Brain’s Working Memory Limits Tied to Learning, Not Just Capacity
·February 25, 2025
Summary: A new study reveals that working memory limitations stem from learning challenges rather than storage capacity. Using a computer model of the basal ganglia and thalamus, researchers showed that holding too much information at once confuses the brain, impairing its ability to learn and use stored data effectively.
The model demonstrated that the brain compensates for this limitation by strategically “chunking” related information together, improving efficiency. These findings also shed light on dopamine-related disorders like Parkinson’s and ADHD, suggesting new treatment approaches targeting the basal ganglia and thalamus.
Key Facts
- Learning Shapes Memory Limits: The brain restricts working memory capacity to prevent overload and optimize learning.
- Chunking Strategy: To conserve space, the brain compresses related information, improving recall efficiency.
- Dopamine’s Role: Disruptions in dopamine delivery impair memory efficiency, linking working memory deficits to conditions like Parkinson’s and ADHD.
Source: Brown University
Working memory is what allows humans to juggle different pieces of information in short-term scenarios, like making a mental grocery list and then going shopping or remembering and then dialing a phone number.
While scientists agree that the capacity of working memory is limited, they offer competing theories about how and why this is true. But new research from scientists at the Carney Institute for Brain Science at Brown University shows why limits on working memory exist.
Michael Frank, a professor of cognitive and psychological sciences affiliated with the Carney Institute, and Aneri Soni, a graduate student in his lab, developed a new computer model of the basal ganglia and the thalamus — the parts of the brain relevant to working memory — that shows why limits on working memory exist.
According to their study published in eLife, the answer has to do with learning.
“The simulations we ran show that if we did hold more than just a few items at a time, it becomes too difficult to learn how to manage so many pieces of information at once, such that the brain gets confused and can’t use the information it does store,” Soni said. “At the same time, our research demonstrates that when faced with these limitations, the brain responds by learning to strategically tap into a mechanism to help conserve space.”
Because the neurotransmitter dopamine plays an important role in how learning relates to working memory, the researchers said these findings shed new light on dopamine-related disorders such as Parkinson’s disease, attention deficit-hyperactivity disorder (ADHD) and schizophrenia.
The team arrived at their discovery by building and testing a new computer model of the brain that replicated the results of an experiment with humans conducted in 2018 by researchers in Frank’s lab and researchers in the lab of Matt Nassar, a Carney Institute assistant professor of neuroscience and cognitive and psychological sciences.
That study established that humans are capable of “chunking” information by compressing related pieces of information together in working memory to conserve space.
Soni knew she’d successfully built a brain-like computer model capable of compressing information when she challenged her model to a version of the 2018 experiment. She showed the model a screen with colored blocks oriented in different directions and then asked it to recall which colored block was pointing in which direction.
Over the course of a number of trials, the model learned how to strategically compress information and began chunking together similar colors, such as blue and light blue.
The lab’s simulations with the new model point to learning, rather than capacity, as the real driver of working memory, Soni said. She established this by running the trials on a model without the ability to chunk but with plenty of space to store items.
She found that while the model with the chunking mechanism was able to strategically store information to its full storage capacity, the model without the chunking mechanism did not seem to realize it had access to such a large amount of storage and was worse at both storing and retrieving the items.
A component critical to the model’s learning process is a mechanism that emulates the human brain’s dopamine delivery system, Soni said. When the model was doing a better job of recalling the orientation of a larger number of blocks because it had chunked similar colors together to save space, the dopamine delivery system kicked in, telling the model to continue to use this strategy when faced with the same set of constraints in subsequent trials.
In another part of the experiments, Soni altered the model’s dopamine delivery system to emulate what is known about dopamine levels in patients with Parkinson’s disease, schizophrenia and ADHD. When she challenged the model to the same series of trials, the results showed that without a healthy dopamine delivery system, the model did not learn how to use its storage space as efficiently and did not chunk items as often.
New findings like this show how computational brain science can advance psychiatry, said Frank, who directs Carney’s Center for Computational Brain Science.
“Take Parkinson’s disease as an example,” Frank said. “Most people think of it as a movement disorder because changes in movement are so obvious. But it turns out that Parkinson’s patients also have changes in working memory. They are generally treated with drugs that target the prefrontal cortex, but our findings suggest that we should be testing whether drugs that target the basal ganglia and thalamus help to improve symptoms.”
Frank said that an increased understanding of what takes place within the basal ganglia and thalamus for people diagnosed with dopamine-related disorders may spur clinicians to adopt different treatment options.
Funding: This research was supported by the Department of Defense (ONR MURI Award N00014-23-1-2792) and the National Institute of Mental Health (R01 MH084840-08A1, T32MH115895). Computing hardware was supported by the National Institutes of Health (S10OD025181).
About this memory and learning research news
Author: Corrie Pikul
Source: Brown University
Contact: Corrie Pikul – Brown University
Image: The image is credited to Neuroscience News
Original Research: Open access.
“Adaptive chunking improves effective working memory capacity in a prefrontal cortex and basal ganglia circuit” by Michael Frank et al. eLife
Abstract
Adaptive chunking improves effective working memory capacity in a prefrontal cortex and basal ganglia circuit
How and why is working memory (WM) capacity limited? Traditional cognitive accounts focus either on limitations on the number or items that can be stored (slots models), or loss of precision with increasing load (resource models).
Here we show that a neural network model of prefrontal cortex and basal ganglia can learn to reuse the same prefrontal populations to store multiple items, leading to resource-like constraints within a slot-like system, and inducing a trade-off between quantity and precision of information.
Such “chunking” strategies are adapted as a function of reinforcement learning and WM task demands, mimicking human performance and normative models.
Moreover, adaptive performance requires a dynamic range of dopaminergic signals to adjust striatal gating policies, providing a new interpretation of WM difficulties in patient populations such as Parkinson’s disease, ADHD and schizophrenia.
These simulations also suggest a computational rather than anatomical limit to WM capacity.
Join our Newsletter
I agree to have my personal information transferred to AWeber for Neuroscience Newsletter ( more information )
Sign up to receive our recent neuroscience headlines and summaries sent to your email once a day, totally free.
We hate spam and only use your email to contact you about newsletters. You can cancel your subscription any time.
Tags
brain researchBrown UniversityMemoryneurobiologyNeuroscienceworking memory
