COURSES:
TOPIC: Learning & Memory
LECTURER: Dr. Andrew Bendrups, School of Human Biosciences
Definition of learning and memory
The terms "learning and memory" are so often used in everyday language that their meaning is taken for granted. However, in science these words are often used in ways that are different to everyday use (eg. we speak of memory in the immune system!) so it is worth defining them quite precisely.
According to Steven Rose* (a memory researcher):
Needless to say, humans are animals too, so this definition is meant to apply to them as well. However, Rose's definition is both complex and narrow because it reflects his research on non-human animals in which visible change in behaviour is the only possible evidence of learning.
My simpler version follows:
To use an example which fits well with Rose's definition: a young child extends his fingers towards a candle flame, which he has never seen before (the novel situation). As a response to burning his fingers, the child subsequently avoids touching any flame (a comparable situation). This change in behaviour which may well persist until death. In other words, a long-lasting memory has been formed. The new behaviour is clearly appropriate to the well-being of the child.
This is an example of what has been called "one-trial avoidance learning" which is a rather extreme form of learning &emdash; more often it takes many repetitions of a similar experience to learn reliably.
A more subtle example is the ability to remember a person's face, which may take more than one experience. Whether this is "appropriate" is less obvious than in the previous example. Also, a change in behaviour may not be obvious, as there may be no outwardly visible sign of the memory: the response may involve no more than thinking "I've seen that person before" (in my definition, this would be covered by the term "acquired knowledge').
So, learning is a process in which, as a result of experience, a memory is formed. Memory is the storage of the result of learning. A person's behaviour is reliably changed as a result of learning, and is generally "more appropriate" in that person's subsequent interaction with the world.
Neural Plasticity
Plasticity means lasting change, a property of plastic materials (eg. plasticine), which can be moulded into new shapes.
Since behaviour is changed as a result of learning, there has to be some change in the biological organ which controls behaviour (ie. the brain). This change is generally called "neural plasticity" or "synaptic plasticity".
It is worth noting that neural plasticity can occur as a result of things other than normal learning (eg. changes occurring after brain damage), but here we are only considering normal learning. The biological basis of learning and memory is really all about neural plasticity.
Many scientists now believe that relatively simple changes in the connections between neurones are the basis of how we learn and remember things like faces and life's daily events, or skills like learning to drive a car. Before we can begin to understand how this could be, we need to look at much simpler types of learning of the sort that can easily be studied in animals.
Most of our understanding of the biology of memory is based on research in animals such as rats, slugs and even fruit flies. There seems to be an amazing consistency in the way learning works, across a wide range of animal species.
There are obvious differences in the learning capabilities of apes and flies but the general principles appear to be similar at the level of the cell. We assume that the difference in capability lies in the size and complexity of the brain, rather than in the fundamental properties of the cells that make up the brain. An analogy could be made with the difference between a $5 calculator and the $5000 computer; the basic electronic components are the same.
Non-associative learning
Perhaps the simplest sort of learning is that which occurs by simple repetition. A good example of this is habituation, in which an animal stops responding to a repeated, non-threatening stimulus. In the sea-slug Aplysia, Eric Kandel and others showed that habituation in a reflex response was due to the progressive weakening of a synapse in the reflex arc. They found that changes in intracellular biochemistry were responsible for reducing the amount of neurotransmitter released by the nerve terminals, therefore reducing the size of the response to the stimulus.
Non-associative learning can also involve strengthening of synapses, leading to a larger response. This has been shown at the cellular level in many parts of the brain, especially those structures known to be involved in learning and memory (eg. hippocampus, cerebellum, cerebral cortex). A strengthening of synapses by repetition alone could be the basis of simple learning by repetition ("rote" learning) in humans.
Associative learning
A pioneer in the field of animal learning, Ivan Pavlov found that if he rang a bell a few seconds before presenting meat to his dogs, after many repetitions the dogs began to salivate just on hearing the bell. In the jargon of this field, the meat is called the unconditioned stimulus (UCS) and the bell the conditioned stimulus (CS), meaning that the bell stimulus initiates a learning process (conditioning). This is a more complex type of learning than non-associative learning since it only occurs when two stimuli occur very close together (are associated) in time. If the two stimuli are not associated, no learning occurs (other than habituation to the ringing of the bell). It could be described as "cause and effect" learning: in Pavlov's example the dogs are subconsciously learning that the bell signals the arrival of food.
Pavlov's finding implied either that a new neural pathway from the hearing system to the salivary control system had been forged, or that a pre-existing but non-functional connection (or near-connection) had become functional. The second alternative is the most likely: it is difficult to conceive of a mechanism whereby a new connection over a distance of some centimetres could be formed as a result of simple pairing of two stimuli.
An influential psychologist, Donald Hebb, proposed a hypothetical way in which such types of learning (associative learning) could occur at the level of the cell (neurone). Hebb proposed (to paraphrase his words using modern jargon) that if a neurone A took part in firing another neurone B, then a plastic change would occur in the synapse between neurone A and B, such that the connection between A and B would be strengthened. We can assume that neurone B has other connections which make it fire, and that neurone A initially produces only a weak or no response. The key point is that neurone A has to be active ("trying to fire neurone B") at the same time that neurone B is undergoing action potentials (ie. "firing")
This model in principle explained conditioning and other sorts of associative learning (eg. learning by reward or punishment), where two sorts of synaptic input might interact. Memory is stored as a change in synaptic strength (or synaptic "weight").
Much later, Eric Kandel and others studied the cellular mechanism of a type of associative learning called "activity-dependent facilitation", involving changes over time in the strength of a withdrawal reflex. This was done in the sea slug Aplysia, because its simple nervous system was amenable to biochemical study. As with non-associative learning in this animal, the ultimate mechanism proved to be biochemical changes in the nerve terminals of the slug.
It is generally assumed that similar sorts of biochemical changes can explain the phenomenon of short-term memory in humans.
The processes which enable lasting long-term changes in neuronal connections (changes which must underlie the long-term storage of memory over years) are less well understood. However, they are known to be dependent on protein synthesis: this is thought to indicate that the processes are more permanent than biochemical and probably involve a morphological (structural) change, such as the formation of new synapses.
* See Rose, S. (1993). The making of memory. Bantam Books, London, p136