Summary: Scientists at the Salk Institute have developed a rigorous, information-theory-based method to quantify synaptic strength, the precision of synaptic plasticity, and the amount of information a single synapse can store. Applying this approach to hippocampal synapses reveals that individual synapses can hold roughly ten times more information than prior estimates suggested.
These results sharpen our understanding of how learning and memory are encoded at the synaptic level and provide a new quantitative tool to study how those processes change during development, aging, and in neurodegenerative or neurodevelopmental disorders.
Key Facts:
- Synaptic plasticity quantified: The study measures synaptic strength, the consistency (precision) of plastic changes, and information storage capacity using information theory.
- Higher capacity: Results indicate synapses can store about 10 times more information than earlier methods estimated.
- Research impact: The method is scalable and can be applied to diverse datasets to advance research on learning, memory, and brain disorders such as Alzheimer’s disease.
Source: Salk Institute
Learning a new word repeatedly strengthens the brain’s connections, making recall faster and easier. That strengthening and weakening of connections is synaptic plasticity, the core mechanism behind learning and memory.
Measuring the anatomy of synapses can indicate relative synaptic strength, but estimating how precisely synapses change (plasticity precision) and how much information a synapse can reliably store has been more difficult. The Salk team introduced an analytical framework that addresses these questions with statistical rigor.
To capture how the brain learns and retains information, researchers need to quantify both how much synaptic strength changes with experience and how many distinct strength levels a synapse can reliably represent. The new method provides a principled way to do both.
When information travels through neural circuits, signals move from one neuron to another across synapses located at the tips of dendritic spines. Each dendritic spine houses a synapse—a tiny electrochemical junction through which neurons communicate. Some axons form synapses on multiple nearby spines of the same dendrite; these coactivated synapse pairs are especially useful because they share activation history and therefore serve as a natural test of plasticity precision.
The research team analyzed synapse pairs in the rat hippocampus, a region critical for learning and memory, using concepts from information theory. Information theory models information transmission through a noisy channel and provides a discrete measure of information in bits, while explicitly accounting for variability—or noise—in biological signals.
Instead of assuming signal distributions as earlier approaches did, the Salk method divided synaptic strengths into 24 distinct categories based on measurable differences in dendritic spine head volumes. By comparing synapse pairs with identical activation histories, the authors assessed how consistently synaptic strength was modulated across those categories.
The analysis revealed that synapse pairs had highly similar spine sizes and strengths, indicating strong precision in synaptic plasticity. Using Shannon entropy to quantify Synaptic Information Storage Capacity (SISC), the investigators found each of the 24 distinguishable strength categories carried between approximately 4.1 and 4.59 bits of information.
Compared with older techniques, this information-theory approach is more comprehensive because it incorporates biological noise and provides a clear bit-based metric for storage capacity. It also proved scalable: the same framework can be applied to larger and more varied datasets across brain regions, species, and developmental stages.
Kristen Harris, coauthor and professor at the University of Texas at Austin, notes that this quantitative view of synaptic strength and plasticity can accelerate research into how memories form and degrade, and can be used to study brains across ages and conditions. Terrence Sejnowski, senior author, emphasizes that large-scale projects—such as atlases of human brain cell types—will gain from incorporating this new analytical tool.
Beyond basic neuroscience, the method offers a way to probe when and how information storage breaks down in diseases like Alzheimer’s, and it may help researchers worldwide study how brains acquire new skills and retain information over short and long time scales.
About this synaptic plasticity research news
Author: Terrence Sejnowski
Source: Salk Institute
Contact: Terrence Sejnowski – Salk Institute
Image: Image credited to Neuroscience News
Original Research: Open access. “Synaptic Information Storage Capacity Measured With Information Theory” by Terrence Sejnowski et al., published in Neural Computation. The study was published on April 23, 2024 and applies Shannon information theory to quantify synaptic storage capacity.
Abstract
Synaptic Information Storage Capacity Measured With Information Theory
Variation in synaptic strength can be quantified by anatomical measures such as dendritic spine head volume. Determining the precision of synaptic plasticity is essential for understanding how neural circuits store and retrieve information.
Pairs of synapses formed by the same axon onto the same dendrite share activation history and therefore serve as precise comparators for plasticity. This study uses Shannon information theory to quantify both the precision of synaptic plasticity and the information stored in synaptic dimensions, extending prior analyses that applied signal detection theory.
Dendritic spine head volumes in hippocampal CA1 were grouped into 24 nonoverlapping size categories corresponding to distinguishable synaptic strengths. Shannon entropy produced a lower bound of 4.1 bits and an upper bound of 4.59 bits of synaptic information storage capacity (SISC) across those categories. Comparison with a uniform distribution via Kullback–Leibler divergence showed a near-uniform distribution of spine head volumes across sizes, suggesting efficient use of available distinguishable states.
SISC is introduced as a generalizable analytic measure to probe synaptic strength, plasticity precision, and storage capacity across brain regions, species, developmental stages, and disease states. This tool provides a quantitative foundation for future investigations into how brains encode, maintain, and lose information.