Understanding the brain with AI-driven explanations and experiments - Microsoft Research
Skip to main content
Research
Publications<br>Code & data<br>People<br>Microsoft Research blog
Artificial intelligence<br>Audio & acoustics<br>Computer vision<br>Graphics & multimedia<br>Human-computer interaction<br>Human language technologies<br>Search & information retrieval
Data platforms and analytics<br>Hardware & devices<br>Programming languages & software engineering<br>Quantum computing<br>Security, privacy & cryptography<br>Systems & networking
Algorithms<br>Mathematics
Ecology & environment<br>Economics<br>Medical, health & genomics<br>Social sciences<br>Technology for emerging markets
Academic programs<br>Events & academic conferences<br>Microsoft Research Forum
Behind the Tech podcast<br>Microsoft Research blog<br>Microsoft Research Forum<br>Microsoft Research podcast
About Microsoft Research<br>Careers & internships<br>People<br>Emeritus program<br>News & awards<br>Microsoft Research newsletter
Africa<br>AI for Science<br>AI Frontiers<br>Asia-Pacific<br>Cambridge<br>Health Futures<br>India<br>Montreal<br>New England<br>New York City<br>Redmond
Applied Sciences<br>Mixed Reality & AI - Cambridge<br>Mixed Reality & AI - Zurich
Register: Research Forum
Microsoft Security<br>Azure<br>Dynamics 365<br>Microsoft 365<br>Microsoft Teams<br>Windows 365
Microsoft AI<br>Azure Space<br>Mixed reality<br>Microsoft HoloLens<br>Microsoft Viva<br>Quantum computing<br>Sustainability
Education<br>Automotive<br>Financial services<br>Government<br>Healthcare<br>Manufacturing<br>Retail
Find a partner<br>Become a partner<br>Partner Network<br>Microsoft Marketplace<br>Software companies
Blog<br>Microsoft Advertising<br>Developer Center<br>Documentation<br>Events<br>Licensing<br>Microsoft Learn<br>Microsoft Research
View Sitemap
Return to Blog Home<br>Microsoft Research Blog
At a glance
LLM-based models can predict the human brain’s responses to language with high accuracy. But what drives that performance is essentially unreadable: a vast collection of learned parameters, not scientific theories anyone can read.
Generative causal testing (GCT), developed in a collaboration between Microsoft Research, the University of California, Berkeley, the University of California, San Francisco, and Columbia University, distills these brain-prediction models into short verbal explanations of what each patch of cortex responds to: phrases like “food preparation” or “location names.”
GCT then closes the loop: an LLM writes new stories designed to activate a targeted brain area, subjects hear them in the scanner, and the region lights up only if the explanation is right.
In experiments, GCT confirmed known selectivity, teased apart neighboring place-processing regions long thought interchangeable, and revealed tiny prefrontal “micro-regions” tuned to specific concepts like dialogue, clock times, and measurements.
The explainability problem in language neuroscience
Over the past decade, LLMs have become the most accurate tools we have for predicting how the human brain responds to language. Feed an LLM the same story a person hears in an fMRI scanner, and the model’s internal representations can predict the activity of individual patches of cortex with remarkable fidelity. But this success comes with a catch: nobody can read these models. They are millions of inscrutable parameters that can’t be directly translated into interpretations. A model that predicts brain activity tells us that a region responds to language, but not what it is actually picking up on, whether it’s food, places, numbers, or something else entirely. As black-box models spread, the gap between prediction and understanding has become one of the central problems in computational neuroscience.
Turning black boxes into testable theories
In a new paper accepted in Nature Neuroscience, Microsoft Research scientists, in collaboration with scientists at the University of California, Berkeley, University of California, San Francisco, and Columbia University, introduce a framework to overcome this explainability crisis: generative causal testing (GCT). GCT distills brain-prediction models into short, readable accounts of what each patch of cortex responds to, then tests those claims. An LLM writes new stories engineered to activate a specific brain area, subjects hear them in the scanner, and if the explanation is correct, the targeted region lights up. The result is a method that translates uninterpretable predictive models back into the currency of science: concise hypotheses that can be confirmed or refuted in a follow-up experiment. An LLM writes new stories engineered to activate a specific brain area, subjects hear them in the scanner, and if the explanation is correct, the targeted region lights up. The result is a method that translates uninterpretable predictive models back into the currency of science: concise hypotheses that can be confirmed or refuted in a follow-up experiment.
Figure 1. The two steps of generative causal testing (GCT). In Step 1, the phrases that most strongly drive a brain region’s predictive model are summarized by an LLM into a...