Using, misusing, and improving online machine learning-based meta-analysis of neuroimaging published data: A perspective on NeuroQuery
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Online, text-based meta-analysis tools for large databases represent a new digital advance for medical, health, and neuroscience research, among other fields. NeuroQuery is an instance of such a tool for neuroimaging research; it employs supervised machine learning to draw from over 13,000 publications and perform a meta-synthesis, generating predictive fMRI scans based on keyword combinations. Although NeuroQuery is a sophisticated tool, a lack of understanding of how it practically works and its limitations may lead to flawed results and conclusions, undermining its potential value. We review potential risks and limitations, including algorithm limitations, potential biases in the database, and user misinterpretation. Simulating the perspective of an end user, we present an example of unreliable but possible metanalysis results on autistic spectrum disorder (ASD). We then report an analysis of the underlying query from a sophisticated user perspective. Using the same examples, we illustrate possible improvements for the use of NeuroQuery and identify how this tool may be valuable in the context of emerging machine-learning meta-analytical approaches. Although a thorough understanding of NeuroQuery is helpful, we conclude that understanding its limitations plays a more critical role in ensuring validity and reliability of its use. While NeuroQuery is currently not appropriate for rigorous scientific analysis, it could be useful for hypothesis development, preliminary fMRI data mining, exploratory and supplemental analysis as well as literature survey.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.050 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.003 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it