Key neurons act as 'ringleaders' to guide decision-making

A recent study by Princeton neuroscientists introduces a novel mathematical model that sheds light on how the brain integrates sensory cues – such as sights and sounds – when making decisions. 

The findings, published in the journal Nature Neuroscience, could deepen our understanding of neurological disorders like Alzheimer’s and may also help improve artificial intelligence systems, such as digital assistants and self-driving cars.

Conflicting sensory information

Everyday decisions require the brain to juggle multiple sensory inputs. A commuter walking to work, for instance, might see a glowing crosswalk signal indicating it is safe to cross. 

At the same time, they may hear the blaring siren of an ambulance approaching the intersection, forcing them to reassess their decision.

While scientists have long studied how the brain processes such competing signals, the precise mechanisms remain unclear.

One crucial region for decision-making is the prefrontal cortex, located just behind the eyes. This area is often regarded as the center of higher cognition. Previous studies have shown that neurons in the prefrontal cortex respond in complex ways during decision-making. 

For example, some neurons may fire only when a green traffic light appears and a car is blocking the crosswalk. However, a comprehensive model explaining how these neurons collectively process sensory information to generate behavioral decisions has remained elusive.

A new model to explain brain circuitry

Various mathematical models have been developed to explain how neural circuits link sensory information to behavior, each with its limitations. 

One commonly used approach involves recurrent neural networks – models with interconnected units designed to simulate decision-making. However, these models are often difficult to interpret due to the complexity of their connections.

In their latest study, postdoctoral researcher Christopher Langdon and assistant professor of neuroscience Tatiana Engel – both from Princeton University – propose a new framework called the latent circuit model.

Instead of analyzing the entire brain network in overwhelming detail, Langdon and Engel suggest focusing on a few key neurons that act as “ringleaders” in decision-making, influencing the behavior of the entire network. 

This approach, known as a low-dimensional mechanism, seeks to identify a small number of core processes that drive neural decision-making.

“The goal of the research was to understand if low-dimensional mechanisms were operating inside large recurrent neural networks,” Langdon said.

To test their hypothesis, the researchers applied the latent circuit model to a context-dependent decision-making task performed by humans, monkeys, and computer models.

Testing the decision-making model 

In the experiment, participants were first shown a context cue – either a square or a triangle. Next, they saw a moving grid, which acted as the sensory cue. 

Depending on the shape they were initially shown, participants had to focus on either the color (red vs. green) or the motion (left vs. right) of the grid to make their decision.

Using the latent circuit model, Langdon and Engel found that when motion was the relevant cue, neurons processing shape information suppressed those that processed color. The opposite occurred when color was the critical cue.

“It was very exciting to find an interpretable, concrete mechanism hiding inside a big network,” Langdon said.

Mapping decision-making in the brain

The latent circuit model allows researchers to predict how decision-making changes when specific neural connections are strengthened or removed. By testing these predictions, scientists can verify whether the underlying connectivity is truly necessary for the task.

Indeed, the study confirmed that task performance deteriorated in predictable ways when certain neural connections were disrupted.

“The cool thing about our new work is that we showed how you can translate all those things that you can do with a circuit onto a big network,” Langdon said. 

“When you build a small neural circuit by hand, there’s lots of things you can do to convince yourself you understand it. You can play with connections and perturb nodes, and have some idea what should happen to behavior when you play with the circuit in this way.”

Implications for neuroscience and AI

With the human brain containing more neurons than there are stars in the Milky Way, understanding its complexity is a daunting challenge. 

However, the latent circuit model opens new possibilities for studying how groups of interconnected neurons give rise to decision-making processes.

Disruptions in decision-making are a hallmark of numerous neurological and mental health disorders, including depression, schizophrenia, and ADHD. By revealing the mathematical computations that drive decision-making, this model may help researchers understand these conditions more effectively.

Beyond neuroscience, these insights could also enhance artificial intelligence. AI-driven systems, from digital assistants like Alexa to self-driving cars, rely on decision-making algorithms. Understanding how the human brain efficiently processes information could lead to more adaptable and intelligent AI systems.

Future research directions 

The next phase of research will involve applying the latent circuit model to other commonly studied decision-making tasks.

“A lot of the tightly controlled decision-making tasks that experimentalists study, I believe that they likely have relatively simple latent mechanisms,” Langdon said. “My hope is that we can start looking for these mechanisms now in those datasets.”

By refining this approach, neuroscientists aim to unlock more secrets of how the brain translates complex sensory inputs into clear, decisive actions.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–