A team of neuroscientists at the Picower Institute for Learning and Memory, Massachusetts Institute of Technology (MIT), has uncovered how the brain maintains distinct memories of options during the decision-making process. Published in the journal iScience, the study reveals that the brain dynamically recruits ensembles of neurons that produce coordinated electrical patterns, enabling it to clearly separate and remember both chosen and unchosen options.
What Happened
The lead researcher Huidi Li and senior author Earl K. Miller, Picower Professor in MIT’s Department of Brain and Cognitive Sciences, conducted experiments where trained animals played a task involving choices between two targets, each associated with a different reward value. The targets and their values were presented sequentially, compelling the brain to hold and distinguish multiple representations simultaneously.
During these sessions, the researchers recorded electrical activity from hundreds of neurons in the lateral prefrontal cortex, a brain region critical for linking options, their values, and actions. Using specialized decoding algorithms called “declassifiers,” the team identified functional ensembles whose collective neural activity patterns distinctly represented decision-relevant information, such as target identities and assigned reward values.
By visualizing these neural patterns as geometric planes within a three-dimensional graph space, the scientists observed that before any values were assigned or decisions made, the ensembles represented options separated by their order of presentation—“target 1” and “target 2” formed orthogonal, or nearly perpendicular, patterns. After decisions, however, the brain reorganized these representations: chosen targets shared parallel neural patterns distinct from those of unchosen targets, showing a clear sorting based on choice outcome rather than presentation order.
Key Facts
The research was conducted at MIT’s Picower Institute, with key contributors including Huidi Li, Earl K. Miller, Nikolaos Chrysanthidis, Scott Brincat, and Jonas Rose. The study appeared in the May 2024 issue of iScience. Animals were trained to respond to two sequentially presented options in a decision-making task, while electrophysiological recordings measured cortex neural activity. The researchers employed computational decoding methods to analyze hundreds of neurons’ coordinated activity during the task.
What This Means
This study advances understanding of how the brain dynamically represents complex information during decision-making. By organizing neural activity into distinct ensembles that shift from encoding the order of options before choice to encoding chosen status afterward, the brain effectively prevents interference between different potential actions. This mechanism ensures both accurate decision-making and the retention of information about unselected options, critical for learning and future choices.
For individuals and researchers, this finding highlights the brain’s remarkable flexibility and sophistication in processing competing information—a process that underpins everyday decisions from simple choices to complex judgments. The ability to maintain clear, separate memories of options and outcomes may also shed light on how the brain tracks rewards and updates behavior, informing fields ranging from behavioral economics to artificial intelligence design.
Background
The lateral prefrontal cortex has long been implicated in integrating information about options, their values, and actions during decision-making. Prior studies suggested that neural ensembles and brain wave patterns support the flexible grouping of neurons to represent transient information. Earl K. Miller’s earlier work revealed that brain waves enable rapid recruitment of neurons into different ensembles to meet changing cognitive demands.
Analysis
According to Miller and Li, the observed reorganization of neural subspaces during different decision phases prevents interference and packages choices for downstream brain areas that transform decisions into actions. The recruitment of multifunctional neurons into distinct ensembles on a round-by-round basis reflects the brain’s adaptability rather than fixed, specialized circuits.
Miller noted that maintaining separate neural representations of chosen and unchosen options supports credit assignment—remembering which choices yielded rewards—to facilitate learning from experience.
What Remains Unclear
The researchers acknowledge that while they have elucidated how neural ensembles represent choices dynamically, they have not fully classified the specific neuronal subtypes involved or detailed the mechanisms that govern ensemble recruitment. The extent to which these findings generalize across different decision types or species remains to be investigated.
What Comes Next
Future work will explore how these ensemble patterns influence downstream brain regions responsible for action execution and learning. The team also plans to investigate how brain wave dynamics interact with ensemble formation during decision-making tasks. These insights may inform novel approaches to understanding cognitive disorders involving impaired decision-making.
Sources
This article is based on reporting and publicly available information from the following sources:
Read more Science Discoveries stories on Goka World News.