diff --git a/freesurfer_post/interfaces/tabular.py b/freesurfer_post/interfaces/tabular.py
index 8c754a2..4875e1e 100644
--- a/freesurfer_post/interfaces/tabular.py
+++ b/freesurfer_post/interfaces/tabular.py
@@ -137,11 +137,42 @@ def sanity_check_columns(reference_column, redundant_column, atol=0):
         output_dir = Path(self.inputs.output_dir) / subject_id
         output_dir.mkdir(parents=True, exist_ok=True)
         output_prefix = f'{subject_id}_{session_id}' if session_id else subject_id
-        cleaned_atlas_name = (
-            atlas.replace('.', '').replace('_order', '').replace('_', '')
-        )
+
+        cleaned_atlas_name = atlas.replace('.', '').replace('_order', '').replace('_', '')
+
+        # Convert column names to snake case
+        out_df.columns = [col.lower().replace('-', '_').replace('.', '_') for col in out_df.columns]
+        # Rename subject_id to participant_id
+        out_df = out_df.rename(columns={'subject_id': 'participant_id'})
+        # Reorder columns to have participant_id first
+        cols = out_df.columns.tolist()
+        cols.remove('participant_id')
+        out_df = out_df[['participant_id'] + cols]
+        # Replace missing values with "n/a"
+        out_df = out_df.fillna('n/a')
+
+        # Create metadata with the same column names
+        metadata = {}
+        for col in out_df.columns:
+            if col == 'participant_id':
+                metadata[col] = {'Description': 'BIDS participant ID'}
+            elif col == 'session_id':
+                metadata[col] = {'Description': 'BIDS session ID'}
+            elif col == 'hemisphere':
+                metadata[col] = {'Description': 'Brain hemisphere (lh or rh)'}
+            elif col == 'atlas':
+                metadata[col] = {'Description': 'Atlas used for parcellation'}
+            else:
+                metadata[col] = {'Description': f'Surface statistic: {col}'}
+
+        # Save metadata
+        out_json = output_dir / f'{output_prefix}_seg-{cleaned_atlas_name}_surfacestats.json'
+        with out_json.open('w') as jsonf:
+            json.dump(metadata, jsonf, indent=2)
+
+        # Save data
         out_df.to_csv(
-            output_dir / f'{output_prefix}_atlas-{cleaned_atlas_name}_surfacestats.tsv',
+            output_dir / f'{output_prefix}_seg-{cleaned_atlas_name}_surfacestats.tsv',
             sep='\t',
             index=False,
         )
@@ -361,11 +392,31 @@ def _run_interface(self, runtime):
         out_tsv = output_dir / f'{output_prefix}_brainmeasures.tsv'
         out_json = output_dir / f'{output_prefix}_brainmeasures.json'
 
-        metadata = {key: value['meta'] for key, value in fs_audit.items()}
+        # Extract just the values from the audit data
+        data_value = {key: value['value'] for key, value in fs_audit.items()}
+        data_df = pd.DataFrame([data_value])
+        # Convert column names to snake case
+        data_df.columns = [col.lower().replace('-', '_').replace('.', '_') for col in data_df.columns]
+        # Rename subject_id to participant_id
+        data_df = data_df.rename(columns={'subject_id': 'participant_id'})
+        # Reorder columns to have participant_id first
+        cols = data_df.columns.tolist()
+        cols.remove('participant_id')
+        data_df = data_df[['participant_id'] + cols]
+        # Replace missing values with "n/a"
+        data_df = data_df.fillna('n/a')
+
+        # Create metadata with the same column names as the TSV
+        metadata = {}
+        for key, value in fs_audit.items():
+            # Convert the key to match TSV column name
+            new_key = key.lower().replace('-', '_').replace('.', '_')
+            if key == 'subject_id':
+                new_key = 'participant_id'
+            metadata[new_key] = {'Description': value['meta']}
+
         with out_json.open('w') as jsonf:
             json.dump(metadata, jsonf, indent=2)
 
-        real_data = {key: value['value'] for key, value in fs_audit.items()}
-        data_df = pd.DataFrame([real_data])
         data_df.to_csv(out_tsv, sep='\t', index=False)
         return runtime
diff --git a/freesurfer_post/utils.py b/freesurfer_post/utils.py
index 4dd6905..0ffa368 100644
--- a/freesurfer_post/utils.py
+++ b/freesurfer_post/utils.py
@@ -38,7 +38,7 @@ def find_freesurfer_dir(
         if warn_no_session:
             warnings.warn(
                 f'{subjects_dir}/{subject_id}_{session_id} not found in {subjects_dir}'
-                f'using {subjects_dir}/{subject_id} instead',
+                f' using {subjects_dir}/{subject_id} instead',
                 stacklevel=2,
             )
         return subjects_dir / subject_id