BUG: Fix bug with ch_name resolution (mne-tools#12086)

Author: larsoner

mne/channels/channels.py

@@ -1556,7 +1556,10 @@ def find_ch_adjacency(info, ch_type):
 
     if conn_name is not None:
         logger.info(f"Reading adjacency matrix for {conn_name}.")
-        return read_ch_adjacency(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."

mne/channels/tests/test_channels.py

@@ -465,14 +465,21 @@ def test_1020_selection():
 @testing.requires_testing_data
 def test_find_ch_adjacency():
     """Test computing the adjacency matrix."""
-    raw = read_raw_fif(raw_fname, preload=True)
+    raw = read_raw_fif(raw_fname)
     sizes = {"mag": 828, "grad": 1700, "eeg": 384}
     nchans = {"mag": 102, "grad": 204, "eeg": 60}
     for ch_type in ["mag", "grad", "eeg"]:
         conn, ch_names = find_ch_adjacency(raw.info, ch_type)
         # Silly test for checking the number of neighbors.
         assert_equal(conn.getnnz(), sizes[ch_type])
         assert_equal(len(ch_names), nchans[ch_type])
+        kwargs = dict(exclude=())
+        if ch_type in ("mag", "grad"):
+            kwargs["meg"] = ch_type
+        else:
+            kwargs[ch_type] = True
+        want_names = [raw.ch_names[pick] for pick in pick_types(raw.info, **kwargs)]
+        assert ch_names == want_names
     pytest.raises(ValueError, find_ch_adjacency, raw.info, None)
 
     # Test computing the conn matrix with gradiometers.
@@ -506,7 +513,7 @@ def test_find_ch_adjacency():
 def test_neuromag122_adjacency():
     """Test computing the adjacency matrix of Neuromag122-Data."""
     nm122_fname = testing_path / "misc" / "neuromag122_test_file-raw.fif"
-    raw = read_raw_fif(nm122_fname, preload=True)
+    raw = read_raw_fif(nm122_fname)
     conn, ch_names = find_ch_adjacency(raw.info, "grad")
     assert conn.getnnz() == 1564
     assert len(ch_names) == 122
@@ -515,7 +522,7 @@ def test_neuromag122_adjacency():
 
 def test_drop_channels():
     """Test if dropping channels works with various arguments."""
-    raw = read_raw_fif(raw_fname, preload=True).crop(0, 0.1)
+    raw = read_raw_fif(raw_fname).crop(0, 0.1)
     raw.drop_channels(["MEG 0111"])  # list argument
     raw.drop_channels("MEG 0112")  # str argument
     raw.drop_channels({"MEG 0132", "MEG 0133"})  # set argument
@@ -535,7 +542,7 @@ def test_drop_channels():
 
 def test_pick_channels():
     """Test if picking channels works with various arguments."""
-    raw = read_raw_fif(raw_fname, preload=True).crop(0, 0.1)
+    raw = read_raw_fif(raw_fname).crop(0, 0.1)
 
     # selected correctly 3 channels
     raw.pick(["MEG 0113", "MEG 0112", "MEG 0111"])

tutorials/stats-sensor-space/40_cluster_1samp_time_freq.py

@@ -5,10 +5,9 @@
 Non-parametric 1 sample cluster statistic on single trial power
 ===============================================================
 
-This script shows how to estimate significant clusters
-in time-frequency power estimates. It uses a non-parametric
-statistical procedure based on permutations and cluster
-level statistics.
+This script shows how to estimate significant clusters in time-frequency power
+estimates. It uses a non-parametric statistical procedure based on permutations and
+cluster level statistics.
 
 The procedure consists of:
 

tutorials/stats-sensor-space/75_cluster_ftest_spatiotemporal.py

@@ -5,12 +5,10 @@
 Spatiotemporal permutation F-test on full sensor data
 =====================================================
 
-Tests for differential evoked responses in at least
-one condition using a permutation clustering test.
-The FieldTrip neighbor templates will be used to determine
-the adjacency between sensors. This serves as a spatial prior
-to the clustering. Spatiotemporal clusters will then
-be visualized using custom matplotlib code.
+Tests for differential evoked responses in at least one condition using a permutation
+clustering test. The FieldTrip neighbor templates will be used to determine the
+adjacency between sensors. This serves as a spatial prior to the clustering.
+Spatiotemporal clusters will then be visualized using custom matplotlib code.
 
 Here, the unit of observation is epochs from a specific study subject.
 However, the same logic applies when the unit observation is