Fix longitudinal functional template

src/rbc/bids/longitudinal/anatomical.py

@@ -43,7 +43,7 @@ def resolve_longitudinal_anat(
         Dict with keys matching ``longitudinal_process`` parameters.
     """
     return {
-        "template": tpl_q.expect(tpl_df, suffix=Suffix.T1W),
+        "template": tpl_q.expect(tpl_df, suffix=Suffix.T1W, without=["res"]),
         "subj_to_template_xfm": tpl_q.expect(
             tpl_df,
             suffix="xfm",

src/rbc/bids/longitudinal/functional.py

@@ -23,6 +23,7 @@ def resolve_longitudinal_func(
     tpl_df: pl.DataFrame,
     *,
     ses: str,
+    task: str,
     regressors: Sequence[str] = ("36-parameter",),
 ) -> dict[str, Path | dict[str, Path]]:
     """Resolve inputs for longitudinal functional processing.
@@ -33,6 +34,7 @@ def resolve_longitudinal_func(
         func_df: DataFrame of functional derivatives.
         tpl_df: DataFrame of longitudinal template files.
         ses: Session label (used for template xfm lookup).
+        task: Task entity value to denote BOLD reference for template resampling.
         regressors: Regressor strategy names to resolve raw regressor
             files for.
 
@@ -51,7 +53,7 @@ def resolve_longitudinal_func(
         )
 
     return {
-        "template": tpl_q.expect(tpl_df, suffix="T1w"),
+        "template": tpl_q.expect(tpl_df, suffix="T1w", res=task),
         "anat_to_template_xfm": tpl_q.expect(
             tpl_df,
             suffix="xfm",

src/rbc/bids/longitudinal/template.py

@@ -7,7 +7,7 @@
 
 import polars as pl
 
-from rbc.bids import Suffix, bids_safe_label
+from rbc.bids import FUNC_GROUP_ENTITIES, Suffix, bids_safe_label
 
 if TYPE_CHECKING:
     from rbc.bids import Bids
@@ -21,11 +21,13 @@ class TemplateInputs(NamedTuple):
         sub: Subject label.
         sessions: Per-input session labels (parallel to ``files``).
         files: Per-session preprocessed T1w brain volumes.
+        bold_ref: Per-session preprocessed BOLD volumes (all tasks).
     """
 
     sub: str
     sessions: list[str]
     files: list[Path]
+    bold_files: dict[str, Path]
 
 
 def discover_template_inputs(
@@ -56,6 +58,13 @@ def discover_template_inputs(
         # the mri_robust_template invocation.
         pl.col("space").is_null(),
     )
+    bold_rows = df.filter(
+        pl.col("ses") != "longitudinal",
+        pl.col("datatype") == "func",
+        pl.col("desc") == "preproc",
+        pl.col("suffix") == "bold",
+        pl.col("space").is_null(),
+    )
 
     inputs: list[TemplateInputs] = []
     skipped: list[str] = []
@@ -68,7 +77,29 @@ def discover_template_inputs(
         files = [
             Path(row["root"]) / row["path"] for row in sub_group.iter_rows(named=True)
         ]
-        inputs.append(TemplateInputs(sub=sub, sessions=sessions, files=files))
+        sub_bold = bold_rows.filter(
+            (pl.col("sub") == sub) & (pl.col("ses") == sessions[0])
+        ).unique(subset=(*FUNC_GROUP_ENTITIES, "root", "path"))
+        # Check each task is unique, otherwise raise assertion error with details
+        if sub_bold.height != sub_bold.unique().height:
+            conflicts = (
+                sub_bold.filter(pl.struct(FUNC_GROUP_ENTITIES).is_duplicated())
+                .group_by(FUNC_GROUP_ENTITIES)
+                .agg(pl.format("{}/{}", "root", "path").alias("paths"))
+            )
+            raise AssertionError(
+                f"Found multiple non-matching grids for subject {sub}:\n"
+                + "\n".join(str(dict(row)) for row in conflicts.iter_rows(named=True))
+            )
+        bold_files = {
+            row["task"]: Path(row["root"]) / row["path"]
+            for row in sub_bold.iter_rows(named=True)
+        }
+        inputs.append(
+            TemplateInputs(
+                sub=sub, sessions=sessions, files=files, bold_files=bold_files
+            )
+        )
     return inputs, skipped
 
 
@@ -81,6 +112,8 @@ def export_template(tpl: Bids, outputs: LongitudinalTemplateOutputs) -> None:
         outputs: Results from the longitudinal template workflow.
     """
     tpl.save(outputs.template, suffix=Suffix.T1W)
+    for btask, bold_template in outputs.bold_templates.items():
+        tpl.save(bold_template, res=btask, suffix=Suffix.T1W)
     for ses, xfm in zip(outputs.sessions, outputs.transforms, strict=True):
         ses_label = bids_safe_label(ses)
         tpl.save(

src/rbc/core/longitudinal/resampling.py

@@ -0,0 +1,42 @@
+"""Resampling utilities for longitudinal templates."""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+import nibabel as nib
+from nibabel.processing import resample_from_to
+
+if TYPE_CHECKING:
+    from pathlib import Path
+
+from rbc.core.niwrap import generate_exec_folder
+
+
+def resample_img_to_bold_grid(bold_ref: Path, img: Path, order: int = 3) -> Path:
+    """Resample template to BOLD grid if shapes differ.
+
+    Args:
+        bold_ref: BOLD reference volume.
+        img: 3D image in target space to resample.
+        order: Interpolation order used during resampling
+
+    Returns:
+        Resampled 3D image with BOLD grid
+    """
+    bold_ref_img = nib.nifti1.load(bold_ref)
+    img_obj = nib.nifti1.load(img)
+
+    # If 4D, extract first volume
+    if len(bold_ref_img.shape) > 3:
+        bold_ref_img = bold_ref_img.slicer[..., 0]
+    # If same shape, no need to resample
+    if bold_ref_img.shape == img_obj.shape:
+        return img
+
+    img_resampled = resample_from_to(img_obj, bold_ref_img, order=order)
+    img_resampled_path = (
+        generate_exec_folder("img_resample_to_bold_grid") / "resampled.nii.gz"
+    )
+    nib.save(img_resampled, img_resampled_path)
+    return img_resampled_path

src/rbc/orchestration/longitudinal/all.py

@@ -18,7 +18,7 @@
 from rbc.bids import FUNC_GROUP_ENTITIES, Datatype, Suffix, extract_entities, load_table
 from rbc.bids.longitudinal.template import discover_template_inputs
 from rbc.bids.metrics import export_metrics
-from rbc.bids.session import iter_session_files
+from rbc.bids.session import _FUNC_ENTITY_KEYS, iter_session_files
 from rbc.context import RunContext
 from rbc.orchestration import Filters, RunnerConfig, init_runner
 from rbc.orchestration.longitudinal._iter import iter_sessions_with_template
@@ -124,7 +124,7 @@ def run(
             )
 
             row = func_df.filter(suffix=Suffix.BOLD).row(0, named=True)
-            ents = extract_entities(row, ["task", "run"])
+            ents = extract_entities(row, _FUNC_ENTITY_KEYS)
             func_q = pipe_ctx.bids(datatype=Datatype.FUNC, entities=ents)
             func_long = func_q.derive(space="longitudinal")
 

src/rbc/orchestration/longitudinal/anatomical.py

@@ -51,7 +51,7 @@ def process_anat(
         Workflow outputs for in-memory handoff to downstream stages.
     """
     anat_df = anat_df.filter(pl.col("space").is_null())
-    ents = extract_entities(anat_df.row(0, named=True), ["run"])
+    ents = extract_entities(anat_df.row(0, named=True), ["run", "acq", "rec", "echo"])
 
     anat_q = pipe_ctx.bids(datatype=Datatype.ANAT)
     tpl_q = anat_q.derive(ses="longitudinal")

src/rbc/orchestration/longitudinal/functional.py

@@ -10,7 +10,7 @@
     export_longitudinal_func,
     resolve_longitudinal_func,
 )
-from rbc.bids.session import iter_session_files
+from rbc.bids.session import _FUNC_ENTITY_KEYS, iter_session_files
 from rbc.orchestration import Filters, RunnerConfig, init_runner
 from rbc.orchestration.longitudinal._iter import iter_sessions_with_template
 from rbc.workflows.longitudinal.functional import (
@@ -53,7 +53,7 @@ def process_func(
         Workflow outputs for in-memory handoff to downstream stages.
     """
     row = func_df.filter(suffix=Suffix.BOLD).row(0, named=True)
-    ents = extract_entities(row, ["task", "run"])
+    ents = extract_entities(row, list(_FUNC_ENTITY_KEYS))
 
     func_q = pipe_ctx.bids(datatype=Datatype.FUNC, entities=ents)
     tpl_q = pipe_ctx.bids(datatype=Datatype.ANAT).derive(ses="longitudinal")
@@ -64,6 +64,7 @@ def process_func(
         func_df,
         tpl_df,
         ses=pipe_ctx.ses,  # type: ignore[arg-type]
+        task=ents["task"],
         regressors=regressors,
     )
     func_outputs = functional_longitudinal(**resolved)  # type: ignore[arg-type]

src/rbc/orchestration/longitudinal/qc.py

@@ -21,6 +21,7 @@
     write_longitudinal_qc_tsv,
 )
 from rbc.bids.session import iter_session_files
+from rbc.core.longitudinal.resampling import resample_img_to_bold_grid
 from rbc.core.niwrap import generate_exec_folder
 from rbc.core.qc.registration import registration_qc_metrics
 from rbc.orchestration import Filters, RunnerConfig, init_runner
@@ -67,6 +68,10 @@ def process_qc(
     Returns:
         QC outputs with overlap metrics and pass/fail flag.
     """
+    # Resample longitudinal anatomical mask to bold grid for QC purposes.
+    # Longitudinal processed data are registered to the longitudinal template with
+    # respective modality's native resolution
+    anat_brain_mask = resample_img_to_bold_grid(bold_mask, anat_brain_mask, order=0)
     anat_mask_arr = nib.nifti1.load(anat_brain_mask).get_fdata()
     bold_mask_arr = nib.nifti1.load(bold_mask).get_fdata()
     reg_metrics = registration_qc_metrics(anat_mask_arr, bold_mask_arr)

src/rbc/orchestration/longitudinal/template.py

@@ -49,6 +49,7 @@ def process_subject(
         sub=inputs.sub,
         sessions=inputs.sessions,
         in_files=inputs.files,
+        bold_files=inputs.bold_files,
     )
     tpl = pipe_ctx.bids(datatype=Datatype.ANAT).derive(ses="longitudinal")
     export_template(tpl, outputs)

src/rbc/workflows/longitudinal/template.py

@@ -14,9 +14,10 @@
     fs_to_itk_xfm,
     generate_robust_template,
 )
+from rbc.core.longitudinal.resampling import resample_img_to_bold_grid
 
 if TYPE_CHECKING:
-    from collections.abc import Sequence
+    from collections.abc import Mapping, Sequence
     from pathlib import Path
 
 _logger = logging.getLogger("rbc")
@@ -27,11 +28,14 @@ class LongitudinalTemplateOutputs(NamedTuple):
 
     Attributes:
         template: Robust within-subject template volume.
+        bold_templates: Within-subject template volumes resampled to task-specific BOLD
+            resolutions.
         sessions: Session labels in the same order as ``transforms``.
         transforms: Per-session ITK-format session-to-template transforms.
     """
 
     template: Path
+    bold_templates: dict[str, Path]
     sessions: list[str]
     transforms: list[Path]
 
@@ -40,13 +44,15 @@ def generate_subject_template(
     sub: str,
     sessions: Sequence[str],
     in_files: Sequence[Path],
+    bold_files: Mapping[str, Path],
 ) -> LongitudinalTemplateOutputs:
     """Build a robust template and ITK transforms for one subject.
 
     Args:
         sub: Subject label (without the ``sub-`` prefix).
         sessions: Session labels parallel to ``in_files``.
         in_files: Per-session preprocessed T1w volumes (e.g. brain-extracted).
+        bold_files: Reference bold volumes to resample template for functional data.
 
     Returns:
         :class:`LongitudinalTemplateOutputs` ready for BIDS export.
@@ -65,8 +71,15 @@ def generate_subject_template(
         in_xfms=robust.transforms,
     )
 
+    _logger.info("Creating reference volumes for each functional task")
+    bold_templates = {
+        btask: resample_img_to_bold_grid(bfile, robust.template)
+        for btask, bfile in bold_files.items()
+    }
+
     return LongitudinalTemplateOutputs(
         template=robust.template,
+        bold_templates=bold_templates,
         sessions=list(sessions),
         transforms=itk_xfms,
     )

tests/unit/bids/test_longitudinal_anatomical.py

@@ -99,6 +99,7 @@ def _anat_row(
         "sub": sub,
         "ses": ses,
         "desc": desc,
+        "res": None,
         "root": "/data",
         "path": path,
         "extra_entities": extra or [],