Summary: A new study reveals that working memory limitations are due to learning challenges rather than storage capacity. Using computer models of the basal ganglia and thalamus, researchers show that holding too much information can disrupt the brain and impair the ability to effectively learn and use stored data. I did.
This model demonstrated that the brain compensates for this limitation and improves efficiency by strategically “chunking” relevant information. These findings also highlight dopamine-related disorders such as Parkinson’s disease and ADHD, suggesting novel therapeutic approaches targeting the basal ganglia and thalamus.
Important facts
Learning to form memory limits: The brain limits the working memory capacity to prevent overload and learning in conditions such as Parkinson’s disease and ADHD.
Source: Brown University
Working memory allows humans to juggle a variety of information in short-term scenarios, such as creating mental grocery lists, shopping, and dialing phone numbers.
Scientists agree that working memory is limited in capacity, but they provide competing theories about how and why this is true. However, new research from scientists at the Kearney Institute of Brain Science at Brown University shows why working memory limitations exist.
Michael Frank, a professor of cognitive and psychological sciences at the Carney Institute, and Anelli Soni, a graduate student in his lab, have developed a new computer model of the basal ganglia and thalamus – in working memory Related brain parts – this shows why working memory limitations exist.
According to their research published in Elife, the answers relate to learning.
“The simulations we performed show that if you hold more than a few items at a time it becomes difficult to learn how to manage so many information at once, and that your brain is confused and unusable. “The information to keep,” Soni said. “At the same time, our research shows that when faced with these limitations, the brain responds by learning to strategically harness the mechanisms to conserve space.”
The neurotransmitter dopamine plays an important role in how learning is related to working memory, and researchers say these findings include Parkinson’s disease, attention deficit hypertrophic disorder (ADHD), and schizophrenia. He said it has shed new light on dopamine-related disorders.
The team built a new computer model of the brain in 2018, which was produced by researchers from Frank’s lab and Matt Nasser, an assistant professor at Carney Institute for Neuroscience, in 2018, to test the results of the brain. By doing so, we arrived to discover. and cognitive and psychological sciences.
The research established that humans can “chunk” information by compressing relevant information together in working memory to save space.
When Soni tried out the model for a version of her 2018 experiment, she knew she could successfully build a brain-like computer model that could compress information. She showed the model a screen with blocks of colours pointed in different directions, then asked to remember which blocks of colours were pointing in which direction.
Over the course of many trials, the model learned how to strategically compress information, and began to play similar colors, such as blue and light blue.
According to Soni, lab simulations with the new model refer to learning rather than capacity, not capacity. She established this by running a trial on the model without the ability to chunk, but there was plenty of room to store the items.
She found that models with chunking mechanisms can strategically store information in full storage capacity, but models without chunking mechanisms can access such large amounts of storage, both in storage and in storage. It doesn’t seem to notice that it’s getting worse. Get the item.
A critical component of the model’s learning process is the mechanisms that emulate the human brain’s dopamine delivery system, Soni said. When the model was doing a better job of recalling more block orientations as they charged similar colors together to save space, the dopamine delivery system kicked off and faced the same strategy on the model We instructed to continue using this strategy to set of constraints in subsequent exams.
In another part of the experiment, Soni changed 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 would not learn how to use its storage space efficiently and did not chunk items frequently. That’s what it was.
These new discoveries show how the brain science of computationality advances psychiatry, said Frank, who directs Kearney’s Center for Computational Brain Science.
“Take Parkinson’s disease as an example,” Frank said. “Most people consider it a movement disorder because the movement changes are so obvious. However, we can see that there are changes in working memory in patients with Parkinson’s disease, which are generally the anterior head. Although it has been treated with drugs targeting the anterior cortex, our findings suggest that there is a need to test whether drugs targeting the basal ganglia and thalamus help to improve symptoms. ”
Frank said there is a growing understanding of what happens within the basal ganglia and thalamus for those diagnosed with dopamine-related disorders.
Funding: This study 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). The 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: Image credited to Neuroscience News
Original research: Open access.
“Adaptive chunking improves effective working memory capacity in the prefrontal cortex and basal ganglia circuits,” Michael Frank et al. Elif
Abstract
Adaptive chunking improves effective working memory capacity in the prefrontal cortex and basal ganglia circuits
How and why is working memory (WM) capacity limited? Traditional cognitive accounts focus on either the limits of the number or items (slot models) that can be saved or the loss of accuracy with increased load (resource models).
Here, we can learn that neural network models of the prefrontal cortex and basal ganglia reuse the same prefrontal cortex population to store multiple items, and are constrained such as resources in the system, such as slots. It shows that connections induce trade-offs between quantity and accuracy. Of information.
Such “chunking” strategies are adapted as a function of reinforcement learning and WM task requirements, mimicking human performance and normative models.
Furthermore, adaptive performance requires dynamic ranges of dopaminergic signals to adjust striatal gating policies, and WM difficulties in patient populations such as Parkinson’s disease, ADHD, and schizophrenia. Provides new interpretations.
These simulations suggest computational limitations rather than anatomical limitations on WM volume.
