Making Memories

Neuroscience's Holy Grail is the engram - a neuron or set of neurons which physically stores a single memory. To a large extent, that search has ended, thanks to scientists like Dr. Eric Kandel, who won the 2000 Nobel Prize for his discoveries in the neurochemistry of learning.

Dr. Kandel's groundbreaking work incorporated the same classical conditioning techniques Pavlov pioneered with dogs to study neural changes in the California sea snail Aplysia as it learned. Aplysia has only 20,000 neurons, which are the largest in the animal kingdom, visible to the naked eye. This makes Aplysia's neural network easy to manipulate and identify, and makes Aplysia for neurobiologists what fruit flies are for geneticists, and rats are for behavioral psychologists.

Kandel began by stimulating Aplysia’s sensory neurons with electrodes, then used the process of elimination, neuron by neuron, to map out the entire neural circuit which controls gill withdrawal - a simple adaptive, learned response to stimuli from the environment.

Just as you would jerk your hand away after touching a hot stove, Aplysia reflexively withdraws its gills in response to aversive (unpleasant) stimulation. In comparison with yours, its brain is primitive, essentially two neural bundles called ganglia; but just as with more complex animals, it can learn: for Aplysia, just as for humans, "practice makes perfect" - repeating a stimulus converts a short-term memory, lasting minutes, into a long-term one, lasting days, weeks or a lifetime.

In Aplysia, the gill withdrawal reflex is controlled by just 24 sensory neurons which enervate (send signals to) six motor neurons. Between them are "middle managers" - interneurons which act as neuromodulators, a flow control system analogous to a water tap, either excitatory, and increasing the signal flow (i.e. likelihood of neural firing) or inhibitory, and decreasing the signal flow (likelihood of neural firing).

Aplysia's neurons (like yours) are hardwired - physically and functionally fixed by instructions encoded in DNA - so they aren't capable of significant change, such as increasing in number or changing in function or location. However, the connections between them are extremely flexible. This allows for adaptations to their signalling efficiency - the heart of learning.

Neurons communicate via these connections, across tiny gaps called synapses, which excrete chemical messengers called neurotransmitters. Lasting memories are preserved by the growth and maintenance of these synapses. In other words, memories are encoded in the connections between neurons.

Stimulating Aplysia's tail activates interneurons, which excrete the neurotransmitter serotonin. This triggers an excitatory response, and the motor neurons fire, causing muscular contractions. The end result is that the animal withdraws its gill in response to a shock.

Sensitized
Dr. Kandel's early experiments involved training Aplysia in a learned-fear response called sensitization, in which repeating an aversive stimulus makes a creature increasingly sensitive to it. This is the principle underlying Post Traumatic Stress Disorder, in which a post-war veteran has become so sensitized to loud, abrupt sounds (like gunfire) that he jumps at the slamming of a car door. Similarly, an Aplysia snail which has become sensitized to shocks on its siphon (mouth organ) will soon also respond just as strongly when its tail is shocked for the first time. Simply stated, "practice makes perfect": repeated exposure converts short-term memory into long-term memory, via physical growth (DNA-directed protein manufacture).

To find the physical basis for this behavioral change, Dr. Kandel first traced the specific neural circuits which control gill withdrawal; then he devised a technique for creating the simplest possible learning circuit out of living tissue: two live neurons, one sensory, one motor. By growing Aplysia larval neurons in a petri dish, he could substitute a direct squirt of serotonin for shocks, and thus isolate and investigate the specific physical and biochemical processes underlying memory formation.

He discovered two separate chemical sequences, one for building short-term memories, and the other for long-term ones. Short-term memories - which last for seconds to minutes - come from increasing a presynaptic neuron's ability to emit neurotransmitters.

This neurotransmitter increase develops via a six-step chemical sequence called a signalling cascade:
when the snail's tail is given a mild shock, interneurons release serotonin. The serotonin binds to special protein receptors embedded in the membrane of a postsynaptic (recipient) neuron.

These serotonergic (serotonin-activated) receptors trigger the conversion of ATP, the cell's natural "fuel", into a special chemical signal, a secondary messenger called cAMP. cAMP is called a secondary messenger because it conveys an initial external signal from outside the cell membrane to protein molecular machines within the cell.

In Aplysia, the cAMP signal activates a protein kinase (PKA), which functions as a biological "on-off" switch . This kinase migrates back to the membrane and causes a shape change (phosphorylation) in special calcium channels, proteins embedded in the membrane which act as selective chemical gateways. The shape change temporarily opens these channels to calcium ions, which flow into the neuron's synaptic terminals (connective neural end points).

Calcium ions are a chemical signal which triggers increased neurotransmitter release. This temporary increase in neurotransmitter release is the physical basis of a short-term memory.

To convert short-term memory into a long-term memory, the stimulation must be repeated within a short window of time, triggering a second chemical cascade that activates physical growth - DNA-guided protein manufacture.

In a neuron, this growth is manifested as a sprouting of new synapses and/or synaptic branches called dendrites, the tendril-like appendages which snake out to connect with neighboring neurons. The anatomical change is the basis of long term potentiation, so-called because the long term potential for a neuron to fire has been enhanced. The postsynaptic neuron has become more sensitive to stimuli and is thus more likely to fire frequently. There has been a long-term modification of the synaptic connection.

In 2010, MIT's Gertler Lab created a spectacular film of this growth process in living, cultured mouse neurons. Here you can see the growth of neurites, neural buds which sprout into dendrites.

In Aplysia, repeated stimulation increases serotonin release, triggering increased synthesis of the secondary messenger cAMP, more persistently activating the kinase (PKA) switch. Then, just as with short-term memory formation, PKA migrates to the cell membrane, protein channels open, and calcium ions rush into the synaptic terminals, triggering neurotransmitter release.

However, repeating the stimulation past a certain threshold has an additional effect, ultimately prompting physical growth of the neuron, through protein synthesis.

When sufficiently stimulated, a subunit of the kinase detaches and moves into the neuron's cell nucleus, where it activates a specialized protein called a transcription factor (CREB protein or cAMP Responsive Element Binding protein). This transcription factor binds to specific DNA sequences (genes), activating them, so they begin manufacturing new synapses, from protein building blocks like neurexin and neuroligin.

Genes are segments of the DNA molecule arranged in specific sequences to guide the synthesis of messenger RNA (mRNA). That mRNA is a molecule which copies the genetic code from DNA and uses it as a template to guide the manufacture of proteins - the most complex molecules on Earth, which carry out virtually all biological functions, from forming tissues to carrying out chemical processes vital for life.

To accomplish this, mRNA peels off from its parent DNA, then travels outside the cell nucleus to a special region of the cell body - a mazelike series of tubes called the endoplasmic reticulum. Here the mRNA is used by mini protein factories called ribosomes as a blueprint for assembling short building blocks (amino acids) into proteins.

Dr. Robert Singer and colleagues at the Albert Einstein College of Medicine developed a means of filming this neural mRNA synthesis. They attached harmless fluorescent chemical tags to mRNA molecules so they could be filmed within live mouse neurons. Allowing us to see a miraculous film of a memory being formed in real-time: https://www.youtube.com/watch?v=6MCf-6It0Zg

Mammals like mice (and humans) have evolved a special memory-creating and storing brain structure called the hippocampus, named after the Greek word for seahorse, because of its curly shape.

Dr. Singer's team stimulated neurons in the mouse's hippocampus, where episodic and spatial memories are formed and stored (episodic memories are the conscious mental record of life events and the sequences in which they occur, while spatial memories are analogous to navigational roadmaps through an organism's environments).

The hippocampus acts as a sort of amplifying loop, receiving signals from the cortex - the area where conscious thought and ultimate brain control resides - and sending signals back. Memories are thought to be consolidated (stabilized into potentially lifelong memories) by gradual transferal from the hippocampus to the cortex during sleep.

Dr. Singer targeted mRNA which carries the code for a structural protein called beta-actin, central to long-term memory formation. In mammals, beta-actin proteins strengthen synaptic connections by building and altering dendritic spines.

Within 10 to 15 minutes, their neural stimulation trigger beta-actin mRNA release, proving the neural stimulation had triggered transcription of the beta-actin gene. The film shows these fluorescently glowing beta-actin mRNA traveling from neural nuclei to their destinations in the dendrites, where they will be used to synthesize beta-actin protein.

Neural mRNA uses a unique mechanism the Einstein team calls masking and unmasking, which allows beta-actin protein synthesis only when and where it is needed. Because neurons are comparatively long cells, the beta-actin mRNA molecules have to be guided to create beta-actin proteins only in specific regions - at the ends of dendrite spines.

According to Dr. Singer, just after beta-actin mRNA forms in the nuclei of hippocampal neurons, its molecules are packed into granules whose genetic code is inaccessible for protein synthesis. These molecules migrate out to the cells' cytoplasm (inner gel-like cellular fluid).

Stimulating the neuron makes these beta-actin granules fall apart, unmasking the mRNA molecules, and making them available for beta-actin protein synthesis. When the stimulation stops, this protein synthesis shuts off: after synthesizing beta-actin proteins for only a few minutes, the mRNA molecules abruptly repack into granules, returning once again to their default inaccessible mode.

In this way, stimulated neurons activate protein synthesis to form memories, then shut it down. Frequent neural stimulation creates frequent, controlled bursts of messenger RNA, resulting in protein synthesis exactly when and where it's needed to strengthen synapses.

Of course, the process involves more than just a single gene - in fact a dizzingly huge number are involved. Learning involves large clusters of genes within huge numbers of cells.

Florida U. Neuroscience Professor Leonid Moroz has tracked specific gene sequence expression in Aplysia neurons, and estimates that any memory formation event will alter the expression of at least 200-400 genes. This is out of over 10,000 which are active every moment of the simple marine snail's life.

Moroz's team zeroed in on genes associated with Aplysia's feeding and defensive reflexes, and found over 100 genes similar to those linked to every major human neurological illness, and over 600 similar genes which control development. This shows these genes emerged in a common ancestor and have remained nearly unchanged for over half a billion years of both human and sea slug evolution.

Human brains contain about one hundred billion neurons, each expressing over 18,000 genes, at varying levels of expression (protein synthesis rates), with over 100 trillion connections between them. Aplysia has a comparatively much simpler nervous system, with only 10,000 easily identifiable, large neurons. However, it is still capable of learning, and its neurons communicate using the same general principles as human neurons.

According to Dr. Moroz, if genes use a chemical alphabet, there is a kind of molecular "grammar", or set of rules controlling the coordinated activity of gene expression across multiple neural genes. To understand memory or neurological diseases at a cellular level he says, scientists need to learn these grammatical rules.

More sophisticated
Like the Einstein team, Dr. Eric Kandel also went on to study the mouse hippocampus, and found it uses a chemical cascade much like Aplysia's. But in mouse - and human - hippocampi, the chemical sequence for memory formation is slightly different.

In the hippocampus the changes which lead to memory formation primarily occur in the receiving (postsynaptic) neuron rather than the sending (presynaptic) neuron. This process occurs in two stages:

During Early LTP, the excitatory neurotransmitter glutamate is released. Glutamate acts upon two types of receptors, NMDA (N-methyl-D-aspartate) and AMPA (a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptors. Normally, only the AMPA receptors respond, but repeated stimulation activates the NMDA receptors as well. These open ion channels in the membrane, allowing calcium ions to flow into the neurons, and this activates a (different) second-messenger kinase, CAMK2A (calcium calmodulin kinase). CAMK2A triggers a chemical cascade that results in the growth of additional AMPA receptors.

In Late LTP, repeated stimulation activates a second system, which releases the modulatory neurotransmitter dopamine. Like serotonin in Aplysia, dopamine acts upon the hippocampus as a sort of volume switch, dialing up neural activity. (Disruptions to the dopamine system are at the heart of disorders like schizophrenia and Parkinson's disease.)

Dopamine binds to specific receptors on the cell membrane, increasing production of cAMP, activating the same PKA-CREB sequence responsible for starting protein synthesis for new dendrite/synapse formation.

But mice are capable of much more sophisticated memory feats than sea slugs, such as memorizing spatial layouts in a manner much like humans. In his experiments with mice, Dr. Kandel found three principles at work in learning. First, when two or more neurons converge to stimulate a third neuron, it acts as a logic circuit known as a coincidence detector. This circuit underlies associative learning - where a creature links object or event A with B, cause with effect, touching a hot stove with pain, etc. Variations of this circuit underlie many forms of complex mental computation.

Dr. Kandel also discovered that mice hippocampi use place cells, special pyramidal neurons which fire in response to specific locations. These place cells act as mental markers, firing at the sight of certain landmarks, allowing mice to encode internal navigational maps.

The permanency of this spatial memory varies based upon how much attention is applied during  navigation, with these levels of attention appearing to correspond to levels of the neurotransmitter dopamine. In other words, dopamine modulates the effects of learning - acting like a volume knob that controls how much data is converted from short-term into long-term memory. This means that,
generally speaking, the more attention (and dopamine) that is applied, the more effectively a route will be memorized.

Dr. Kandel went on to study motor learning, concentrating his search in the cerebellum (latin for the "Little Brain"), known to be the master coordinator of complex movements.

All animals practice complex sets of coordinated movements until they perfect them, from learning to drive a car (for humans) to catching flies with the tongue (for frogs). Their first attempts may be unsuccessful, but with perseverance, eventually they master each skill, by creating procedural memories, a form of motor learning central to animal behavior.

Just as with episodic and spatial memory, procedural memory is dependent upon changes in synaptic strength - with repeated practice, there is growth (of new dendrites and synapses) in a superior medial (midline) strip atop the brain's outer surface, the motor cortex, which plans and issues commands for voluntary movements - sequences of muscle contractions.

Coordinating the process, the cerebellum sits behind the brainstem at the top of the spine. It acts as a kind of coach, comparing actual physical performance - the changing positions of limbs, trunk and head in space - to intentions. Through this feed forward mechanism, the cerebellum corrects motion, ensuring smooth timing and orchestration of the many muscle contraction signals needed to engage in complex movements. Learning to orchestrate these movements depends to a large extent upon special Purkinje cells.

Neurons fire when special protein channels embedded in the cell membranes open, allowing electrically-charged ions to flow inward. Among these, the HCN1 ion channel (Hyperpolarization-activated, Cyclic Nucleotide-regulated nonselective cation), is of central importance to motor learning in cerebellar Purkinje cells.

To study HCN1 ion channels during motor learning, Dr. Kandel and his team bred mutant mice with neurons that lacked HCN1 channels in varying brain regions.

They then ran their mutant mice (along with a control group of normal mice for comparison) through a series of complex motor tests, which included swimming through water mazes and balancing on rods. These tests required complex, repetitive and coordinated motor output. The mice were also conditioned with simpler motor behaviors like eye blinking.

Mice with no HCN1 channels in their Purkinje cells could still perform simple movements like eye blinking, but had extreme difficulty in performing complex behaviors like swimming and balancing. In contrast, mice without HCN1 channels in their forebrain but with normal cerebella had no problem in performing complex behaviors, demonstrating that HCN1 channels specifically within Purkinje cells are the key to complex motor learning.

When currents were applied to Purkinje cells, those without HCN1 channels took longer to return to normal firing activity levels than normal Purkinje cells. This indicates that HCN1 channels function as stabilizing units for Purkinje cells, allowing them to quickly recover from activation and return to normal functioning. During complex, repetitive behaviors, Purkinje cells receive repetitive bursts of input, and need to recover quickly to form the feed forward function that guides motor activity.

In the end, while it is good to understand the principles by which your brain manufactures memory, here is some more specifically practical advice on how to make the most of what you've got, based on the latest scientific studies:

1. Sleep at least seven to eight hours at a regular time every night

2. Take fish oil supplements

3. Exercise every day

4. Limit or eliminate your intake of alcohol, tobacco, and fatty or sugary foods

5. Pay full attention to things you wish to remember

6. Space your learning out; cramming does not help long-term memory retention, especially all-night study sessions.

7. Think of how the information you want to learn specifically relates to you or to things you already know

8. Study a foreign language and a musical instrument

9. Study in the same environment, at the same time every day

10. Keep your study environment free from clutter and distractions

11. Physically stand up and take a break from studying to do something different about every 15 minutes before returning

12. Spend time in mentally stimulating environments where you have new experiences and interact with many different sorts of people from many cultures and backgrounds

13. Reduce your life stresses to a minimum