[DSRP Evidence](https://dsrpevidence.org/)

# Adaptive chunking improves effective working memory capacity in a prefrontal cortex and basal ganglia circuit

## Details

**Authors** Soni & Frank

**Year** 2025

**Publisher** eLife

**Discipline** Neuroscience

[Read it at the publisher](https://doi.org/10.7554/eLife.97894) 
10.7554/eLife.97894

## In authors' words

### Abstract

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.

### What they set out to do (purpose)

To test how neural circuits form chunks and whether adaptive part–whole grouping increases usable working memory.

### What they found (results)

A biologically grounded PFC–basal ganglia network learned to adaptively chunk nearby items — binding parts into wholes — increasing effective working-memory capacity and matching human set-size behavior.

## Commentary

### In short

The mind binds individual parts into higher-order wholes (chunks).

### In more detail

Grouping individual items into higher-order chunks (part→whole binding) expands the mind's capacity to hold structured wholes. Purpose: To test how neural circuits form chunks and whether adaptive part–whole grouping increases usable working memory. Finding: A biologically grounded PFC–basal ganglia network learned to adaptively chunk nearby items — binding parts into wholes — increasing effective working-memory capacity and matching human set-size behavior. [Added 2026-07-30 as new evidence beyond the original 109; DSRP structural basis: S axiom (Containment / Compositional Identity): chunking literally composes parts into an emergent whole with new capacity properties.]

**Patterns it shows** S

**Added** 2026-07-30

**How to cite this** Soni & Frank (2025). Adaptive chunking improves effective working memory capacity in a prefrontal cortex and basal ganglia circuit. eLife.