Review Papers Claim by Claim

 Quick Explanation Copied

 Long Explanation

What the Paper Reports

Surf_2_Volume chains Connectome Workbench, FreeSurfer, AFNI/SUMA, neuromaps, and Python to convert CIFTI parcellations into NIfTI volumes, separating cortical and subcortical components, transferring cortical labels via fsaverage to a surface representation of the MNI152 target, and assigning voxels within a user-adjustable gray-matter probability mask Surf_2_Volume preprint. The core quantitative results are adjusted Dice scores for two atlas tests:

The adjusted Dice combines surface label overlap (D) with two voxel terms: E (proportion of labeled voxels outside gray matter) and O (proportion of gray-matter voxels labeled). Threshold comparisons (>0.25 vs >0.01) were qualitative only Surf_2_Volume results.

Critical Assessment

• Strengths: Honest, well-scoped limitations section; open code, Docker container, and precomputed CAB-NP outputs in two MNI152 spaces; demonstrated file compatibility in SPM, FSL, MRIcroGL, and freeview. The gap addressed is real—CIFTI atlases are not directly usable in volume-based pipelines Data availability and compatibility.
• Weaknesses: Descriptive-only evaluation with no statistics, confidence intervals, or participant-level variation; only two atlases and a restricted set of Workbench settings; the adjusted Dice metric itself is unvalidated. The comparison may also be somewhat circular: the Schaefer reference was itself projected to the surface and back, so reconstruction error conflates method error with round-trip error.
• Blind spot: The choice of Workbench settings (ribbon, nearest-vertex 2 and 7) is presented as illustrative; an exhaustive parameter sweep could narrow the gap. No per-label error or boundary-distance analysis was done.

Verification Path

The claim would weaken if independent evaluation on additional CIFTI parcellations found Surf_2_Volume no better than tuned Workbench settings—the authors' own falsification criterion. Confidence in the reported scores is moderate (single-run descriptive values); confidence in generalizability across atlases and analyses is low until replicated.

Deeper context: the volume-vs-surface localization problem the paper acknowledges is well documented—volumetric registration can mix signal across sulcal banks Coalson et al. surface vs volume localization, and the CIFTI grayordinate framework underlying CAB-NP was designed to avoid it CAB-NP parcellation paper.

 BGPT Paper Review

Study Novelty

Pipelines for surface-to-volume atlas conversion exist (e.g., HCP-MMP projections, extended MNI152 atlases); the contribution is a generalizable, threshold-controlled workflow rather than a new method concept.

Scientific Quality

Transparent methods, honest limitations, open data and container. But evaluation is purely descriptive: no statistics, only two atlases, limited comparator settings, unvalidated composite metric, and a partially circular Schaefer round-trip test.

Study Generality

The workflow is designed for any CIFTI parcellation and two standard MNI152 spaces, but demonstrated generalization is limited to two group-level atlases.

Study Usefulness

Solves a concrete, recurring practical gap: using CIFTI parcellations in NIfTI-only pipelines (SPM, FSL) with explicit gray-matter control, plus precomputed outputs and a Docker image.

Study Reproducibility

GitHub code, Docker container, precomputed outputs, and detailed tool-by-tool sequence support reproduction, but no statistical or participant-level protocol to verify reported scores.

Explanatory Depth

The paper explains the pipeline mechanics and metric construction but offers no theoretical or mechanistic analysis of why mappings fail or how errors propagate through registration steps.