NIFD Skill (Dataset-Orchestration Layer)

Overview

nifd-skill is the NeuroClaw orchestration skill for the Neuroimaging in Frontotemporal Dementia (NIFD) dataset, collected at the UCSF Memory and Aging Center.

It strictly follows the NeuroClaw hierarchical design principles:

• This skill only describes WHAT needs to be done and which tool skill to delegate to.
• It contains no implementation code or concrete commands.
• All concrete execution is delegated to existing base/tool skills via claw-shell.
• Companion scripts in scripts/ provide reference implementations for BIDS validation, phenotype extraction, and QC.

Core workflow (never bypassed):

1. Identify input NIFD data and target modalities.
2. Generate a numbered execution plan clearly stating WHAT needs to be done and which tool skill will handle each step.
3. Present the full plan, estimated runtime, resource requirements, and risks to the user and wait for explicit confirmation ("YES" / "execute" / "proceed").
4. On confirmation, delegate every step to the appropriate skill via claw-shell.
5. After execution, save all outputs in a clean directory structure (nifd_output/).

Research use only.

---

Quick Reference

Task	What needs to be done	Delegate to	Expected output
Data access	Apply for NIFD/FTLDNI access and accept the DUA in LONI IDA	Manual access gate	Approved IDA project access
BIDS validation	Validate NIFD BIDS structure	scripts/validate_nifd.py	Validation report
sMRI processing	Brain extraction, tissue segmentation, cortical thickness	smri-skill	smri_output/ derivatives
rs-fMRI processing	Preprocessing, denoising, connectivity	fmri-skill	fmri_output/ connectivity
dMRI processing	Diffusion preprocessing, tensor metrics, tractography	dwi-skill	dwi_output/ metrics
Phenotype extraction	Diagnosis, cognitive scores, clinical measures	scripts/extract_nifd_phenotype.py	Merged phenotype CSV
QC summary	Per-subject quality control	scripts/nifd_qc_summary.py	QC summary + exclusion list

---

Access Stage

• Official route: https://ida.loni.usc.edu/collaboration/access/appApply.jsp?project=NIFD
• NIFD/FTLDNI is a controlled LONI IDA project. Create an IDA account, submit the requested use information, and accept the project DUA before downloading.
• The previously cited OpenNeuro dataset ds004403 has been deleted as a duplicate. Do not use it as a download source or claim that NIFD is anonymously downloadable.
• After approval, record the IDA project, export date, query, subject/session selection, and checksum manifest before staging.
• Treat the IDA export as source data. Convert or organize it into BIDS before running the validation and modality workflows below.

---

Dataset Characteristics

• Cohort: ~120 participants
  • bvFTD: Behavioral variant frontotemporal dementia
  • svPPA: Semantic variant primary progressive aphasia
  • nfvPPA: Nonfluent variant primary progressive aphasia
  • Healthy controls: Age-matched
• Scanner: 3T Siemens TIM Trio
• Modalities: T1w sMRI, rs-fMRI, dMRI/DTI
• Clinical: CDR, MMSE, neuropsychological battery
• Access: Controlled download through the NIFD/FTLDNI project in LONI IDA
• Format: IDA export; BIDS staging and validation are required before BIDS-native tools

---

Supported Modalities

Modality	Description	Details
T1w	High-resolution structural MRI	1mm isotropic, cortical thickness/atrophy
rs-fMRI	Resting-state functional MRI	Functional connectivity, network degeneration
dMRI	Diffusion-weighted imaging	DTI, white matter tract integrity

---

NIFD Diagnostic Groups

Group	Description	Typical N
bvFTD	Behavioral variant FTD	~40
svPPA	Semantic variant PPA	~20
nfvPPA	Nonfluent variant PPA	~15
Control	Healthy age-matched controls	~45

---

BIDS Preparation

Script: scripts/validate_nifd.py

Validates NIFD BIDS structure and generates a compliance report.

python skills/nifd-skill/scripts/validate_nifd.py \
  --input /path/to/NIFD/bids \
  --output /path/to/nifd_output/qc/bids_validation.csv

Features:

• BIDS directory structure validation
• Diagnostic group completeness check
• Modality completeness (T1w, rs-fMRI, dMRI)
• Missing data identification

---

Core Workflow (Never Bypassed)

1. Identify user target: full NIFD processing, imaging subset, phenotype extraction, or BIDS validation only.
2. Generate a numbered plan with tools, outputs, runtime, storage, and risks.
3. Wait for explicit confirmation (YES / execute / proceed).
4. On confirmation, run BIDS validation using scripts/validate_nifd.py.
5. Delegate to smri-skill for structural MRI processing.
6. Delegate to fmri-skill for rs-fMRI processing (functional connectivity).
7. Delegate to dwi-skill for dMRI processing (white matter integrity).
8. If phenotype extraction is requested, run scripts/extract_nifd_phenotype.py.
9. If QC summary is requested, run scripts/nifd_qc_summary.py.
10. Save outputs into nifd_output/.

---

Modality Processing Delegation

Modality	Delegated skill	Typical tasks	Main outputs
sMRI (T1w)	smri-skill	brain extraction, tissue segmentation, cortical thickness	smri_output/ derivatives
rs-fMRI	fmri-skill	preprocessing, denoising, connectivity	fmri_output/ connectivity
dMRI	dwi-skill	diffusion preprocessing, tensor metrics, tractography	dwi_output/ metrics

---

Standard Output Layout

nifd_output/
├── bids/                   # BIDS-staged data (or validation report)
├── smri/                   # Structural MRI derivatives
├── fmri/                   # Functional MRI derivatives (rs-fMRI connectivity)
├── dwi/                    # Diffusion MRI derivatives (DTI metrics)
├── phenotype/              # Merged phenotype tables (diagnosis, cognitive)
├── qc/                     # QC summaries and exclusion lists
└── logs/                   # Processing logs

---

Benchmark Adapter Guidance

For benchmark-style prompts, do not force the full orchestration when the task only asks for local NIFD data validation.

• If the task starts from NIFD data already present on disk and only asks for BIDS validation:
  • Skip the download stage
  • Default to the narrow path local NIFD discovery -> BIDS validation -> report
• In benchmark mode, do not require explicit confirmation before presenting the validation solution.

---

Safety and Execution Policy

• No execution before explicit plan confirmation.
• All execution must be routed via claw-shell.
• Missing dependencies must be resolved by dependency-planner before running.

---

Important Notes and Limitations

• NIFD is a clinical cohort; patient data requires careful handling.
• NIFD data may not be redistributed; preserve the DUA, required acknowledgement, release/export date, and local access controls.
• Diagnostic groups (bvFTD, svPPA, nfvPPA) have distinct atrophy patterns; group-level analyses should account for heterogeneity.
• Cortical thickness and voxel-based morphometry are commonly used structural measures.
• Network degeneration hypothesis: FTD targets specific large-scale networks.
• nifd-skill is orchestration-only; detailed preprocessing logic remains in modality skills.

---

When to Call This Skill

• User asks for end-to-end NIFD workflow.
• User asks to process NIFD neuroimaging data.
• User needs BIDS validation for NIFD data.
• User asks to extract NIFD phenotype data (diagnosis, cognitive scores).
• User asks for frontotemporal dementia neuroimaging analysis.

---

Complementary / Related Skills

• smri-skill → structural MRI preprocessing
• fmri-skill → functional MRI preprocessing and analysis
• dwi-skill → diffusion MRI preprocessing
• pet-skill → PET imaging (tau-PET, amyloid-PET if available)
• bids-organizer → BIDS validation and organization
• brain-visualization → visualization of derivatives
• dependency-planner → dependency resolution
• conda-env-manager → environment management
• claw-shell → command execution

---

Reference

• NIFD/FTLDNI LONI DUA: https://ida.loni.usc.edu/collaboration/access/appApply.jsp?project=NIFD
• NIFD: UCSF Memory and Aging Center
• Frontotemporal Dementia: FTDC clinical diagnostic criteria

Created At: 2026-05-06 13:55 HKT Last Updated At: 2026-08-11 HKT Author: chengwang96