We build brain imaging processing tools, benchmarking suites, and pipelines that make neuroimaging analyses more robust. By streamlining complex workflows and increasing transparency and reproducibility, we help democratize brain imaging research across the scientific community.
Advancing Methods
Center for Data Analytics, Innovation, and Rigor (DAIR)
Improving access to mental health care and the rigorous tools clinicians and researchers rely on.
About Our Work
DAIR brings together experts in neuroscience, psychology, data science, and engineering to pursue projects that have long been inaccessible due to their technical complexity and interdisciplinary nature.
We develop novel computational methods, experimental techniques, and open-source tools to improve the reproducibility, reliability, and transparency of research findings. Alongside this foundational work, we collaborate with clinical experts at the Child Mind Institute to apply data science directly to real-world challenges — improving the quality of front-line care, creating resources used globally, and engaging with communities in need.
Our Projects
Ask Kai
Ask Kai is a mental health symptom checker built in collaboration between clinicians and data scientists. It can be flexibly adapted across diverse contexts through easy integration of new questions, resources, and conversation flows.
Clinician Toolkit
The Clinician Toolkit is an AI-supported portal that reduces clinical administrative burden by semi-automating report writing and data retrieval. It offers auto-populating templates, synthesizes assessment data, and responsibly applies LLMs to draft clinical summaries.
Synthetic Data Generation
To overcome barriers in openly sharing and working with rich text data (e.g., journal entries and clinical notes), we are developing methods to generate synthetic datasets. This can expand the reach of clinical data without any risk to patients.
Styx & NiWrap
Styx is a compiler that transforms tool documentation into modern data science ecosystems. NiWrap is a collection of nearly 2,000 neuroimaging tools compiled using Styx, representing the largest unified collection of brain imaging software available.
Automated Mental Status Evaluation
Our Automated Mental Status Evaluation (autoMSE) turns accessible, everyday signals — such as data from wearables and smartphone apps — into reliable measures of well-being to create a more holistic, accurate picture of mental health and strengthen care.
NMIND
The NMIND consortium promotes reproducible neuroimaging research through community-driven software development standards. Our team leads the development of these standards, along with transparent tool review and certification processes.
Our Publications
Why Experimental Variation in Neuroimaging Should be Embraced
Brain imaging, like many areas of research, lacks accessible ground truths, leading scientists to debate comparable but conceptually distinct approaches. This work demonstrates how embracing variability in data analysis can improve the generalizability of results, and that these disagreements are a source of strength for the field rather than a weakness.
How Data Science Competitions Accelerate Brain Health Discovery
Data science competitions provide opportunities to crowdsource solutions for brain data research and clinical application development. Here, we discuss how researchers can leverage data science competitions to explore data heterogeneity, assess modelling challenges, generate hypotheses, and promote inclusive research.
Moving Beyond Processing- and Analysis-Related Variation
This study evaluates five independently developed fMRI preprocessing pipelines, showing that even when handling identical data, low-to-moderate inter-pipeline agreement limits cross-study reproducibility. We highlight the importance of comparing and transparently reporting analytic configurations, since widely discussed and commonly overlooked decisions can lead to marked variation.
Styx: A Multi-Language API Generator for Command-Line Tools
Many powerful brain imaging tools use complex command-line interfaces that are difficult to integrate into modern workflows. Here, we describe Styx, a compiler that automatically creates easy-to-use wrapper functions for command-line tools. Additionally, we introduce NiWrap, a collection of over 1,900 neuroimaging tool interfaces. Together, these tools make it easier to build reproducible analysis pipelines across different software packages in Python, R, and TypeScript.
If you are interested in exploring a collaboration, contact us at [email protected].
Collaborate With Us
Collaboration is central to our work. From hackathon-style sprints to long-term initiatives, we actively seek opportunities to engage with others. While the projects highlighted above represent only a fraction of our work, our general areas of interest include:
We focus on closing gaps in mental health care by using technology to streamline existing workflows or expand access to resources. Our solutions aim to increase the fidelity of mental health assessments while reducing the burden on clinicians and patients. Our goal is a community with easy access to relevant resources, a deepened understanding of their symptoms, and the ability to meaningfully take ownership over their mental health care.
We improve data infrastructures, accessibility, and availability within the Child Mind Institute and for the broader community. By fostering open science practices and engaging with youth and global data science communities, we help accelerate and democratize mental health research.
We integrate data from multiple sources — including actigraphy, free-form text, and video — to build a more comprehensive picture of mental health. This holistic approach enables richer patient phenotyping, ultimately leading to deeper insights into the complex factors that influence mental health outcomes.
Our Team
Gregory Kiar, PhD
Director, Center for Data Analytics, Innovation, and Rigor (DAIR)
Daniel Low, PhD
Research Scientist, Center for Data Analytics, Innovation, and Rigor (DAIR)
Laura Newman, PhD
Program Manager, Center for Data Analytics, Innovation, and Rigor (DAIR)
Reinder Vos de Wael, PhD
Senior Scientific Software Developer, Center for Data Analytics, Innovation, and Rigor (DAIR)
Florian Rupprecht, MSc
Senior Scientific Software Developer, Center for Data Analytics, Innovation, and Rigor (DAIR)
Tamsin Rogers
Junior Scientific Software Developer, Center for Data Analytics, Innovation, and Rigor (DAIR)
David LoBue, MSc
Senior Machine Learning Engineer, Center for Data Analytics, Innovation, and Rigor (DAIR)
Azaadeh Goharzad, PhD
Scientific Software and Engineering Generalist, Center for Data Analytics, Innovation, and Rigor (DAIR)
Jason Kai, PhD
Postdoctoral Fellow, Centers for Integrative Developmental Neuroscience (IDN) and Data Analytics, Innovation, & Rigor (DAIR)
Maya Roberts
Data Analyst, Center for Data Analytics, Innovation, and Rigor (DAIR)
Alp Erkent, MD, MA
Data Analyst, Strategic Data Initiatives (SDI)
Science Today for Better Care Tomorrow
Learn more about our science team’s other initiatives
About Our Science
Across six interconnected focus areas, our scientists build foundational knowledge of the developing brain, generate better tools for discovery, and more.
Healthy Brain Network
The goal of the Healthy Brain Network is bold: to find biological markers of childhood mental health and learning disorders.
On the Shoulders of Giants Scientific Symposium
An annual celebration of scientific collaboration and breakthroughs in our understanding of the brain.
Mirror: A Journaling App for Self-Reflection
Backed by robust clinical research, Mirror offers the privacy teens need to reflect, build resilience, and become more comfortable with who they are.
