Oasis Skill

OASIS Skill (Dataset-Orchestration Layer)

Overview

oasis-skill is the NeuroClaw orchestration skill for the OASIS (Open Access Series of Imaging Studies) dataset, founded by the Knight Alzheimer Disease Research Center at Washington University.

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 OASIS 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 (oasis_output/).

Research use only.

Quick Reference

Dataset Characteristics

• OASIS-1 (Cross-sectional): ~416 subjects aged 18-96
  • T1w MRI, some with very mild to mild AD
• OASIS-2 (Longitudinal): ~150 subjects aged 60-96
  • Multiple sessions, non-demented and demented
• OASIS-3: ~1,000+ subjects, longitudinal spanning ~20 years
  • T1w, PET (amyloid, tau), ASL
• Scanner: 1.5T Siemens Vision
• Clinical: CDR (Clinical Dementia Rating), MMSE, demographics
• Access: OpenNeuro ds000014 (OASIS-1), www.oasis-brains.org (OASIS-3)
• Format: BIDS-compliant (OASIS-1), raw + derivatives (OASIS-3)

Supported Modalities

OASIS Clinical Measures

BIDS Preparation

Script: scripts/validate_oasis.py

Validates OASIS BIDS structure and generates a compliance report.

python skills/oasis-skill/scripts/validate_oasis.py \
  --input /path/to/OASIS/bids \
  --output /path/to/oasis_output/qc/bids_validation.csv

Features:

• BIDS directory structure validation
• Version detection (OASIS-1 vs OASIS-2 vs OASIS-3)
• Modality completeness check
• CDR/MMSE availability verification

Core Workflow (Never Bypassed)

1. Identify user target: full OASIS processing, sMRI only, 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_oasis.py.
5. Delegate to smri-skill for structural MRI processing.
6. If PET data available (OASIS-3), delegate to pet-skill.
7. If ASL data available (OASIS-3), delegate to asl-skill.
8. If phenotype extraction is requested, run scripts/extract_oasis_phenotype.py.
9. If QC summary is requested, run scripts/oasis_qc_summary.py.
10. Save outputs into oasis_output/.

Modality Processing Delegation

Standard Output Layout

oasis_output/
├── bids/                   # BIDS-staged data (or validation report)
├── smri/                   # Structural MRI derivatives
├── pet/                    # PET derivatives (SUVR maps)
├── asl/                    # ASL derivatives (CBF maps)
├── phenotype/              # Merged phenotype tables (CDR, MMSE, demographics)
├── 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 OASIS data validation.

• If the task starts from OASIS data already present on disk and only asks for BIDS validation:
  • Skip the download stage
  • Default to the narrow path local OASIS 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

• OASIS-1 is the most commonly used version; OASIS-3 adds PET and longitudinal data.
• Age range is very wide (18-96); analyses should account for age effects.
• CDR is the primary dementia staging tool; MMSE provides additional cognitive screening.
• OASIS-1 uses 1.5T scanner; resolution is lower than modern 3T/7T datasets.
• oasis-skill is orchestration-only; detailed preprocessing logic remains in modality skills.

When to Call This Skill

• User asks for end-to-end OASIS workflow.
• User asks to process OASIS structural MRI data.
• User needs BIDS validation for OASIS data.
• User asks to extract OASIS phenotype data (CDR, MMSE, demographics).
• User asks for Alzheimer's disease neuroimaging analysis.

Complementary / Related Skills

• smri-skill → structural MRI preprocessing
• pet-skill → PET imaging (amyloid, tau)
• asl-skill → arterial spin labeling perfusion
• fmri-skill → functional MRI (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

• OASIS: https://www.oasis-brains.org/
• Marcus et al. (2007): Open Access Series of Imaging Studies (OASIS). Journal of Cognitive Neuroscience.
• OpenNeuro ds000014

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