs(self):
        """Keep only projectors which apply to at least 1 data channel."""
        drop_idx = []
        for idx, proj in enumerate(self.info["projs"]):
            if not set(self.info["ch_names"]) & set(proj["data"]["col_names"]):
                drop_idx.append(idx)

        for idx in drop_idx:
            logger.info(f"Removing projector {self.info['projs'][idx]}")

        if drop_idx and hasattr(self, "del_proj"):
            self.del_proj(drop_idx)

        return self

    def add_channels(self, add_list, force_update_info=False):
        """Append new channels from other MNE objects to the instance.

        Parameters
        ----------
        add_list : list
            A list of MNE objects to append to the current instance.
            The channels contained in the other instances are appended to the
            channels of the current instance. Therefore, all other instances
            must be of the same type as the current object.
            See notes on how to add data coming from an array.
        force_update_info : bool
            If True, force the info for objects to be appended to match the
            values of the current instance. This should generally only be
            used when adding stim channels for which important metadata won't
            be overwritten.

            .. versionadded:: 0.12

        Returns
        -------
        inst : instance of Raw, Epochs, or Evoked
            The modified instance.

        See Also
        --------
        drop_channels

        Notes
        -----
        If ``self`` is a Raw instance that has been preloaded into a
        :obj:`numpy.memmap` instance, the memmap will be resized.

        This function expects an MNE object to be appended (e.g. :class:`~mne.io.Raw`,
        :class:`~mne.Epochs`, :class:`~mne.Evoked`). If you simply want to add a
        channel based on values of an np.ndarray, you need to create a
        :class:`~mne.io.RawArray`.
        See <https://mne.tools/mne-project-template/auto_examples/plot_mne_objects_from_arrays.html>`_
        """
        # avoid circular imports
        from ..epochs import BaseEpochs
        from ..io import BaseRaw
        from ..time_frequency import EpochsTFR

        _validate_type(add_list, (list, tuple), "Input")

        # Object-specific checks
        for inst in add_list + [self]:
            _check_preload(inst, "adding channels")
        if isinstance(self, BaseRaw):
            con_axis = 0
            comp_class = BaseRaw
        elif isinstance(self, BaseEpochs):
            con_axis = 1
            comp_class = BaseEpochs
        elif isinstance(self, EpochsTFR):
            con_axis = 1
            comp_class = EpochsTFR
        else:
            con_axis = 0
            comp_class = type(self)
        for inst in add_list:
            _validate_type(inst, comp_class, "All input")
        data = [inst._data for inst in [self] + add_list]

        # Make sure that all dimensions other than channel axis are the same
        compare_axes = [i for i in range(data[0].ndim) if i != con_axis]
        shapes = np.array([dat.shape for dat in data])[:, compare_axes]
        for shape in shapes:
            if not ((shapes[0] - shape) == 0).all():
                raise ValueError(
                    "All data dimensions except channels must match, got "
                    f"{shapes[0]} != {shape}"
                )
        del shapes

        # Create final data / info objects
        infos = [self.info] + [inst.info for inst in add_list]
        new_info = _merge_info(infos, force_update_to_first=force_update_info)

        # Now update the attributes
        if (
            isinstance(self._data, np.memmap)
            and con_axis == 0
            and sys.platform != "darwin"
        ):  # resizing not available--no mremap
            # Use a resize and fill in other ones
            out_shape = (sum(d.shape[0] for d in data),) + data[0].shape[1:]
            n_bytes = np.prod(out_shape) * self._data.dtype.itemsize
            self._data.flush()
            self._data.base.resize(n_bytes)
            self._data = np.memmap(
                self._data.filename, mode="r+", dtype=self._data.dtype, shape=out_shape
            )
            assert self._data.shape == out_shape
            assert self._data.nbytes == n_bytes
            offset = len(data[0])
            for d in data[1:]:
                this_len = len(d)
                self._data[offset : offset + this_len] = d
                offset += this_len
        else:
            self._data = np.concatenate(data, axis=con_axis)
        self.info = new_info
        if isinstance(self, BaseRaw):
            self._cals = np.concatenate(
                [getattr(inst, "_cals") for inst in [self] + add_list]
            )
            # We should never use these since data are preloaded, let's just
            # set it to something large and likely to break (2 ** 31 - 1)
            extra_idx = [2147483647] * sum(info["nchan"] for info in infos[1:])
            assert all(len(r) == infos[0]["nchan"] for r in self._read_picks)
            self._read_picks = [
                np.concatenate([r, extra_idx]) for r in self._read_picks
            ]
            assert all(len(r) == self.info["nchan"] for r in self._read_picks)
            for other in add_list:
                self._orig_units.update(other._orig_units)
        elif isinstance(self, BaseEpochs):
            self.picks = np.arange(self._data.shape[1])
            if hasattr(self, "_projector"):
                activate = False if self._do_delayed_proj else self.proj
                self._projector, self.info = setup_proj(
                    self.info, False, activate=activate
                )

        return self

    @fill_doc
    def add_reference_channels(self, ref_channels):
        """Add reference channels to data that consists of all zeros.

        Adds reference channels to data that were not included during
        recording. This is useful when you need to re-reference your data
        to different channels. These added channels will consist of all zeros.

        Parameters
        ----------
        %(ref_channels)s

        Returns
        -------
        inst : instance of Raw | Epochs | Evoked
               The modified instance.
        """
        return add_reference_channels(self, ref_channels, copy=False)


class InterpolationMixin:
    """Mixin class for Raw, Evoked, Epochs."""

    @verbose
    def interpolate_bads(
        self,
        reset_bads=True,
        mode="accurate",
        origin="auto",
        method=None,
        exclude=(),
        verbose=None,
    ):
        """Interpolate bad MEG and EEG channels.

        Operates in place.

        Parameters
        ----------
        reset_bads : bool
            If True, remove the bads from info.
        mode : str
            Either ``'accurate'`` or ``'fast'``, determines the quality of the
            Legendre polynomial expansion used for interpolation of channels
            using the minimum-norm method.
        origin : array-like, shape (3,) | str
            Origin of the sphere in the head coordinate frame and in meters.
            Can be ``'auto'`` (default), which means a head-digitization-based
            origin fit.

            .. versionadded:: 0.17
        method : dict | str | None
            Method to use for each channel type.

            - ``"meg"`` channels support ``"MNE"`` (default) and ``"nan"``
            - ``"eeg"`` channels support ``"spline"`` (default), ``"MNE"`` and ``"nan"``
            - ``"fnirs"`` channels support ``"nearest"`` (default) and ``"nan"``
            - ``"ecog"`` channels support ``"spline"`` (default) and ``"nan"``
            - ``"seeg"`` channels support ``"spline"`` (default) and ``"nan"``

            None is an alias for::

                method=dict(meg="MNE", eeg="spline", fnirs="nearest")

            If a :class:`str` is provided, the method will be applied to all channel
            types supported and available in the instance. The method ``"nan"`` will
            replace the channel data with ``np.nan``.

            .. warning::
                Be careful when using ``method="nan"``; the default value
                ``reset_bads=True`` may not be what you want.

            .. versionadded:: 0.21
        exclude : list | tuple
            The channels to exclude from interpolation. If excluded a bad
            channel will stay in bads.
        %(verbose)s

        Returns
        -------
        inst : instance of Raw, Epochs, or Evoked
            The modified instance.

        Notes
        -----
        The ``"MNE"`` method uses minimum-norm projection to a sphere and back.

        .. versionadded:: 0.9.0
        """
        from .interpolation import (
            _interpolate_bads_ecog,
            _interpolate_bads_eeg,
            _interpolate_bads_meeg,
            _interpolate_bads_nan,
            _interpolate_bads_nirs,
            _interpolate_bads_seeg,
        )

        _check_preload(self, "interpolation")
        _validate_type(method, (dict, str, None), "method")
        method = _handle_default("interpolation_method", method)
        ch_types = self.get_channel_types(unique=True)
        # figure out if we have "mag" for "meg", "hbo" for "fnirs", ... to filter the
        # "method" dictionary and keep only keys that correspond to existing channels.
        for ch_type in ("meg", "fnirs"):
            for sub_ch_type in _second_rules[ch_type][1].values():
                if sub_ch_type in ch_types:
                    ch_types.remove(sub_ch_type)
                    if ch_type not in ch_types:
                        ch_types.append(ch_type)
        keys2delete = set(method) - set(ch_types)
        for key in keys2delete:
            del method[key]
        valids = {
            "eeg": ("spline", "MNE", "nan"),
            "meg": ("MNE", "nan"),
            "fnirs": ("nearest", "nan"),
            "ecog": ("spline", "nan"),
            "seeg": ("spline", "nan"),
        }
        for key in method:
            _check_option("method[key]", key, tuple(valids))
            _check_option(f"method['{key}']", method[key], valids[key])
        logger.info("Setting channel interpolation method to %s.", method)
        idx = _picks_to_idx(self.info, list(method), exclude=(), allow_empty=True)
        if idx.size == 0 or len(pick_info(self.info, idx)["bads"]) == 0:
            warn("No bad channels to interpolate. Doing nothing...")
            return self
        for ch_type in method.copy():
            idx = _picks_to_idx(self.info, ch_type, exclude=(), allow_empty=True)
            if len(pick_info(self.info, idx)["bads"]) == 0:
                method.pop(ch_type)
        logger.info("Interpolating bad channels.")
        needs_origin = [key != "seeg" and val != "nan" for key, val in method.items()]
        if any(needs_origin):
            origin = _check_origin(origin, self.info)
        for ch_type, interp in method.items():
            if interp == "nan":
                _interpolate_bads_nan(self, ch_type=ch_type, exclude=exclude)
        if method.get("eeg", "") == "spline":
            _interpolate_bads_eeg(self, origin=origin, exclude=exclude)
        meg_mne = method.get("meg", "") == "MNE"
        eeg_mne = method.get("eeg", "") == "MNE"
        if meg_mne or eeg_mne:
            _interpolate_bads_meeg(
                self,
                mode=mode,
                meg=meg_mne,
                eeg=eeg_mne,
                origin=origin,
                exclude=exclude,
                method=method,
            )
        if method.get("fnirs", "") == "nearest":
            _interpolate_bads_nirs(self, exclude=exclude)
        if method.get("ecog", "") == "spline":
            _interpolate_bads_ecog(self, origin=origin, exclude=exclude)
        if method.get("seeg", "") == "spline":
            _interpolate_bads_seeg(self, exclude=exclude)

        if reset_bads is True:
            if "nan" in method.values():
                warn(
                    "interpolate_bads was called with method='nan' and "
                    "reset_bads=True. Consider setting reset_bads=False so that the "
                    "nan-containing channels can be easily excluded from later "
                    "computations."
                )
            self.info["bads"] = [ch for ch in self.info["bads"] if ch in exclude]

        return self

    def interpolate_to(self, sensors, origin="auto", method="spline", reg=0.0):
        """Interpolate EEG data onto a new montage.

        .. warning::
            Be careful, only EEG channels are interpolated. Other channel types are
            not interpolated.

        Parameters
        ----------
        sensors : DigMontage
            The target montage containing channel positions to interpolate onto.
        origin : array-like, shape (3,) | str
            Origin of the sphere in the head coordinate frame and in meters.
            Can be ``'auto'`` (default), which means a head-digitization-based
            origin fit.
        method : str
            Method to use for EEG channels.
            Supported methods are 'spline' (default) and 'MNE'.
        reg : float
            The regularization parameter for the interpolation method
            (only used when the method is 'spline').

        Returns
        -------
        inst : instance of Raw, Epochs, or Evoked
            The instance with updated channel locations and data.

        Notes
        -----
        This method is useful for standardizing EEG layouts across datasets.
        However, some attributes may be lost after interpolation.

        .. versionadded:: 1.10.0
        """
        from ..epochs import BaseEpochs, EpochsArray
        from ..evoked import Evoked, EvokedArray
        from ..forward._field_interpolation import _map_meg_or_eeg_channels
        from ..io import RawArray
        from ..io.base import BaseRaw
        from .interpolation import _make_interpolation_matrix
        from .montage import DigMontage

        # Check that the method option is valid.
        _check_option("method", method, ["spline", "MNE"])
        _validate_type(sensors, DigMontage, "sensors")

        # Get target positions from the montage
        ch_pos = sensors.get_positions().get("ch_pos", {})
        target_ch_names = list(ch_pos.keys())
        if not target_ch_names:
            raise ValueError(
                "The provided sensors configuration has no channel positions."
            )

        # Get original channel order
        orig_names = self.info["ch_names"]

        # Identify EEG channel
        picks_good_eeg = pick_types(self.info, meg=False, eeg=True, exclude="bads")
        if len(picks_good_eeg) == 0:
            raise ValueError("No good EEG channels available for interpolation.")
        # Also get the full list of EEG channel indices (including bad channels)
        picks_remove_eeg = pick_types(self.info, meg=False, eeg=True, exclude=[])
        eeg_names_orig = [orig_names[i] for i in picks_remove_eeg]

        # Identify non-EEG channels in original order
        non_eeg_names_ordered = [ch for ch in orig_names if ch not in eeg_names_orig]

        # Create destination info for new EEG channels
        sfreq = self.info["sfreq"]
        info_interp = create_info(
            ch_names=target_ch_names,
            sfreq=sfreq,
            ch_types=["eeg"] * len(target_ch_names),
        )
        info_interp.set_montage(sensors)
        info_interp["bads"] = [ch for ch in self.info["bads"] if ch in target_ch_names]
        # Do not assign "projs" directly.

        # Compute the interpolation mapping
        if method == "spline":
            origin_val = _check_origin(origin, self.info)
            pos_from = self.info._get_channel_positions(picks_good_eeg) - origin_val
            pos_to = np.stack(list(ch_pos.values()), axis=0)

            def _check_pos_sphere(pos):
                d = np.linalg.norm(pos, axis=-1)
                d_norm = np.mean(d / np.mean(d))
                if np.abs(1.0 - d_norm) > 0.1:
                    warn("Your spherical fit is poor; interpolation may be inaccurate.")

            _check_pos_sphere(pos_from)
            _check_pos_sphere(pos_to)
            mapping = _make_interpolation_matrix(pos_from, pos_to, alpha=reg)

        else:
            assert method == "MNE"
            info_eeg = pick_info(self.info, picks_good_eeg)
            # If the original info has an average EEG reference projector but
            # the destination info does not,
            # update info_interp via a temporary RawArray.
            if _has_eeg_average_ref_proj(self.info) and not _has_eeg_average_ref_proj(
                info_interp
            ):
                # Create dummy data: shape (n_channels, 1)
                temp_data = np.zeros((len(info_interp["ch_names"]), 1))
                temp_raw = RawArray(temp_data, info_interp, first_samp=0)
                # Using the public API, add an average reference projector.
                temp_raw.set_eeg_reference(
                    ref_channels="average", projection=True, verbose=False
                )
                # Extract the updated info.
                info_interp = temp_raw.info
            mapping = _map_meg_or_eeg_channels(
                info_eeg, info_interp, mode="accurate", origin=origin
            )

        # Interpolate EEG data
        data_good = self.get_data(picks=picks_good_eeg)
        data_interp = mapping @ data_good

        # Create a new instance for the interpolated EEG channels
        # TODO: Creating a new instance leads to a loss of information.
        #       We should consider updating the existing instance in the future
        #       by 1) drop channels, 2) add channels, 3) re-order channels.
        if isinstance(self, BaseRaw):
            inst_interp = RawArray(data_interp, info_interp, first_samp=self.first_samp)
        elif isinstance(self, BaseEpochs):
            inst_interp = EpochsArray(data_interp, info_interp)
        else:
            assert isinstance(self, Evoked)
            inst_interp = EvokedArray(data_interp, info_interp)

        # Merge only if non-EEG channels exist
        if not non_eeg_names_ordered:
            return inst_interp

        inst_non_eeg = self.copy().pick(non_eeg_names_ordered).load_data()
        inst_out = inst_non_eeg.add_channels([inst_interp], force_update_info=True)

        # Reorder channels
        # Insert the entire new EEG block at the position of the first EEG channel.
        orig_names_arr = np.array(orig_names)
        mask_eeg = np.isin(orig_names_arr, eeg_names_orig)
        if mask_eeg.any():
            first_eeg_index = np.where(mask_eeg)[0][0]
            pre = orig_names_arr[:first_eeg_index]
            new_eeg = np.array(info_interp["ch_names"])
            post = orig_names_arr[first_eeg_index:]
            post = post[~np.isin(orig_names_arr[first_eeg_index:], eeg_names_orig)]
            new_order = np.concatenate((pre, new_eeg, post)).tolist()
        else:
            new_order = orig_names
        inst_out.reorder_channels(new_order)
        return inst_out


@verbose
def rename_channels(
    info, mapping, allow_duplicates=False, *, on_missing="raise", verbose=None
):
    """Rename channels.

    Parameters
    ----------
    %(info_not_none)s Note: modified in place.
    %(mapping_rename_channels_duplicates)s
    %(on_missing_ch_names)s

        .. versionadded:: 1.11.0
    %(verbose)s

    See Also
    --------
    mne.channels.equalize_channels
    mne.channels.unify_bad_channels
    mne.channels.combine_channels
    """
    _validate_type(info, Info, "info")
    info._check_consistency()
    bads = list(info["bads"])  # make our own local copies
    ch_names = list(info["ch_names"])

    # first check and assemble clean mappings of index and name
    if isinstance(mapping, dict):
        if on_missing in ["warn", "ignore"]:
            new_mapping = {
                ch_old: ch_new
                for ch_old, ch_new in mapping.items()
                if ch_old in ch_names
            }
        else:
            new_mapping = mapping

        if new_mapping != mapping and on_missing == "warn":
            warn(
                "Channel rename map contains keys that are not present in the object "
                "to be renamed. These will be ignored."
            )

        _check_dict_keys(
            new_mapping,
            ch_names,
            key_description="channel name(s)",
            valid_key_source="info",
        )
        new_names = [
            (ch_names.index(ch_name), new_name)
            for ch_name, new_name in new_mapping.items()
        ]
    elif callable(mapping):
        new_names = [(ci, mapping(ch_name)) for ci, ch_name in enumerate(ch_names)]
    else:
        raise ValueError(f"mapping must be callable or dict, not {type(mapping)}")

    # check we got all strings out of the mapping
    for new_name in new_names:
        _validate_type(new_name[1], "str", "New channel mappings")

    # do the remapping locally
    for c_ind, new_name in new_names:
        for bi, bad in enumerate(bads):
            if bad == ch_names[c_ind]:
                bads[bi] = new_name
        ch_names[c_ind] = new_name

    # check that all the channel names are unique
    if len(ch_names) != len(np.unique(ch_names)) and not allow_duplicates:
        raise ValueError("New channel names are not unique, renaming failed")

    # do the remapping in info
    info["bads"] = []
    ch_names_mapping = dict()
    for ch, ch_name in zip(info["chs"], ch_names):
        ch_names_mapping[ch["ch_name"]] = ch_name
        ch["ch_name"] = ch_name
    # .get b/c fwd info omits it
    _rename_comps(info.get("comps", []), ch_names_mapping)
    if "projs" in info:  # fwd might omit it
        for proj in info["projs"]:
            proj["data"]["col_names"][:] = _rename_list(
                proj["data"]["col_names"], ch_names_mapping
            )
    info._update_redundant()
    info["bads"] = bads
    info._check_consistency()


def _recursive_flatten(cell, dtype):
    """Unpack mat files in Python."""
    if len(cell) > 0:
        while not isinstance(cell[0], dtype):
            cell = [c for d in cell for c in d]
    return cell


@dataclass
class _BuiltinChannelAdjacency:
    name: str
    description: str
    fname: str
    source_url: str | None


_ft_neighbor_url_t = string.Template(
    "https://github.com/fieldtrip/fieldtrip/raw/master/template/neighbours/$fname"
)

_BUILTIN_CHANNEL_ADJACENCIES = [
    _BuiltinChannelAdjacency(
        name="biosemi16",
        description="Biosemi 16-electrode cap",
        fname="biosemi16_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="biosemi16_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="biosemi32",
        description="Biosemi 32-electrode cap",
        fname="biosemi32_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="biosemi32_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="biosemi64",
        description="Biosemi 64-electrode cap",
        fname="biosemi64_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="biosemi64_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="bti148",
        description="BTI 148-channel system",
        fname="bti148_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="bti148_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="bti248",
        description="BTI 248-channel system",
        fname="bti248_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="bti248_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="bti248grad",
        description="BTI 248 gradiometer system",
        fname="bti248grad_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="bti248grad_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="ctf64",
        description="CTF 64 axial gradiometer",
        fname="ctf64_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="ctf64_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="ctf151",
        description="CTF 151 axial gradiometer",
        fname="ctf151_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="ctf151_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="ctf275",
        description="CTF 275 axial gradiometer",
        fname="ctf275_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="ctf275_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="easycap32ch-avg",
        description="",
        fname="easycap32ch-avg_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="easycap32ch-avg_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="easycap64ch-avg",
        description="",
        fname="easycap64ch-avg_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="easycap64ch-avg_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="easycap128ch-avg",
        description="",
        fname="easycap128ch-avg_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="easycap128ch-avg_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="easycapM1",
        description="Easycap M1",
        fname="easycapM1_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="easycapM1_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="easycapM11",
        description="Easycap M11",
        fname="easycapM11_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="easycapM11_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="easycapM14",
        description="Easycap M14",
        fname="easycapM14_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="easycapM14_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="easycapM15",
        description="Easycap M15",
        fname="easycapM15_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="easycapM15_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="KIT-157",
        description="",
        fname="KIT-157_neighb.mat",
        source_url=None,
    ),
    _BuiltinChannelAdjacency(
        name="KIT-208",
        description="",
        fname="KIT-208_neighb.mat",
        source_url=None,
    ),
    _BuiltinChannelAdjacency(
        name="KIT-NYU-2019",
        description="",
        fname="KIT-NYU-2019_neighb.mat",
        source_url=None,
    ),
    _BuiltinChannelAdjacency(
        name="KIT-UMD-1",
        description="",
        fname="KIT-UMD-1_neighb.mat",
        source_url=None,
    ),
    _BuiltinChannelAdjacency(
        name="KIT-UMD-2",
        description="",
        fname="KIT-UMD-2_neighb.mat",
        source_url=None,
    ),
    _BuiltinChannelAdjacency(
        name="KIT-UMD-3",
        description="",
        fname="KIT-UMD-3_neighb.mat",
        source_url=None,
    ),
    _BuiltinChannelAdjacency(
        name="KIT-UMD-4",
        description="",
        fname="KIT-UMD-4_neighb.mat",
        source_url=None,
    ),
    _BuiltinChannelAdjacency(
        name="neuromag306mag",
        description="Neuromag306, only magnetometers",
        fname="neuromag306mag_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="neuromag306mag_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="neuromag306planar",
        description="Neuromag306, only planar gradiometers",
        fname="neuromag306planar_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="neuromag306planar_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="neuromag122cmb",
        description="Neuromag122, only combined planar gradiometers",
        fname="neuromag122cmb_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="neuromag122cmb_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="neuromag306cmb",
        description="Neuromag306, only combined planar gradiometers",
        fname="neuromag306cmb_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="neuromag306cmb_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="ecog256",
        description="ECOG 256channels, average referenced",
        fname="ecog256_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="ecog256_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="ecog256bipolar",
        description="ECOG 256channels, bipolar referenced",
        fname="ecog256bipolar_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="ecog256bipolar_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="eeg1010_neighb",
        description="",
        fname="eeg1010_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="eeg1010_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="elec1005",
        description="Standard 10-05 system",
        fname="elec1005_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="elec1005_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="elec1010",
        description="Standard 10-10 system",
        fname="elec1010_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="elec1010_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="elec1020",
        description="Standard 10-20 system",
        fname="elec1020_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="elec1020_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="itab28",
        description="ITAB 28-channel system",
        fname="itab28_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="itab28_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="itab153",
        description="ITAB 153-channel system",
        fname="itab153_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="itab153_neighb.mat"),
    ),
    _BuiltinChannelAdjacency(
        name="language29ch-avg",
        description="MPI for Psycholinguistic: Averaged 29-channel cap",
        fname="language29ch-avg_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="language29ch-avg_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="mpi_59_channels",
        description="MPI for Psycholinguistic: 59-channel cap",
        fname="mpi_59_channels_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="mpi_59_channels_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="yokogawa160",
        description="",
        fname="yokogawa160_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="yokogawa160_neighb.mat"),  # noqa: E501
    ),
    _BuiltinChannelAdjacency(
        name="yokogawa440",
        description="",
        fname="yokogawa440_neighb.mat",
        source_url=_ft_neighbor_url_t.substitute(fname="yokogawa440_neighb.mat"),  # noqa: E501
    ),
]


@fill_doc
def get_builtin_ch_adjacencies(*, descriptions=False):
    """Get a list of all FieldTrip neighbor definitions shipping with MNE.

    The names of the these neighbor definitions can be passed to
    :func:`read_ch_adjacency`.

    Parameters
    ----------
    descriptions : bool
        Whether to return not only the neighbor definition names, but also
        their corresponding descriptions. If ``True``, a list of tuples is
        returned, where the first tuple element is the neighbor definition name
        and the second is the description. If ``False`` (default), only the
        names are returned.

    Returns
    -------
    neighbor_name : list of str | list of tuple
        If ``descriptions=False``, the names of all builtin FieldTrip neighbor
        definitions that can be loaded directly via :func:`read_ch_adjacency`.

        If ``descriptions=True``, a list of tuples ``(name, description)``.

    Notes
    -----
    .. versionadded:: 1.1
    """
    if descriptions:
        return sorted(
            [(m.name, m.description) for m in _BUILTIN_CHANNEL_ADJACENCIES],
            key=lambda x: x[0].casefold(),  # only sort based on name
        )
    else:
        return sorted([m.name for m in _BUILTIN_CHANNEL_ADJACENCIES], key=str.casefold)


@fill_doc
def read_ch_adjacency(fname, picks=None):
    """Read a channel adjacency ("neighbors") file that ships with MNE.

    More information on these neighbor definitions can be found on the related
    `FieldTrip documentation pages
    <http://www.fieldtriptoolbox.org/template/neighbours/>`__.

    Parameters
    ----------
    fname : path-like | str
        The path to the file to load, or the name of a channel adjacency
        matrix that ships with MNE-Python.

        .. note::
            You can retrieve the names of all
            built-in channel adjacencies via
            :func:`mne.channels.get_builtin_ch_adjacencies`.
    %(picks_all_notypes)s

    Returns
    -------
    ch_adjacency : scipy.sparse.csr_array, shape (n_channels, n_channels)
        The adjacency matrix.
    ch_names : list
        The list of channel names present in adjacency matrix.

    See Also
    --------
    get_builtin_ch_adjacencies
    mne.viz.plot_ch_adjacency
    find_ch_adjacency
    mne.stats.combine_adjacency

    Notes
    -----
    If the neighbor definition you need is not shipped by MNE-Python,
    you may use :func:`find_ch_adjacency` to compute the
    adjacency matrix based on your 2D sensor locations.

    Note that depending on your use case, you may need to additionally use
    :func:`mne.stats.combine_adjacency` to prepare a final "adjacency"
    to pass to the eventual function.
    """
    if op.isabs(fname):
        fname = str(
            _check_fname(
                fname=fname,
                overwrite="read",
                must_exist=True,
            )
        )
    else:  # built-in FieldTrip neighbors
        ch_adj_name = fname
        del fname
        if ch_adj_name.endswith("_neighb.mat"):  # backward-compat
            ch_adj_name = ch_adj_name.replace("_neighb.mat", "")

        if ch_adj_name not in get_builtin_ch_adjacencies():
            raise ValueError(
                f"No built-in channel adjacency matrix found with name: "
                f"{ch_adj_name}. Valid names are: "
                f"{', '.join(get_builtin_ch_adjacencies())}"
            )

        ch_adj = [a for a in _BUILTIN_CHANNEL_ADJACENCIES if a.name == ch_adj_name][0]
        fname = ch_adj.fname
        templates_dir = Path(__file__).resolve().parent / "data" / "neighbors"
        fname = str(
            _check_fname(  # only needed to convert to a string
                fname=templates_dir / fname,
                overwrite="read",
                must_exist=True,
            )
        )

    nb = loadmat(fname)["neighbours"]
    ch_names = _recursive_flatten(nb["label"], str)
    temp_info = create_info(ch_names, 1.0)
    picks = _picks_to_idx(temp_info, picks, none="all")
    neighbors = [_recursive_flatten(c, str) for c in nb["neighblabel"].flatten()]
    assert len(ch_names) == len(neighbors)
    adjacency = _ch_neighbor_adjacency(ch_names, neighbors)
    # picking before constructing matrix is buggy
    adjacency = adjacency[picks][:, picks]
    ch_names = [ch_names[p] for p in picks]

    return adjacency, ch_names


def _ch_neighbor_adjacency(ch_names, neighbors):
    """Compute sensor adjacency matrix.

    Parameters
    ----------
    ch_names : list of str
        The channel names.
    neighbors : list of list
        A list of list of channel names. The neighbors to
        which the channels in ch_names are connected with.
        Must be of the same length as ch_names.

    Returns
    -------
    ch_adjacency : scipy.sparse.spmatrix
        The adjacency matrix.
    """
    if len(ch_names) != len(neighbors):
        raise ValueError("`ch_names` and `neighbors` must have the same length")
    set_neighbors = {c for d in neighbors for c in d}
    rest = set_neighbors - set(ch_names)
    if len(rest) > 0:
        raise ValueError(
            "Some of your neighbors are not present in the list of channel names"
        )

    for neigh in neighbors:
        if not isinstance(neigh, list) and not all(isinstance(c, str) for c in neigh):
            raise ValueError("`neighbors` must be a list of lists of str")

    ch_adjacency = np.eye(len(ch_names), dtype=bool)
    for ii, neigbs in enumerate(neighbors):
        ch_adjacency[ii, [ch_names.index(i) for i in neigbs]] = True
    ch_adjacency = csr_array(ch_adjacency)
    return ch_adjacency


@fill_doc
def find_ch_adjacency(info, ch_type):
    """Find the adjacency matrix for the given channels.

    This function tries to infer the appropriate adjacency matrix template
    for the given channels. If a template is not found, the adjacency matrix
    is computed using Delaunay triangulation based on 2D sensor locations.

    Parameters
    ----------
    %(info_not_none)s
    ch_type : str | None
        The channel type for computing the adjacency matrix. Currently
        supports ``'mag'``, ``'grad'``, ``'eeg'`` and ``None``.
        If ``None``, the info must contain only one channel type.

    Returns
    -------
    ch_adjacency : scipy.sparse.csr_array, shape (n_channels, n_channels)
        The adjacency matrix.
    ch_names : list
        The list of channel names present in adjacency matrix.

    See Also
    --------
    mne.viz.plot_ch_adjacency
    mne.stats.combine_adjacency
    get_builtin_ch_adjacencies
    read_ch_adjacency

    Notes
    -----
    .. versionadded:: 0.15

    Automatic detection of an appropriate adjacency matrix template only
    works for MEG data at the moment. This means that the adjacency matrix
    is always computed for EEG data and never loaded from a template file. If
    you want to load a template for a given montage use
    :func:`read_ch_adjacency` directly.

    .. warning::
        If Delaunay triangulation is used to calculate the adjacency matrix it
        may yield partially unexpected results (e.g., include unwanted edges
        between non-adjacent sensors). Therefore, it is recommended to check
        (and, if necessary, manually modify) the result by inspecting it
        via :func:`mne.viz.plot_ch_adjacency`.

    Note that depending on your use case, you may need to additionally use
    :func:`mne.stats.combine_adjacency` to prepare a final "adjacency"
    to pass to the eventual function.
    """
    from ..io.kit.constants import KIT_NEIGHBORS

    if ch_type is None:
        picks = channel_indices_by_type(info)
        if sum([len(p) != 0 for p in picks.values()]) != 1:
            raise ValueError(
                "info must contain only one channel type if ch_type is None."
            )
        ch_type = channel_type(info, 0)
    else:
        _check_option("ch_type", ch_type, ["mag", "grad", "eeg"])
    (
        has_vv_mag,
        has_vv_grad,
        is_old_vv,
        has_4D_mag,
        ctf_other_types,
        has_CTF_grad,
        n_kit_grads,
        has_any_meg,
        has_eeg_coils,
        has_eeg_coils_and_meg,
        has_eeg_coils_only,
        has_neuromag_122_grad,
        has_csd_coils,
    ) = _get_ch_info(info)
    conn_name = None
    if has_vv_mag and ch_type == "mag":
        conn_name = "neuromag306mag"
    elif has_vv_grad and ch_type == "grad":
        conn_name = "neuromag306planar"
    elif has_4D_mag:
        if "MEG 248" in info["ch_names"]:
            idx = info["ch_names"].index("MEG 248")
            grad = info["chs"][idx]["coil_type"] == FIFF.FIFFV_COIL_MAGNES_GRAD
            mag = info["chs"][idx]["coil_type"] == FIFF.FIFFV_COIL_MAGNES_MAG
            if ch_type == "grad" and grad:
                conn_name = "bti248grad"
            elif ch_type == "mag" and mag:
                conn_name = "bti248"
        elif "MEG 148" in info["ch_names"] and ch_type == "mag":
            idx = info["ch_names"].index("MEG 148")
            if info["chs"][idx]["coil_type"] == FIFF.FIFFV_COIL_MAGNES_MAG:
                conn_name = "bti148"
    elif has_CTF_grad and ch_type == "mag":
        if info["nchan"] < 100:
            conn_name = "ctf64"
        elif info["nchan"] > 200:
            conn_name = "ctf275"
        else:
            conn_name = "ctf151"
    elif n_kit_grads > 0:
        conn_name = KIT_NEIGHBORS.get(info["kit_system_id"])

    if conn_name is not None:
        logger.info(f"Reading adjacency matrix for {conn_name}.")
        adjacency, ch_names = read_ch_adjacency(conn_name)
        if conn_name.startswith("neuromag") and info["ch_names"][0].startswith("MEG "):
            ch_names = [ch_name.replace("MEG", "MEG ") for ch_name in ch_names]
        return adjacency, ch_names
    logger.info(
        "Could not find a adjacency matrix for the data. "
        "Computing adjacency based on Delaunay triangulations."
    )
    return _compute_ch_adjacency(info, ch_type)


@fill_doc
def _compute_ch_adjacency(info, ch_type):
    """Compute channel adjacency matrix using Delaunay triangulations.

    Parameters
    ----------
    %(info_not_none)s
    ch_type : str
        The channel type for computing the adjacency matrix. Currently
        supports ``'mag'``, ``'grad'`` and ``'eeg'``.

    Returns
    -------
    ch_adjacency : scipy.sparse.csr_array, shape (n_channels, n_channels)
        The adjacency matrix.
    ch_names : list
        The list of channel names present in adjacency matrix.
    """
    from ..channels.layout import _find_topomap_coords, _pair_grad_sensors
    from ..source_estimate import spatial_tris_adjacency

    combine_grads = ch_type == "grad" and any(
        [
            coil_type in [ch["coil_type"] for ch in info["chs"]]
            for coil_type in [FIFF.FIFFV_COIL_VV_PLANAR_T1, FIFF.FIFFV_COIL_NM_122]
        ]
    )

    picks = dict(_picks_by_type(info, exclude=[]))[ch_type]
    ch_names = [info["ch_names"][pick] for pick in picks]
    if combine_grads:
        pairs = _pair_grad_sensors(info, topomap_coords=False, exclude=[])
        if len(pairs) != len(picks):
            raise RuntimeError(
                "Cannot find a pair for some of the "
                "gradiometers. Cannot compute adjacency "
                "matrix."
            )
        # only for one of the pair
        xy = _find_topomap_coords(info, picks[::2], sphere=HEAD_SIZE_DEFAULT)
    else:
        xy = _find_topomap_coords(info, picks, sphere=HEAD_SIZE_DEFAULT)
    tri = Delaunay(xy)
    neighbors = spatial_tris_adjacency(tri.simplices)

    if combine_grads:
        ch_adjacency = np.eye(len(picks), dtype=bool)
        for idx, neigbs in zip(neighbors.row, neighbors.col):
            for ii in range(2):  # make sure each pair is included
                for jj in range(2):
                    ch_adjacency[idx * 2 + ii, neigbs * 2 + jj] = True
                    ch_adjacency[idx * 2 + ii, idx * 2 + jj] = True  # pair
        ch_adjacency = csr_array(ch_adjacency)
    else:
        ch_adjacency = lil_array(neighbors)
        ch_adjacency.setdiag(np.repeat(1, ch_adjacency.shape[0]))
        ch_adjacency = ch_adjacency.tocsr()

    return ch_adjacency, ch_names


@fill_doc
def fix_mag_coil_types(info, use_cal=False):
    """Fix magnetometer coil types.

    Parameters
    ----------
    %(info_not_none)s Corrections are done in-place.
    use_cal : bool
        If True, further refine the check for old coil types by checking
        ``info['chs'][ii]['cal']``.

    Notes
    -----
    This function changes magnetometer coil types 3022 (T1: SQ20483N) and
    3023 (T2: SQ20483-A) to 3024 (T3: SQ20950N) in the channel definition
    records in the info structure.

    Neuromag Vectorview systems can contain magnetometers with two
    different coil sizes (3022 and 3023 vs. 3024). The systems
    incorporating coils of type 3024 were introduced last and are used at
    the majority of MEG sites. At some sites with 3024 magnetometers,
    the data files have still defined the magnetometers to be of type
    3022 to ensure compatibility with older versions of Neuromag software.
    In the MNE software as well as in the present version of Neuromag
    software coil type 3024 is fully supported. Therefore, it is now safe
    to upgrade the data files to use the true coil type.

    .. note:: The effect of the difference between the coil sizes on the
              current estimates computed by the MNE software is very small.
              Therefore the use of ``fix_mag_coil_types`` is not mandatory.
    """
    old_mag_inds = _get_T1T2_mag_inds(info, use_cal)
    n_mag = len(pick_types(info, meg="mag", exclude=[]))
    for ii in old_mag_inds:
        info["chs"][ii]["coil_type"] = FIFF.FIFFV_COIL_VV_MAG_T3
    logger.info(f"{len(old_mag_inds)} of {n_mag} magnetometer types replaced with T3.")
    info._check_consistency()


def _get_T1T2_mag_inds(info, use_cal=False):
    """Find T1/T2 magnetometer coil types."""
    picks = pick_types(info, meg="mag", exclude=[])
    old_mag_inds = []
    # From email exchanges, systems with the larger T2 coil only use the cal
    # value of 2.09e-11. Newer T3 magnetometers use 4.13e-11 or 1.33e-10
    # (Triux). So we can use a simple check for > 3e-11.
    for ii in picks:
        ch = info["chs"][ii]
        if ch["coil_type"] in (FIFF.FIFFV_COIL_VV_MAG_T1, FIFF.FIFFV_COIL_VV_MAG_T2):
            if use_cal:
                if ch["cal"] > 3e-11:
                    old_mag_inds.append(ii)
            else:
                old_mag_inds.append(ii)
    return old_mag_inds


def _get_ch_info(info):
    """Get channel info for inferring acquisition device."""
    chs = info["chs"]
    # Only take first 16 bits, as higher bits store CTF comp order
    coil_types = {ch["coil_type"] & 0xFFFF for ch in chs}
    channel_types = {ch["kind"] for ch in chs}

    has_vv_mag = any(
        k in coil_types
        for k in [
            FIFF.FIFFV_COIL_VV_MAG_T1,
            FIFF.FIFFV_COIL_VV_MAG_T2,
            FIFF.FIFFV_COIL_VV_MAG_T3,
        ]
    )
    has_vv_grad = any(
        k in coil_types
        for k in [
            FIFF.FIFFV_COIL_VV_PLANAR_T1,
            FIFF.FIFFV_COIL_VV_PLANAR_T2,
            FIFF.FIFFV_COIL_VV_PLANAR_T3,
        ]
    )
    has_neuromag_122_grad = any(k in coil_types for k in [FIFF.FIFFV_COIL_NM_122])

    is_old_vv = " " in chs[0]["ch_name"]

    has_4D_mag = FIFF.FIFFV_COIL_MAGNES_MAG in coil_types
    ctf_other_types = (
        FIFF.FIFFV_COIL_CTF_REF_MAG,
        FIFF.FIFFV_COIL_CTF_REF_GRAD,
        FIFF.FIFFV_COIL_CTF_OFFDIAG_REF_GRAD,
    )
    has_CTF_grad = FIFF.FIFFV_COIL_CTF_GRAD in coil_types or (
        FIFF.FIFFV_MEG_CH in channel_types
        and any(k in ctf_other_types for k in coil_types)
    )
    # hack due to MNE-C bug in IO of CTF
    # only take first 16 bits, as higher bits store CTF comp order
    n_kit_grads = sum(
        ch["coil_type"] & 0xFFFF == FIFF.FIFFV_COIL_KIT_GRAD for ch in chs
    )

    has_any_meg = any([has_vv_mag, has_vv_grad, has_4D_mag, has_CTF_grad, n_kit_grads])
    has_eeg_coils = (
        FIFF.FIFFV_COIL_EEG in coil_types and FIFF.FIFFV_EEG_CH in channel_types
    )
    has_eeg_coils_and_meg = has_eeg_coils and has_any_meg
    has_eeg_coils_only = has_eeg_coils and not has_any_meg
    has_csd_coils = (
        FIFF.FIFFV_COIL_EEG_CSD in coil_types and FIFF.FIFFV_EEG_CH in channel_types
    )

    return (
        has_vv_mag,
        has_vv_grad,
        is_old_vv,
        has_4D_mag,
        ctf_other_types,
        has_CTF_grad,
        n_kit_grads,
        has_any_meg,
        has_eeg_coils,
        has_eeg_coils_and_meg,
        has_eeg_coils_only,
        has_neuromag_122_grad,
        has_csd_coils,
    )


@fill_doc
def make_1020_channel_selections(info, midline="z", *, return_ch_names=False):
    """Map hemisphere names to corresponding EEG channel names or indices.

    This function uses a simple heuristic to separate channel names into three
    Region of Interest-based selections: ``Left``, ``Midline`` and ``Right``.

    The heuristic is that any of the channel names ending
    with odd numbers are filed under ``Left``; those ending with even numbers
    are filed under ``Right``; and those ending with the character(s) specified
    in ``midline`` are filed under ``Midline``. Other channels are ignored.

    This is appropriate for 10/20, 10/10, 10/05, …, sensor arrangements, but
    not for other naming conventions.

    Parameters
    ----------
    %(info_not_none)s If channel locations are present, the channel lists will
        be sorted from posterior to anterior; otherwise, the order specified in
        ``info["ch_names"]`` will be kept.
    midline : str
        Names ending in any of these characters are stored under the
        ``Midline`` key. Defaults to ``'z'``. Capitalization is ignored.
    return_ch_names : bool
        Whether to return channel names instead of channel indices.

        .. versionadded:: 1.4.0

    Returns
    -------
    selections : dict
        A dictionary mapping from region of interest name to a list of channel
        indices (if ``return_ch_names=False``) or to a list of channel names
        (if ``return_ch_names=True``).
    """
    _validate_type(info, "info")

    try:
        from .layout import find_layout

        layout = find_layout(info)
        pos = layout.pos
        ch_names = layout.names
    except RuntimeError:  # no channel positions found
        ch_names = info["ch_names"]
        pos = None

    selections = dict(Left=[], Midline=[], Right=[])
    for pick, channel in enumerate(ch_names):
        last_char = channel[-1].lower()  # in 10/20, last char codes hemisphere
        if last_char in midline:
            selection = "Midline"
        elif last_char.isdigit():
            selection = "Left" if int(last_char) % 2 else "Right"
        else:  # ignore the channel
            continue
        selections[selection].append(pick)

    if pos is not None:
        # sort channels from front to center
        # (y-coordinate of the position info in the layout)
        selections = {
            selection: np.array(picks)[pos[picks, 1].argsort()]
            for selection, picks in selections.items()
        }

    # convert channel indices to names if requested
    if return_ch_names:
        for selection, ch_indices in selections.items():
            selections[selection] = [info.ch_names[idx] for idx in ch_indices]

    return selections


@verbose
def combine_channels(
    inst,
    groups,
    method="mean",
    keep_stim=False,
    drop_bad=False,
    *,
    on_missing="raise",
    verbose=None,
):
    """Combine channels based on specified channel grouping.

    Parameters
    ----------
    inst : instance of Raw, Epochs, or Evoked
        An MNE-Python object to combine the channels for. The object can be of
        type Raw, Epochs, or Evoked.
    groups : dict
        Specifies which channels are aggregated into a single channel, with
        aggregation method determined by the ``method`` parameter. One new
        pseudo-channel is made per dict entry; the dict values must be lists of
        picks (integer indices of ``ch_names``). For example::

            groups=dict(Left=[1, 2, 3, 4], Right=[5, 6, 7, 8])

        Note that within a dict entry all channels must have the same type.
    method : str | callable
        Which method to use to combine channels. If a :class:`str`, must be one
        of 'mean', 'median', or 'std' (standard deviation). If callable, the
        callable must accept one positional input (data of shape ``(n_channels,
        n_times)``, or ``(n_epochs, n_channels, n_times)``) and return an
        :class:`array <numpy.ndarray>` of shape ``(n_times,)``, or ``(n_epochs,
        n_times)``. For example with an instance of Raw or Evoked::

            method = lambda data: np.mean(data, axis=0)

        Another example with an instance of Epochs::

            method = lambda data: np.median(data, axis=1)

        Defaults to ``'mean'``.
    keep_stim : bool
        If ``True``, include stimulus channels in the resulting object.
        Defaults to ``False``.
    drop_bad : bool
        If ``True``, drop channels marked as bad before combining. Defaults to
        ``False``.
    %(on_missing_epochs)s
        .. versionadded:: 1.11.0
    %(verbose)s

    Returns
    -------
    combined_inst : instance of Raw, Epochs, or Evoked
        An MNE-Python object of the same type as the input ``inst``, containing
        one virtual channel for each group in ``groups`` (and, if ``keep_stim``
        is ``True``, also containing stimulus channels).

    See Also
    --------
    mne.channels.equalize_channels
    mne.channels.rename_channels
    mne.channels.unify_bad_channels
    """
    from ..epochs import BaseEpochs, EpochsArray
    from ..evoked import Evoked, EvokedArray
    from ..io import BaseRaw, RawArray

    ch_axis = 1 if isinstance(inst, BaseEpochs) else 0
    ch_idx = list(range(inst.info["nchan"]))
    ch_names = inst.info["ch_names"]
    ch_types = inst.get_channel_types()
    kwargs = dict()
    if isinstance(inst, BaseEpochs):
        kwargs["copy"] = False
    inst_data = inst.get_data(**kwargs)
    groups = OrderedDict(deepcopy(groups))

    # Convert string values of ``method`` into callables
    # XXX Possibly de-duplicate with _make_combine_callable of mne/viz/utils.py
    if isinstance(method, str):
        method_dict = {
            key: partial(getattr(np, key), axis=ch_axis)
            for key in ("mean", "median", "std")
        }
        try:
            method = method_dict[method]
        except KeyError:
            raise ValueError(
                '"method" must be a callable, or one of "mean", '
                f'"median", or "std"; got "{method}".'
            )

    # Instantiate channel info and data
    new_ch_names, new_ch_types, new_data = [], [], []
    if not isinstance(keep_stim, bool):
        raise TypeError(f'"keep_stim" must be of type bool, not {type(keep_stim)}.')
    if keep_stim:
        stim_ch_idx = list(pick_types(inst.info, meg=False, stim=True))
        if stim_ch_idx:
            new_ch_names = [ch_names[idx] for idx in stim_ch_idx]
            new_ch_types = [ch_types[idx] for idx in stim_ch_idx]
            new_data = [np.take(inst_data, idx, axis=ch_axis) for idx in stim_ch_idx]
        else:
            warn("Could not find stimulus channels.")

    # Get indices of bad channels
    ch_idx_bad = []
    if not isinstance(drop_bad, bool):
        raise TypeError(f'"drop_bad" must be of type bool, not {type(drop_bad)}.')
    if drop_bad and inst.info["bads"]:
        ch_idx_bad = pick_channels(ch_names, inst.info["bads"])

    # Check correctness of combinations
    for this_group, this_picks in groups.items():
        # Check if channel indices are out of bounds
        if not all(idx in ch_idx for idx in this_picks):
            raise ValueError("Some channel indices are out of bounds.")
        # Check if heterogeneous sensor type combinations
        this_ch_type = np.array(ch_types)[this_picks]
        if len(set(this_ch_type)) > 1:
            types = ", ".join(set(this_ch_type))
            raise ValueError(
                "Cannot combine sensors of different types; "
                f'"{this_group}" contains types {types}.'
            )
        # Remove bad channels
        these_bads = [idx for idx in this_picks if idx in ch_idx_bad]
        this_picks = [idx for idx in this_picks if idx not in ch_idx_bad]
        if these_bads:
            logger.info(
                f"Dropped the following channels in group {this_group}: {these_bads}"
            )
        #  Check if combining less than 2 channel
        if len(set(this_picks)) < 2:
            warn(
                f'Less than 2 channels in group "{this_group}" when '
                f'combining by method "{method}".'
            )
        # If all good create more detailed dict without bad channels
        groups[this_group] = dict(picks=this_picks, ch_type=this_ch_type[0])

    # Combine channels and add them to the new instance
    for this_group, this_group_dict in groups.items():
        new_ch_names.append(this_group)
        new_ch_types.append(this_group_dict["ch_type"])
        this_picks = this_group_dict["picks"]
        this_data = np.take(inst_data, this_picks, axis=ch_axis)
        new_data.append(method(this_data))
    new_data = np.swapaxes(new_data, 0, ch_axis)
    info = create_info(
        sfreq=inst.info["sfreq"], ch_names=new_ch_names, ch_types=new_ch_types
    )
    # create new instances and make sure to copy important attributes
    if isinstance(inst, BaseRaw):
        combined_inst = RawArray(new_data, info, first_samp=inst.first_samp)
    elif isinstance(inst, BaseEpochs):
        combined_inst = EpochsArray(
            new_data,
            info,
            events=inst.events,
            event_id=inst.event_id,
            tmin=inst.times[0],
            baseline=inst.baseline,
            on_missing=on_missing,
        )
        if inst.metadata is not None:
            combined_inst.metadata = inst.metadata.copy()
    elif isinstance(inst, Evoked):
        combined_inst = EvokedArray(
            new_data, info, tmin=inst.times[0], baseline=inst.baseline
        )

    return combined_inst


# NeuroMag channel groupings
_SELECTIONS = [
    "Vertex",
    "Left-temporal",
    "Right-temporal",
    "Left-parietal",
    "Right-parietal",
    "Left-occipital",
    "Right-occipital",
    "Left-frontal",
    "Right-frontal",
]
_EEG_SELECTIONS = ["EEG 1-32", "EEG 33-64", "EEG 65-96", "EEG 97-128"]


def _divide_to_regions(info, add_stim=True):
    """Divide channels to regions by positions."""
    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)

    occipital = picks[np.argsort(y[picks])[:chs_in_lobe]]
    picks = np.setdiff1d(picks, occipital)

    temporal = picks[np.argsort(z[picks])[:chs_in_lobe]]
    picks = np.setdiff1d(picks, temporal)

    lt, rt = _divide_side(temporal, x)
    lf, rf = _divide_side(frontal, x)
    lo, ro = _divide_side(occipital, x)
    lp, rp = _divide_side(picks, x)  # Parietal lobe from the remaining picks.

    # Because of the way the sides are divided, there may be outliers in the
    # temporal lobes. Here we switch the sides for these outliers. For other
    # lobes it is not a big problem because of the vicinity of the lobes.
    with np.errstate(invalid="ignore"):  # invalid division, greater compare
        zs = np.abs(zscore(x[rt]))
        outliers = np.array(rt)[np.where(zs > 2.0)[0]]
    rt = list(np.setdiff1d(rt, outliers))

    with np.errstate(invalid="ignore"):  # invalid division, greater compare
        zs = np.abs(zscore(x[lt]))
        outliers = np.append(outliers, (np.array(lt)[np.where(zs > 2.0)[0]]))
    lt = list(np.setdiff1d(lt, outliers))

    l_mean = np.mean(x[lt])
    r_mean = np.mean(x[rt])
    for outlier in outliers:
        if abs(l_mean - x[outlier]) < abs(r_mean - x[outlier]):
            lt.append(outlier)
        else:
            rt.append(outlier)

    if add_stim:
        stim_ch = _get_stim_channel(None, info, raise_error=False)
        if len(stim_ch) > 0:
            for region in [lf, rf, lo, ro, lp, rp, lt, rt]:
                region.append(info["ch_names"].index(stim_ch[0]))
    return OrderedDict(
        [
            ("Left-frontal", lf),
            ("Right-frontal", rf),
            ("Left-parietal", lp),
            ("Right-parietal", rp),
            ("Left-occipital", lo),
            ("Right-occipital", ro),
            ("Left-temporal", lt),
            ("Right-temporal", rt),
        ]
    )


def _divide_side(lobe, x):
    """Make a separation between left and right lobe evenly."""
    lobe = np.asarray(lobe)
    median = np.median(x[lobe])

    left = lobe[np.where(x[lobe] < median)[0]]
    right = lobe[np.where(x[lobe] > median)[0]]
    medians = np.where(x[lobe] == median)[0]

    left = np.sort(np.concatenate([left, lobe[medians[1::2]]]))
    right = np.sort(np.concatenate([right, lobe[medians[::2]]]))
    return list(left), list(right)


@verbose
def read_vectorview_selection(name, fname=None, info=None, verbose=None):
    """Read Neuromag Vector View channel selection from a file.

    Parameters
    ----------
    name : str | list of str
        Name of the selection. If a list, the selections are combined.
        Supported selections are: ``'Vertex'``, ``'Left-temporal'``,
        ``'Right-temporal'``, ``'Left-parietal'``, ``'Right-parietal'``,
        ``'Left-occipital'``, ``'Right-occipital'``, ``'Left-frontal'`` and
        ``'Right-frontal'``. Selections can also be matched and combined by
        spcecifying common substrings. For example, ``name='temporal`` will
        produce a combination of ``'Left-temporal'`` and ``'Right-temporal'``.
    fname : path-like
        Filename of the selection file (if ``None``, built-in selections are
        used).
    %(info)s Used to determine which channel naming convention to use, e.g.
        ``'MEG 0111'`` (with space) for old Neuromag systems and ``'MEG0111'``
        (without space) for new ones.
    %(verbose)s

    Returns
    -------
    sel : list of str
        List with channel names in the selection.
    """
    # convert name to list of string
    if not isinstance(name, list | tuple):
        name = [name]
    if isinstance(info, Info):
        picks = pick_types(info, meg=True, exclude=())
        if len(picks) > 0 and " " not in info["ch_names"][picks[0]]:
            spacing = "new"
        else:
            spacing = "old"
    elif info is not None:
        raise TypeError(f"info must be an instance of Info or None, not {type(info)}")
    else:  # info is None
        spacing = "old"

    # use built-in selections by default
    if fname is None:
        fname = op.join(op.dirname(__file__), "..", "data", "mne_analyze.sel")

    fname = str(_check_fname(fname, must_exist=True, overwrite="read"))

    # use this to make sure we find at least one match for each name
    name_found = {n: False for n in name}
    with open(fname) as fid:
        sel = []
        for line in fid:
            line = line.strip()
            # skip blank lines and comments
            if len(line) == 0 or line[0] == "#":
                continue
            # get the name of the selection in the file
            pos = line.find(":")
            if pos < 0:
                logger.info('":" delimiter not found in selections file, skipping line')
                continue
            sel_name_file = line[:pos]
            # search for substring match with name provided
            for n in name:
                if sel_name_file.find(n) >= 0:
                    sel.extend(line[pos + 1 :].split("|"))
                    name_found[n] = True
                    break

    # make sure we found at least one match for each name
    for n, found in name_found.items():
        if not found:
            raise ValueError(f'No match for selection name "{n}" found')

    # 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