summary: For years, studies of “London taxi drivers” have suggested that elite navigators have physically larger brain areas. However, new research suggests this may not be the case in the general population.
Researchers used advanced deep learning and convolutional neural networks to analyze MRI scans of healthy young people to see if AI could find subtle structural patterns that predict navigational ability. Surprisingly, even the most sophisticated AI tools found no measurable association between the shape or size of brain regions, including the hippocampus, and how well a person is able to find their way.
important facts
- AI vs. Tradition: Simple size measurements sometimes suggested an association, but deep learning models that can detect complex structural patterns were unable to find “navigation signals” in healthy young people.
- Hippocampus and thalamus: The study specifically compared the hippocampus (the brain’s traditional “GPS”) and the thalamus (a control region). Navigation performance was indistinguishable between the two based on structure alone.
- healthy young people: Data were collected from 90 participants (mean age 23 years) who learned the route in a virtual environment.
- Prediction of disease and behavior: The researchers noted that while AI is good at predicting disease states (such as Alzheimer’s disease), it is difficult to map everyday behavioral functions such as spatial navigation.
- Redefining research: The results of this study challenge the idea that “more areas” in the brain equals improved cognitive ability and suggest that: function and connectivity It may be more important than macroscopic structure.
sauce: UT Arlington
University of Texas at Arlington researcher Stephen Weisberg found that advanced artificial intelligence tools fail to reveal a clear link between brain structure and navigational abilities in healthy young people, challenging long-held ideas about how the brain helps us find our way.
For decades, many in the scientific community believed that people with good navigation skills, such as learning and recalling complex routes quickly, may have larger or differently shaped brain areas than other people.
For example, a famous study of London taxi drivers suggested that intensive navigation training can increase “real estate” in certain parts of the brain.
In the new study, Dr. Weisberg and his team, including University of Florida Ph.D. candidate Ashish Sahoo, tested these assumptions using new analytical techniques, including deep convolutional neural networks and other machine learning models that can go beyond simple size measurements to detect subtle patterns in brain scans.
Despite these advanced methods, researchers failed to find a measurable link between brain structure and navigation ability in healthy young people.
Understanding navigation is important given its practical impact on daily life, including independence, memory, and risk of dementia.
“Given the quality of the data obtained from MRI scans and healthy young adult populations, there appears to be no detectable signal using these advanced metrics,” said Professor Weisberg, who conducted the study at the University of Florida before joining UT Arlington last fall as part of the RISE 100 initiative.
This study was published in a peer-reviewed journal neuropsychologyanalyzed data from 90 participants with an average age of 23.1 years. Participants learned two routes using a virtual environment.
The results showed that there was little difference in navigation performance when comparing two brain regions: the thalamus, which acts as a control region, and the hippocampus, an area traditionally associated with navigation and memory.
Although this finding shows that there are limits to what AI can currently reveal about everyday cognitive skills, the technology remains a powerful research tool. Weisberg said more robust models could detect differences in future studies.
“Our study should be one data point in a larger picture of what AI can tell us about how brain structure and function maps to behavior,” Weisberg said.
“Machine learning and AI have been fairly successful in predicting disease states. What we’re interested in is whether these models can help with behavioral functions such as cognitive training and education.”
Future studies will focus on larger samples and older populations, Weisberg said.
“Our ability to move allows us to do basically everything we do. Studying how the brain supports navigation can help us understand what we need when things go well and what we lack when things go wrong.”
Answers to key questions:
answer: Not necessarily. Years of intensive training (like London taxi drivers) can still cause structural changes. But for an average healthy person, being a “good” or “bad” navigator is not determined by the physical shape of the brain. It probably has something to do with how your neurons work fire It’s not about how much space it takes up.
answer: On a physical, macroscopic level, yes. This study suggests that a large hippocampus is not required for “elite” navigation skills in young people. The difference likely lies “under the hood” in minute wiring and chemical signals that cannot yet be detected by AI-processed MRI scans.
answer: Navigation is often the first skill to decline in Alzheimer’s disease. If a structural “baseline” for good navigation in healthy people cannot be found, we suggest that early diagnosis should focus on: behavioral change and functional Rather than just looking at the size of the hippocampus in a scan, we measure connectivity.
Editorial note:
- This article was edited by the editors of Neuroscience News.
- Journal articles were reviewed in full text.
- Additional context added by staff.
About this AI and neuroscience research news
author: Drew Davison
sauce: UT Arlington
contact: Drew Davison – UT Arlington
image: Image credited to Neuroscience News
Original research: Open access.
“A Deep Learning Approach to Mapping Individual Differences in Macroscopic Neural Structure with Changes in Spatial Navigation Behavior” by Ashish K. Sahoo, Hajymyrat Geldimurodov, Kaleb E. Smith, Aaron Zygala, Yiming Cui, Mahsa Lotfollahi, Kuang Gong, Alina Zare, and Steven M. Weisberg. neuropsychology
DOI:10.1016/j.neuropsychologia.2025.109352
abstract
A deep learning approach that maps individual differences in macroscopic neural structure to changes in spatial navigation behavior
Understanding the link between structural properties of the human brain and individual differences in behavior is an ongoing endeavor challenged by the complexity of the brain.
Previous approaches were limited by simple neural structural measures such as brain volume and cortical thickness, giving way to more sophisticated modeling approaches.
Empirical evidence using these simpler metrics may indicate that hippocampal structure is involved in individual differences in spatial navigation ability, particularly in older adults and expert navigators (e.g., London taxi drivers). However, a preregistered, high-powered study in typical young adults found no association between hippocampal volume and navigational ability.
Here, we follow a data-driven approach to develop and compare deep learning techniques (graph convolutional neural networks, GCNNs, 3DCNNs) and analyze whether complex aspects of brain structure predict spatial navigation abilities in young populations.
To do so, we trained a GCNN and a 3DCNN on a T1 MRI dataset (N = 90) to predict navigation ability as measured by an objective virtual reality test of spatial memory in which participants create as accurate a map as possible of a highly realistic virtual environment.
We found that all approaches had weak predictive values on the retained test data, despite a good fit with the training data.
Although these results may indicate the need for much larger datasets that include more comprehensive behavioral measures (as this study was limited to one measure) to improve predictability, they may also support the notion that structural features of the hippocampus may not be the primary factor associated with navigational ability in healthy young adults.
