diff --git a/src/bids_validator/context.py b/src/bids_validator/context.py
index 7b940a1..2427a59 100644
--- a/src/bids_validator/context.py
+++ b/src/bids_validator/context.py
@@ -219,8 +219,8 @@ def walk_back(
             yield file_group
         elif len(file_group) == 1:
             yield file_group[0]
-        else:
-            raise ValidationError('Multiple matching files.')
+        elif file_group:
+            raise ValidationError(f'Multiple matching files: {file_group}')
 
 
 def _walk_back(
@@ -230,7 +230,7 @@ def _walk_back(
     target_suffix: str | None,
     target_entities: tuple[str, ...],
 ) -> Generator[list[FileTree, ...]]:
-    file_parts = FileParts.from_file(source.relative_path)
+    file_parts = FileParts.from_file(source)
 
     if target_suffix is None:
         target_suffix = file_parts.suffix
@@ -238,11 +238,11 @@ def _walk_back(
     tree = source.parent
     while tree:
         matches = []
-        for child in tree.children:
+        for child in tree.children.values():
             if child.is_dir:
                 continue
-            parts = FileParts.from_file(child.relative_path)
-            if parts.extension != target_extensions:
+            parts = FileParts.from_file(child)
+            if parts.extension not in target_extensions:
                 continue
             if parts.suffix != target_suffix:
                 continue
@@ -270,7 +270,7 @@ class FileParts:
     extension: str | None
 
     @classmethod
-    def from_file(cls, file: FileTree, schema: Namespace) -> t.Self:
+    def from_file(cls, file: FileTree, schema: Namespace | None = None) -> t.Self:
         """Parse file parts from FileTree object."""
         stem, _, extension = file.name.partition('.')
 
@@ -280,7 +280,7 @@ def from_file(cls, file: FileTree, schema: Namespace) -> t.Self:
             extension = f'{extension}/'
 
         datatype = None
-        if file.parent:
+        if file.parent and schema:
             if any(file.parent.name == dtype.value for dtype in schema.objects.datatypes.values()):
                 datatype = file.parent.name
 
diff --git a/tests/test_context.py b/tests/test_context.py
index 0cc6c45..234f393 100644
--- a/tests/test_context.py
+++ b/tests/test_context.py
@@ -1,37 +1,35 @@
+import pytest
+
 from bids_validator import context
 from bids_validator.types.files import FileTree
 
-import pytest
 
-def test_load(examples, schema):
-    tree = FileTree.read_from_filesystem(examples / 'synthetic')
-    ds = context.Dataset(tree, schema)
+@pytest.fixture
+def synthetic_dataset(examples):
+    return FileTree.read_from_filesystem(examples / 'synthetic')
+
+
+def test_load(synthetic_dataset, schema):
+    ds = context.Dataset(synthetic_dataset, schema)
 
     assert ds.dataset_description.Name.startswith('Synthetic dataset')
     assert ds.subjects.participant_id == [f'sub-{i:02d}' for i in range(1, 6)]
     assert sorted(ds.subjects.sub_dirs) == [f'sub-{i:02d}' for i in range(1, 6)]
-    assert sorted(ds.datatypes) == ["anat", "beh", "func"]
-    assert sorted(ds.modalities) == ["beh", "mri"]
-
-
-@pytest.mark.parametrize(
-        "depth, expected",
-        [
-            (2, {"anat", "beh", "func"}),
-            (1, set())
-        ])
-def test_find_datatypes(examples, schema, depth, expected):
-    tree = FileTree.read_from_filesystem(examples / 'synthetic')
+    assert sorted(ds.datatypes) == ['anat', 'beh', 'func']
+    assert sorted(ds.modalities) == ['beh', 'mri']
+
+
+@pytest.mark.parametrize(('depth', 'expected'), [(2, {'anat', 'beh', 'func'}), (1, set())])
+def test_find_datatypes(synthetic_dataset, schema, depth, expected):
     datatypes = schema.objects.datatypes
 
-    result = context.find_datatypes(tree, datatypes, max_depth=depth)
+    result = context.find_datatypes(synthetic_dataset, datatypes, max_depth=depth)
 
     assert result == expected
 
-def test_fileparts(examples, schema):
-    tree = FileTree.read_from_filesystem(examples / 'synthetic')
 
-    T1w = tree / 'sub-01' / 'ses-01' / 'anat' / 'sub-01_ses-01_T1w.nii'
+def test_fileparts(synthetic_dataset, schema):
+    T1w = synthetic_dataset / 'sub-01' / 'ses-01' / 'anat' / 'sub-01_ses-01_T1w.nii'
     parts = context.FileParts.from_file(T1w, schema)
     assert parts == context.FileParts(
         path='/sub-01/ses-01/anat/sub-01_ses-01_T1w.nii',
@@ -41,3 +39,16 @@ def test_fileparts(examples, schema):
         suffix='T1w',
         extension='.nii',
     )
+
+
+def test_walkback(synthetic_dataset, schema):
+    bold = (
+        synthetic_dataset
+        / 'sub-01'
+        / 'ses-01'
+        / 'func'
+        / 'sub-01_ses-01_task-nback_run-01_bold.nii'
+    )
+    sidecars = list(context.walk_back(bold, inherit=True))
+    assert len(sidecars) == 1
+    assert sidecars[0] is synthetic_dataset / 'task-nback_bold.json'