Neurons and the Symphony of Learning: How Timing Shapes Information Encoding

A timing mismatch

Learning can occur over several seconds to minutes. However, the most understood mechanisms by which the brain encodes information operate at speeds much faster—approximately 1000 times quicker. These processes, referred to as Hebbian plasticity, imply that if two neurons that are interconnected both activate within a hundredth of a second, the connection between them gets stronger. This enables the information received by these linked neurons within this brief interval to be associated together. Yet, during activities, the information that must be encoded together is often spaced out by multiple seconds or minutes. So, how can neurons combine information over durations relevant for learning?

A new model of learning

Recently, researchers introduced a new neural framework for information encoding called behavioral timescale synaptic plasticity (BTSP), which addresses this inconsistency by showing that neurons can assimilate information over seconds, thus matching behavioral timescales. Indeed, during activities like navigation, neurons use BTSP to encode distinct locations. However, the biological processes through which neurons enact BTSP were previously unidentified.

This week, a team led by Dr. Anant Jain, Dr. Yoshihisa Nakahata, and Scientific Director Dr. Ryohei Yasuda unveiled key components of how BTSP functions in neurons, summarizing their extensive research into this significant plasticity model.

Dr. Yasuda shared the team’s motivation for this project, stating, “Understanding the specific molecules and mechanisms that neurons deploy to encode information is essential for grasping brain function and health. Investigations in this field have mainly concentrated on traditional plasticity models, which might be less applicable to learning through experience. It is vital to investigate the molecular mechanisms underpinning new plasticity models like BTSP.”

The team first aimed to model BTSP in isolated brain tissue, allowing them to accurately observe the neuronal changes that occurred. They successfully triggered BTSP with inputs spaced about 1 second apart, confirming the prolonged time span for information storage. They also discovered that BTSP happens at individual synapses, a crucial factor for the specificity of information encoding. By merging electrophysiological measurements of neuronal activity with specialized microscopy and biosensors, they were able to visualize real-time molecular changes occurring during BTSP to ascertain their functions.

CaMKII: Same player, different role

The research team concentrated on a molecule known as CaMKII, which is recognized for its significant role in various types of plasticity in neurons.

“We suspected that CaMKII would play a crucial role in BTSP. This molecule activates at synapses and can remain active for many seconds. It seemed like the ideal candidate to extend the timeframe of information integration in neurons,” explained Dr. Jain. “As it turns out, we were correct in identifying CaMKII as essential for BTSP, but we were completely mistaken about its function.”

After disrupting CaMKII’s function, the team noticed that BTSP was impaired. Aiming to observe CaMKII activity in neurons throughout the BTSP process, they developed a biosensor that indicates when CaMKII is active. With this newly refined sensor, which had nearly double the sensitivity of earlier versions, the researchers could measure CaMKII activity during BTSP. Surprisingly, they did not find what they had anticipated.

Contrary to their initial expectations, they detected no observable CaMKII activation during the induction of BTSP. Instead, a delayed and random activation of CaMKII was seen many seconds after BTSP was initiated. Furthermore, while the plasticity was occurring at a specific synapse, CaMKII activation took place across a broader area of the neuron. This study demonstrated that CaMKII serves as an instructive signal for BTSP, but does not establish the synaptic specificity of plasticity, suggesting a wider timeframe for synaptic plasticity and a new perspective on how synapse-specific and instructive signals can integrate over several seconds.

“This constitutes a fundamental shift in our understanding of CaMKII functions and the mechanisms of plasticity. The activity of CaMKII throughout the dendrite indicates that it does not determine the synapse specificity of plasticity; instead, it contributes to dendritic information processing. Our findings raise many new questions for further exploration, including what defines the specificity of information coding at individual synapses or the delay observed in CaMKII activation,” stated Dr. Jain. “These unexpected results highlight the necessity of behaviorally relevant models for understanding how the brain encodes information, aiding our ultimate goal of connecting molecular activity to memory formation and addressing neurological disorders related to learning and memory impairment.”