Mariya Toneva is a faculty member at the Max Planck Institute for Software Systems, where she leads the Bridging AI and Neuroscience (BrAIN) group. Her research bridges natural language processing, machine learning and cognitive neuroscience to develop computational models that deepen our understanding of how the brain processes language and guide the creation of more human-aligned AI systems. Her pioneering work at this intersection has been recognized and supported by the U.S. National Science Foundation, the German Research Foundation and the European Research Council through an ERC Starting Grant.
Mariya Toneva
Faculty member
Max Planck Institute for Software Systems
Selected articles
- “Improving semantic understanding in speech language models via brain-tuning” | arXiv
- “Interpreting and improving natural-language processing (in machines) with natural language-processing (in the brain)” | NeurIPS
- “Combining computational controls with natural text reveals aspects of meaning composition” | Nature Computational Science
- “Joint processing of linguistic properties in brains and language models” | NeurIPS
Explore more from The Transmitter
Leucovorin saga, and more
Here is a roundup of autism-related news and research spotted around the web for the week of 15 June.
Leucovorin saga, and more
Here is a roundup of autism-related news and research spotted around the web for the week of 15 June.
By
Jill Adams
16 June 2026 | 2 min read
Models at the speed of thought: How AI coding is reshaping theoretical neuroscience
Agentic coding makes it possible to specify a neuroscience model in hours instead of months. Seven neuroscientists weigh in on what that tectonic change may bring to the field.
Models at the speed of thought: How AI coding is reshaping theoretical neuroscience
Agentic coding makes it possible to specify a neuroscience model in hours instead of months. Seven neuroscientists weigh in on what that tectonic change may bring to the field.
By
Brian DePasquale
16 June 2026 | 17 min read
Writing science that humans and machines can read
Large language models are now routinely used to search, summarize and synthesize the literature at scales impossible for any individual researcher—yet scientific publishing has not adapted to that reality.
Writing science that humans and machines can read
Large language models are now routinely used to search, summarize and synthesize the literature at scales impossible for any individual researcher—yet scientific publishing has not adapted to that reality.
By
Rachel Parkinson
15 June 2026 | 7 min read