Hippocampus – the neurons involved in Pavlovian learning become more synchronized when a memory is being formed


The phrase “Pavlov’s dogs” has long evoked images of bells, food and salivating dogs.

Even though this tried-and-true model of repetitive patterns mimics a variety of learning processes, what happens on a cellular level in the brain isn’t clear.

Researchers at the University of New Hampshire took a closer look at the hippocampus, the part of the brain critical for long-term memory formation, and found that the neurons involved in so-called Pavlovian learning shift their behavior during the process and become more synchronized when a memory is being formed – a finding that helps better understand memory mechanisms and provides clues for the development of future therapies for memory-related diseases like dementia, autism and post-traumatic stress disorder (PTSD).

“There are tens of millions of neurons in the hippocampus but only a small fraction of them are involved in this learning process” said Xuanmao (Mao) Chen, assistant professor of neurobiology.

“Before engaging in Pavlovian conditioning, these neurons are highly active, almost chaotic, without much coordination with each other, but during memory formation they change their pattern from random to synchronized, likely forging new connecting circuits in the brain to bridge two unrelated events.

In the study, recently published in The FASEB Journal, researchers looked at Pavlovian learning patterns, or respondent conditioning, in mice.

In the beginning, before any repetitive learning exercises, the mice did not know what to expect and using special imaging with an endomicroscope the researchers saw that the neural activity was disorderly.

But after repeating different tasks associated with a conditional stimulus, like a tone or bell, the mice began to recognize the pattern and the highly active neurons became more synchronized.

The researchers hypothesize that without forming synchronization, animals cannot form or retrieve this type of memory.

This shows hippocampal cells

On the left is an enlarged image showing many hippocampal neurons, most of which are silent and only a few are active.

On the right are close ups of three highly active neurons, or memory cells, which become synchronized after memory formation. Image is credited to UNH.

In the 1890’s, Russian psychologist, Ivan Pavlov discovered classical conditioning through repetitive patterns of bell ringing which signaled to his dogs that food was on its way and stimulated salivation.

This same learned behavior is important for episodic knowledge which is the basis for such things as learning vocabulary, textbook knowledge, and memorizing account passwords. Abnormal learning processing and memory formation are associated with a number of diseases like dementia, autism, and PTSD.

People who struggle with these cognitive dysfunction-related disorders may have trouble retaining memories or can even form too strong a memory, as with PTSD patients.

The UNH researchers believe that understanding the fundamentals of how classical conditioning shape neural connections in the brain could speed up the development of treatments for these disorders in the future.

Contributing to these findings are Yuxin Zhou, doctoral candidate; Liyan Qiu, research scientist; both at UNH, and Haiying Wang, assistant professor at the University of Connecticut.

Funding: This work was supported by the National Institutes of Health (NIH) and the Cole Neuroscience and Behavioral Faculty Research Awards.

Memory Engram Allocation in the Brain

Memory formation and storage relies on structural changes that occur in the connectivity between neurons. Richard Semon, an early advocate of a physical theory of memory, was the first to term the neural substrate containing memories as the memory engram(Semon, 1904; Schacter, 2001).

He defined the engram as the lasting modification produced by experience and stimulation in the brain. Attempts in the 20th century to find and identify the engram were intense.

Karl Lashley is most famous for performing a long series of lesion experiments in the cerebral cortex of rats in an attempt to find associations between carefully targeted lesions and the ability of animals to solve maze tasks in which they were trained.

While the lesions did cause memory impairments, Lashley’s studies showed that impairments occurred irrespective of the location of the lesion, leading him to conclude that the memory engram is not localized, but it is spread broadly and indiscriminately throughout the brain (Lashley, 1950).

Other experiments identified specific loci of stimulation that evoke memories, such as the experiments of Penfield and Rasmussen who applied electrical stimulation to epileptic patients in order to identify the centers of seizures.

They found that electrical stimulation in specific brain regions could cause a vivid recall of various memories (Penfield and Rasmussen, 1950).

Another historically significant case is the famous patient Henry Molaison whose bilateral removal of the medial temporal cortex led to anterograde amnesia. This bolstered the idea that episodic memories may be processed in the hippocampus (Scoville and Milner, 1957) before being consolidated.

These earlier studies hinted at the existence of cellular memory engrams but lacked the tools to visualize or manipulate them. Thus, a causal link between memories and the neural substrate that contains them remained elusive for decades.

It was not until a few years ago that optogenetics, molecular labeling methods and 2-photon imaging allowed scientists to precisely identify the neurons which are involved in the learning and recall of specific memories (reviewed in Rogerson et al., 2014; Josselyn et al., 2015; Tonegawa et al., 2015; Poo et al., 2016).

The first study that revealed the cellular memory engram, i.e., the existence and identification of populations of neurons that are sufficient and necessary for learning was (Han et al., 2007). Since then, numerous studies have characterized the properties and even manipulated cellular memory engram in various ways.

These manipulations include the ligand- and light-driven neuronal activation of neurons, pharmacological activators/suppressors of plasticity, combined with a variety of imaging techniques for the identification and exploration of the properties of cellular populations engaged in the long-term storage of memories (Guzowski et al., 1999; Zhang and Linden, 2003; Han et al., 2007; Reijmers et al., 2007; Silva et al., 2009; Bergstrom et al., 2011; Liu et al., 2012; Mayford, 2014; Lai et al., 2018).

The formation of memory engrams is associated with the action of multiple mechanisms that alter the functional properties of neuronal circuits during learning processes, which are collectively grouped as plasticity phenomena and act on multiple spatial and temporal scales (Bhalla, 2014).

On the neuronal circuit level, plasticity is associated with changes in the intrinsic excitability of neurons (Zhou et al., 2009) and/or the excitability of their dendritic domains (Zhang and Linden, 2003; Losonczy et al., 2008), as well as on homeostatic phenomena that shape neuronal responses (Turrigiano, 2008).

In addition, memory engrams are subject to the effects of epigenetic mechanisms such as the contribution of DNA methylation to memory storage (Day and Sweatt, 2011) and histone acetylation to memory maintenance and reconsolidation (Gräff et al., 2014).

However, the main (and best studied) mechanism viawhich memories are believed to be encoded in these neuronal populations is the plasticity of synaptic strengths, which occurs primarily within the dendritic regions of excitatory neurons.

At the level of the synapse, plasticity is expressed both presynaptically, and post-synaptically (Larkman et al., 1992), and even affects the local excitability properties of dendrites (Sjöström et al., 2008).

Thus, to understand memory engram formation we need to understand how sub-cellular, primarily synaptic, modifications occur as a consequence of learning, and how these modifications relate to cellular engrams.

It is important, first, to identify patterns of spatial allocation of synapses within dendrites during learning.

University of New Hampshire


Please enter your comment!
Please enter your name here

Questo sito usa Akismet per ridurre lo spam. Scopri come i tuoi dati vengono elaborati.