Handle channel name mismatches

CITATION.cff

@@ -225,6 +225,10 @@ authors:
       family-names: Aristimunha
       affiliation: 'Université Paris-Saclay, LISN, Inria TAU/INRIA Nerv, France'
       orcid: 'https://orcid.org/0000-0001-5258-2995'
+    - given-names: Kalle
+      family-names: Mäkelä
+      affiliation: 'University of Helsinki, Helsinki, Finland'
+      orcid: 'https://orcid.org/0009-0005-5706-0842'
     - given-names: Alexandre
       family-names: Gramfort
       affiliation: 'Université Paris-Saclay, Inria, CEA, Palaiseau, France'

doc/authors.rst

@@ -61,3 +61,4 @@
 .. _William Turner: https://bootstrapbill.github.io/
 .. _Yorguin Mantilla: https://github.com/yjmantilla
 .. _Julius Welzel: https://github.com/JuliusWelzel
+.. _Kalle Mäkelä: https://github.com/Kallemakela

doc/whats_new.rst

@@ -22,6 +22,7 @@ The following authors contributed for the first time. Thank you so much! 🤩
 * `Julius Welzel`_
 * `Alex Lopez Marquez`_
 * `Bruno Aristimunha`_
+* `Kalle Mäkelä`_
 
 The following authors had contributed before. Thank you for sticking around! 🤘
 
@@ -34,6 +35,7 @@ Detailed list of changes
 🚀 Enhancements
 ^^^^^^^^^^^^^^^
 
+- TODO
 - :func:`mne_bids.write_raw_bids()` has a new parameter `electrodes_tsv_task` which allows adding the `task` entity to the `electrodes.tsv` filepath, by `Alex Lopez Marquez`_ (:gh:`1424`)
 - Extended the configuration to recognise `motion` as a valid BIDS datatype by `Julius Welzel`_ (:gh:`1430`)
 - Better control of verbosity in several functions, by `Bruno Aristimunha`_ (:gh:`1449`)

mne_bids/read.py

@@ -766,7 +766,7 @@ def _get_bads_from_tsv_data(tsv_data):
     return bads
 
 
-def _handle_channels_reading(channels_fname, raw):
+def _handle_channels_reading(channels_fname, raw, ch_name_mismatch="raise"):
     """Read associated channels.tsv and populate raw.
 
     Updates status (bad) and types of channels.
@@ -847,7 +847,23 @@ def _handle_channels_reading(channels_fname, raw):
             f"set channel names."
         )
     else:
-        raw.rename_channels(dict(zip(raw.ch_names, ch_names_tsv)))
+        orig_names = list(raw.ch_names)
+        if orig_names != ch_names_tsv:
+            if ch_name_mismatch == "raise":
+                raise RuntimeError(
+                    f"Channel mismatch between {channels_fname} and the raw data file detected."
+                )
+            warn(
+                f"Channel mismatch between {channels_fname} and the raw data file detected. Using mismatch strategy: {ch_name_mismatch}."
+            )
+            if ch_name_mismatch == "reorder":
+                raw.reorder_channels(ch_names_tsv)
+            elif ch_name_mismatch == "rename":
+                raw.rename_channels(dict(zip(raw.ch_names, ch_names_tsv)))
+            else:
+                raise ValueError(
+                    "ch_name_mismatch must be one of {'reorder','raise','rename'}"
+                )
 
     # Set the channel types in the raw data according to channels.tsv
     channel_type_bids_mne_map_available_channels = {
@@ -887,7 +903,12 @@ def _handle_channels_reading(channels_fname, raw):
 
 @verbose
 def read_raw_bids(
-    bids_path, extra_params=None, *, return_event_dict=False, verbose=None
+    bids_path,
+    extra_params=None,
+    *,
+    return_event_dict=False,
+    ch_name_mismatch="raise",
+    verbose=None,
 ):
     """Read BIDS compatible data.
 
@@ -918,6 +939,12 @@ def read_raw_bids(
         event IDs, in addition to the :class:`~mne.io.Raw` object. If a ``value`` column
         is present in the ``*_events.tsv`` file, it will be used as the source of the
         integer event ID values (events with ``value="n/a"`` will be omitted).
+    ch_name_mismatch : str
+        What to do if the channel names in the channels.tsv file do not match the channel names in the raw data file.
+        Must be one of {'raise', 'reorder', 'rename'}. Default is 'raise'.
+        - 'raise' will raise a RuntimeError if there is a channel mismatch.
+        - 'reorder' will reorder the channels in the raw data file to match the channel names in the channels.tsv file.
+        - 'rename' will rename the channels in the raw data file to match the channel names in the channels.tsv file.
     %(verbose)s
 
     Returns
@@ -944,6 +971,8 @@ def read_raw_bids(
     ValueError
         If the specified ``datatype`` cannot be found in the dataset.
 
+    RuntimeError
+        If channels.tsv and raw data file have a channel name mismatch and ch_name_mismatch is 'raise'.
     """
     if not isinstance(bids_path, BIDSPath):
         raise RuntimeError(
@@ -1082,7 +1111,9 @@ def read_raw_bids(
         bids_path, suffix="channels", extension=".tsv", on_error="warn"
     )
     if channels_fname is not None:
-        raw = _handle_channels_reading(channels_fname, raw)
+        raw = _handle_channels_reading(
+            channels_fname, raw, ch_name_mismatch=ch_name_mismatch
+        )
 
     # Try to find an associated electrodes.tsv and coordsystem.json
     # to get information about the status and type of present channels

mne_bids/tests/test_read.py

@@ -33,6 +33,7 @@
 )
 from mne_bids.path import _find_matching_sidecar
 from mne_bids.read import (
+    _handle_channels_reading,
     _handle_events_reading,
     _handle_scans_reading,
     _read_raw,
@@ -1463,7 +1464,13 @@ def test_channels_tsv_raw_mismatch(tmp_path):
     raw.reorder_channels(ch_names_new)
     raw.save(raw_path, overwrite=True)
 
-    raw = read_raw_bids(bids_path)
+    with (
+        pytest.warns(
+            RuntimeWarning,
+            match=r"Channel mismatch between .*channels\.tsv and the raw data file detected\. Using mismatch strategy: reorder\.",
+        ),
+    ):
+        raw = read_raw_bids(bids_path, ch_name_mismatch="reorder")
     assert raw.ch_names == ch_names_orig
 
 
@@ -1509,6 +1516,109 @@ def test_gsr_and_temp_reading():
     assert raw.get_channel_types(["Temperature"]) == ["temperature"]
 
 
+def _setup_nirs_channel_mismatch(tmp_path):
+    ch_order_snirf = ["S1_D1 760", "S1_D2 760", "S1_D1 850", "S1_D2 850"]
+    ch_types = ["fnirs_cw_amplitude"] * len(ch_order_snirf)
+    info = mne.create_info(ch_order_snirf, sfreq=10, ch_types=ch_types)
+    data = np.arange(len(ch_order_snirf) * 10.0).reshape(len(ch_order_snirf), 10)
+    raw = mne.io.RawArray(data, info)
+
+    for i, ch_name in enumerate(raw.ch_names):
+        loc = np.zeros(12)
+        if "S1" in ch_name:
+            loc[3:6] = np.array([0, 0, 0])
+        if "D1" in ch_name:
+            loc[6:9] = np.array([1, 0, 0])
+        elif "D2" in ch_name:
+            loc[6:9] = np.array([0, 1, 0])
+        loc[9] = int(ch_name.split(" ")[1])
+        loc[0:3] = (loc[3:6] + loc[6:9]) / 2
+        raw.info["chs"][i]["loc"] = loc
+
+    orig_name_to_loc = {
+        name: raw.info["chs"][i]["loc"].copy() for i, name in enumerate(raw.ch_names)
+    }
+    orig_name_to_data = {
+        name: raw.get_data(picks=i).copy() for i, name in enumerate(raw.ch_names)
+    }
+
+    ch_order_bids = ["S1_D1 760", "S1_D1 850", "S1_D2 760", "S1_D2 850"]
+    ch_types_bids = ["NIRSCWAMPLITUDE"] * len(ch_order_bids)
+    channels_dict = OrderedDict([("name", ch_order_bids), ("type", ch_types_bids)])
+    channels_fname = tmp_path / "channels.tsv"
+    _to_tsv(channels_dict, channels_fname)
+
+    return (
+        raw,
+        ch_order_snirf,
+        ch_order_bids,
+        channels_fname,
+        orig_name_to_loc,
+        orig_name_to_data,
+    )
+
+
+def test_channel_mismatch_raise(tmp_path):
+    raw, _, _, channels_fname, _, _ = _setup_nirs_channel_mismatch(tmp_path)
+    with pytest.raises(
+        RuntimeError,
+        match=r"Channel mismatch between .*channels\.tsv and the raw data file detected\.",
+    ):
+        _handle_channels_reading(channels_fname, raw.copy(), ch_name_mismatch="raise")
+
+
+def test_channel_mismatch_reorder(tmp_path):
+    raw, _, ch_order_bids, channels_fname, orig_name_to_loc, orig_name_to_data = (
+        _setup_nirs_channel_mismatch(tmp_path)
+    )
+    with (
+        pytest.warns(
+            RuntimeWarning,
+            match=r"Channel mismatch between .*channels\.tsv and the raw data file detected\. Using mismatch strategy: reorder\.",
+        ),
+    ):
+        raw_out = _handle_channels_reading(
+            channels_fname, raw, ch_name_mismatch="reorder"
+        )
+    assert raw_out.ch_names == ch_order_bids
+    for i, new_name in enumerate(raw_out.ch_names):
+        np.testing.assert_allclose(
+            raw_out.info["chs"][i]["loc"], orig_name_to_loc[new_name]
+        )
+        np.testing.assert_allclose(
+            raw_out.get_data(picks=i), orig_name_to_data[new_name]
+        )
+
+
+def test_channel_mismatch_rename(tmp_path):
+    (
+        raw,
+        ch_order_snirf,
+        ch_order_bids,
+        channels_fname,
+        orig_name_to_loc,
+        orig_name_to_data,
+    ) = _setup_nirs_channel_mismatch(tmp_path)
+    with (
+        pytest.warns(
+            RuntimeWarning,
+            match=r"Channel mismatch between .*channels\.tsv and the raw data file detected\. Using mismatch strategy: rename\.",
+        ),
+    ):
+        raw_out_rename = _handle_channels_reading(
+            channels_fname, raw.copy(), ch_name_mismatch="rename"
+        )
+    assert raw_out_rename.ch_names == ch_order_bids
+    for i in range(len(ch_order_bids)):
+        orig_name_at_i = ch_order_snirf[i]
+        np.testing.assert_allclose(
+            raw_out_rename.info["chs"][i]["loc"], orig_name_to_loc[orig_name_at_i]
+        )
+        np.testing.assert_allclose(
+            raw_out_rename.get_data(picks=i), orig_name_to_data[orig_name_at_i]
+        )
+
+
 def test_events_file_to_annotation_kwargs(tmp_path):
     """Test that events file is read correctly."""
     bids_path = BIDSPath(

mne_bids/tests/test_write.py

@@ -96,6 +96,7 @@
     no_hand=r"ignore:Unable to map.*\n.*subject handedness.*:RuntimeWarning:mne",
     no_montage=r"ignore:Not setting position of.*channel found in "
     r"montage.*:RuntimeWarning:mne",
+    channel_mismatch="ignore:Channel mismatch between .*channels\\.tsv and the raw data file detected\\.:RuntimeWarning:mne",
 )
 
 
@@ -1398,6 +1399,7 @@ def test_vhdr(_bids_validate, tmp_path):
     warning_str["cnt_warning2"],
     warning_str["no_hand"],
     warning_str["no_montage"],
+    warning_str["channel_mismatch"],
 )
 @testing.requires_testing_data
 def test_eegieeg(dir_name, fname, reader, _bids_validate, tmp_path):
@@ -1488,7 +1490,7 @@ def test_eegieeg(dir_name, fname, reader, _bids_validate, tmp_path):
     # Reading the file back should still work, even though we've renamed
     # some channels (there's now a mismatch between BIDS and Raw channel
     # names, and BIDS should take precedence)
-    raw_read = read_raw_bids(bids_path=bids_path)
+    raw_read = read_raw_bids(bids_path=bids_path, ch_name_mismatch="rename")
     assert raw_read.ch_names[0] == "EOGtest"
     assert raw_read.ch_names[1] == "EMG"