Skip to content

neurostuff/neurometabench

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 

Repository files navigation

neurometabench

neurometabench is a structured dataset of neuroimaging meta-analyses designed to benchmark methods for automatically reconstructing meta-analyses using AI.

The initial set of meta-analyses was sourced primarily from Dr. Angie Laird’s laboratory, with metadata manually curated for each study. Our long-term goal is to expand this resource to include a broader range of meta-analytic datasets and to develop methods that leverage large language models (LLMs) for automatic extraction of meta-analytic information from full-text articles.


Dataset contents

For each study, the dataset includes:

  • Publication metadata (e.g., PMID)
  • Identification criteria (e.g., search parameters, methods used to find studies, total number of records identified)
  • Study objectives and inclusion/exclusion criteria
  • Final list and count of included studies

For studies sourced from Dr. Laird’s lab, we additionally provide the original analyses and corresponding coordinate data, enabling direct replication of published results. (TODO: Convert these analyses to NiMADS format for replication within NiMARE.)


Repository structure

All structured data files are located under the data/ directory.

Metadata tables

  • included_studies.tsv – Links meta-analyses to their included studies.
  • meta_datasets_pmid.txt – Reference list of PMIDs/PMCIDs for meta-analyses.
  • meta_datasets.tsv – Curated annotations for each meta-analysis (study characteristics, inclusion/exclusion criteria, search details, etc.).

Full-text study data

In addition to metadata tables, neurometabench provides the fetched text and extracted data for both meta-analyses and their included studies:

  • Meta-analysis study datadata/meta-studies/
  • Individual included studiesdata/studies/

Each contains two subfolders:

  • pmc-oa/ – Studies available in PubMed Central Open Access (included in this repository).
  • ace/ – Non–open-access studies (not distributed on GitHub; must be obtained manually).

Directory overview

data/
├── included_studies.tsv
├── meta_datasets_annotated_pmid_pmcids.txt
├── meta_datasets_annotated.tsv 
│
├── meta-studies/
│   ├── pmc-oa/   # Open-access meta-analysis full texts and extracted tables
│   └── ace/      # Non–open-access meta-analyses (not included in repo)
│
└── studies/
    ├── pmc-oa/   # Open-access included study full texts and extracted tables
    └── ace/      # Non–open-access included studies (not included in repo)

Availability

  • A subset of meta-analyses are openly available in PubMed Central (PMC). For these, full text and tables have been fetched and processed.
  • For studies not available in PMC, the full text must be downloaded manually and information extracted by hand.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages