Alejandro de la Vega
neurozorro.bsky.social
Alejandro de la Vega
@neurozorro.bsky.social
Building tools for better neuroscience - Research Prof, UT Austin 🤘

I like cities and bikes 🚴 - Chair, Austin BAC
We showcase these tools by evaluating a heuristic based, and LLM based method for extracting participant demographics, namely sample size
September 11, 2025 at 10:24 PM
Next up, we present labelbuddy, a simple, effective and flexible tool for annotating documents. We use this tool to generate manual labels for evaluating automated extraction methods
September 11, 2025 at 10:24 PM
pubget makes it easy to downloads thousands of studies using a single command line operation. you also seamlesslyfit a neurosynth style meta-analysis to the results of your search.
September 11, 2025 at 10:24 PM
We present a set of tools to facilitate mining biomedical texts, and apply it to neuroimaging as an example.

First up, we make it easy to download texts from the open-access subset of PubMedCentral using pubget
September 11, 2025 at 10:24 PM
Yep, Cornyn basically told me "deal with it" when I wrote to his office about the attack on science.
February 16, 2025 at 3:07 PM
Red states with big public universities right now
February 9, 2025 at 6:02 PM
Cha ching 🪙
December 18, 2024 at 3:50 AM
December 13, 2024 at 9:23 PM
Obligatory NOA post.

Happy to share that Neurosynth Compose lives on!

Stay tuned for exciting NLP/LLM information extraction and a researcher-in-the-loop interface for “living” meta-analysis.
August 21, 2024 at 10:20 PM
Pretty sure fMRI is not the right tool for this
May 22, 2024 at 9:00 PM
The published articles that *do* make it to NeuroVault, represent a diverse slice of our field!

But, some topics are under-represented, particularly clinical research.

We can do better at reaching out to encourage them to share their valuable data!
March 8, 2024 at 2:45 PM
But, is everyone using NeuroVault?

Despite steady growth, the majority of researchers still don't upload to NeuroVault, and instead only report peak coordinates in tables

Help spread the word & take our survey to improve adoption!
March 8, 2024 at 2:44 PM
Most collections only feature group-level statistics, but a many also include subject-level statistics, enabling exciting future avenues for re-analysis!
March 8, 2024 at 2:44 PM
What kind of images are uploaded?

The majority represent statistical maps from fMRI-BOLD studies, and most are tagged with the Cognitive Atlas Ontology, facilitating re-use.

A growing number of beta maps are also present!
March 8, 2024 at 2:44 PM
Hundreds of collections are uploaded yearly, and after some publication lag, many collections are associated with a published record
March 8, 2024 at 2:43 PM
I'm in the wrong field
February 15, 2024 at 8:54 PM
Spoiler: even though regions like ACC activate for social processing tasks, this rate of activation is no higher than other studies in the database.

But, activity in regions like DMPFC, TPJ & precuneous *is* strongly associated with social processing.
February 2, 2024 at 10:00 PM
What's up with your handle
December 1, 2023 at 9:23 PM
The name inspired these fun DALL-E designed stickers!
November 29, 2023 at 6:52 PM
The results are automatically uploaded to NeuroVault, making it super easy to share your results
November 28, 2023 at 10:54 PM
... and run this bundle in the cloud in Google Colab, by copying and pasting your unique ID into our interactive notebook
November 28, 2023 at 10:54 PM
You can then specify a full meta-analysis to generate fully reproducible workflow...
November 28, 2023 at 10:53 PM
You can also add, edit and organize imaging coordinates of studies to make sure the data is 100% correct.
November 28, 2023 at 10:53 PM
We generate a PRISMA workflow for you automatically, making it easy to share your thought process with reviewers and other scientists
November 28, 2023 at 10:53 PM
One of the main features is a pretty slick Curation workflow. You can import studies from our database of over 20,000 studies (or from PubMed), and systematically select the ones you want to keep
November 28, 2023 at 10:53 PM