Learning


Table of Contents


The learning process results in a relatively permanent change in the behavioural potential of the individual. There are various theories of learning.

Classical Conditioning

This is also known as Respondent or Pavlovian Conditioning after the Russian Scientist Pavlov.

In this form of learning, the repeated pairing of a new stimulus (for example the ringing of a bell or a buzzer) with an unconditioned stimulus, for example, food which is known to elicit a given, unconditioned response (e.g. Salivation) leads to learning that the sound of a bell will announce the introduction of food.

The new stimulus is the conditioned stimulus (sound of a bell) that will elicit the response (salivation) which is renamed the conditioned stimulus, partly because it is usually diminished compared with the original if the food is not presented when the bell rings repeatedly.

So, the unconditioned stimulus (e.g., food) naturally elicit an unconditioned response (salivation)

when the unconditioned stimulus is paired with a conditioned stimulus (bell) we get the same unconditioned response (salivation).

If food is withdrawn, the mere sound of the bell (conditioned stimulus) will lead to a conditioned response (salivation)

In conditional learning, the subject is passive, and the responses are typically autonomic or emotional. This type of learning essentially involves the prediction of the event by the subject.

The conditioned stimulus should be simultaneous with or just anticipate the unconditioned stimulus (i.e., The bell rings at the same time the food is presented or just before it appears)

The unconditioned response and the conditioned response are essentially the same (salivation).

The conditioned stimulus can have an effect only if the stimulus-response link already exists.

Subjects are more likely to acquire some conditioned stimuli than others. For example, in Garcia’s rats’ experiments, nausea link with stimuli was more easily established than an electric shock. This might be due to cognitive evaluation by the subject or biological preparedness.

Extinction

if the conditioned stimulus is presented repeatedly without the unconditioned stimulus, it becomes weaker and disappears. The conditioned Stimulus-Response link can be recovered if we repeat pairing with the unconditioned stimulus. This can also occur if we allow a respite period without presenting the conditioned stimulus (spontaneous recovery)

Stimulus generalisation

A similar but not identical stimulus to the conditioned stimulus may also elicit a conditioned response even though it was never paired with the conditioned stimulus. For example, different bell sounds can elicit the same conditioned response. However, the conditioned response diminishes proportionally according to the extent the new conditioned stimulus differs from the original stimulus. This includes semantic generalisation, thus a conditioned stimulus like the word “sea” can be replaced by the word “wave” and we still obtain the same conditioned response.

Higher order condition

A new conditioned stimulus is paired with the first conditioned stimulus and both act similarly as the unconditioned stimulus. Thus, words may have similar connotations and induce the same emotional response.

Clinical relevance

Conditional learning is the mechanism whereby emotional and attitudinal responses become linked to new stimuli. This is the model for the acquisition of phobias. Watson and Rayner (1920) produced a new fear of a white rat in baby Albert by banging an iron bar behind him each time the rat presented. This Rabbit Phobia was treated in another child by Jones (1924) later by feeding him during repeated presentations of rabbit at progressively shorter distances.

In real life, findings in humans must include also the consequences of Incubation and Preparedness.

Incubation is the increase in strength of emotional conditioned response as a consequence of repeated brief exposure to a conditioned stimulus. Thus repeated feeling of nausea following a certain type of food will make that food or similar dishes abhorrent. Preparedness means that some stimuli are more likely to become good conditioned stimuli than others

Conditional learning is a necessary conceptual framework for behaviour therapy, in particular concepts of exposure, desensitisation stimulus fading etc.

Operant (instrumental) conditioning:

Operant conditioning is based on Thorndike’s law of effect, which dictates that successful behaviour will be repeated. It is mainly concerned with the establishment of new stimulus-response links.

Subjects learn to do something under certain stimulus conditions by being reinforced according to the consequences. Repetition of behaviour increases the likelihood of its recurrence (habit strength). The subject is active, likely to be using consciously controlled behaviours. Reinforcement alters the probability of a response being made. Timing is crucial: short-term interests can thus outweigh long-term.

Primary reinforcement is the natural reinforcement through the decrease of the assumed basic drive (food, drink, sex etc.). Secondary reinforcement is reinforcing properties derived from association with primary reinforcement (money, tokens etc.)

Reinforcement schedules:

Intermittent reinforcement takes longer to establish learning but learned response is less likely to extinguish. Reinforcement can be continuous (on every occasion), or fixed ratio (e.g. reinforce one in 3 correct responses) or fixed interval (e.g. reinforce after every 10 seconds of continuous responses). Alternatively, reinforcement can be variable ratio or variable interval. The variable-ratio reinforcement type is the most resistant to Extinction (like the case in gambling). The cognitive appraisal may alter apparent or observed reinforcement contingencies. Reinforcers and associated responses may compete with each other in real life, so the subject acts amid an array of stimulus conditions with a repertoire of responses for which there are several more or less powerful reinforcers, primary and secondary.

The reward is a consequence of a response that is assumed to be pleasurable but may not actually be a reinforcer.

Negative reinforcement:

Negative reinforcement can mean two things:

Reinforcement through withdrawal of unpleasant conditions. The punishment is an aversive consequence of a response that suppresses the response. Negative reinforcement is most effective if it is immediate and of sufficient intensity to suppress the response on the first occasion rather than starting with low intensity and building up over repeated trials. It is referred to the subject’s conscious understanding i.e. makes sense to them (Different reinforcement of Other behaviour (DRO).

Escape conditioning is a special variety of negative reinforcement. The response learned provides a complete escape from an unpleasant situation (rather than an alteration in it). Responses which are acquired in this way are very resistant to extinction. In avoidance conditioning, the response learned prevents an unpleasant situation occurs. Extinction is the cessation of a learned response when reinforcement is no longer provided.

Clinical relevance:

Operant conditioning is the basis of much behaviour modification, especially with children and the mentally handicapped but also in social skills groups and ward management schemes such as token economies.

Shaping is the reinforcement of successive approximations to desired and effective behaviour. This occurs when a complete response is complex. With humans it is used in teaching and often accompanied by instruction, prompting and encouragement. Chaining: is breaking a complex behaviour into a sequence of steps, establish learning of each step, then learn the chain by linking steps forward or backwards. In backward chaining, the satisfaction of achieving the desired final links in the chain (or the increasing drive as the final accomplishment and goal of the entire behaviour is approached ) reinforces the learning of successively earlier links.

Imitation learning (modelling)

Both classical and operant learning can be accomplished vicariously without effort on the part of either the model or the learner:

Observational learning:

This type of learning occurs through the demonstration of skills. Disinhibition or inhibition of behaviour is the basis of learning e.g., consequences (or not) of action promotes behaviour.

Observational learning works in social facilitation like infectious laughter, mass hysteria, and crowd behaviour. It is also the way of setting standards and institutional norms.

The model of observational learning has to be of high status, high competence, and high social power. It will be most influential if it has some characteristics in common with the observer. Low self-esteem and low self-confidence in the observer will enhance their perception and valuation of these characteristics.

Vicarious Learning:

This form of learning occurs with reinforcement through observation of rewards to model. Observational learning increases the chances of the behaviour being manifest, i.e. the learning is not necessarily immediately revealed in behaviour or performance.

There are five important functions in observational learning :

Attention to relevant aspects of the model’s behaviour The visual image of the model. 3.Remembering or rehearsal of behaviour.

Refinement by reproduction of learned behaviour. Anticipation of consequences Factual transmission This is a neglected aspect of learning

Sign learning theory

To explain how familiarity with a maze facilitates learning how to run it. Formation of cognitive maps which are expectations about what will happen next.

Insight learning:

Rapid restructuring of perceptual field or concept to derive sudden insight into a problem. Learning of a cognitive relationship between means and end.

Social learning theory:

A somewhat vague term originally applied to attempts to integrate psychoanalysis and learning theory. Now associated with Bandura and integration of this basic learning theory with cognitive factors and observational learning.

Conceptualises people as active, thinking, problem solvers who learn by a variety of mechanisms and whose learning is affected by such factors as cognitive appraisal, inference, goal seeking, affiliation, striving for meaning etc.

Assumptions that classical and operant conditioning operate but cannot explain the appearance of new behaviours.

Memory

Memory is the retention of learned associations and stored information, skills

It is intertwined with learning. We need memory to assess whether learning has occurred. Learning steps involves the recall of information and recognition of existing sensory stimuli. The components of learning increase in difficulty from Recall to Recognition and finally Relearning. It is unclear whether these are different processes or a difference in the strength of memorisation

Traditionally, memory has three aspects:

1- Registration (which is associated with selective attention )

2- Storage

3- Retrieval

Storage is also classically assumed to have three components:

Sense Organ Memory:

This allows comparison of stimulus with Long Term Memory to assign significance. The stimulus can be echoic or iconic according to sensory modality. What is compared is mainly the physical features of the stimulus. This component is revealed in for example. dichotic listening tests. The Sense organ memory is ephemeral as it fades and is lost in about 0.5 secs. Perhaps, it is another way of talking about registration as it is a very limited form of storage indeed.

Short-Term (primary or working) Memory (STM):

Items entering STM will be lost in about 30 secs unless rehearsal or repetition is used. Typically, but not exclusively, verbal or acoustic stimuli are the prevailing coding modality. The subject is conscious of the store contents. It has a capacity of about 7 items (bits of information). This can be increased by chunking of information to allow one entry to cover several items. Chunking is performed by imposing meaning or rule on the information. It is increased by using several subsystems: recent auditory, visual input, with speech and motor output. This allows storage of information for long enough to be used. Clinicians use of term “short-term memory “ when they mean memory which lasts for about 20 mins). During the transfer of selected STM contents to LTM, the remainder is lost.

Long-Term (secondary) Memory (LTM):

This is the storehouse where information is stored relatively permanently. We are usually not conscious of the store though may know what is stored in there. Theoretically, it has unlimited capacity but there may be limitations on retrieval, i.e. items may be “available” but not “accessible”. Long-Term Memory requires consolidation. Some loss occurs by forgetting but this is slow. The coding of memory is mainly visual, semantic, acoustic (probably also olfactory).

The organisation of LTM is under study. The information stored systematically irrespective of presentation. Lexical memory (own stored vocabulary or internal lexicon) has structure but uncertain whether hierarchy or matrix.

We need to differentiate between Episodic memory (events) from semantic memory (facts). This also includes memories for visuospatial knowledge and motor skills.

Clinical relevance of memory is evident in amnesic syndrome (anterograde amnesia):

There are two possibilities:

1- Inability to transfer information from STM to LTM, so there is no new material in LTM though other information can be retrieved from LTM. We can show that STM is intact by digits recall.

2-It may be a retrieval deficit rather than an encoding problem. Pooled patients in most studies, such as head injuries, alcoholic Wernicke-Korsakoff’s, nutritional deficiencies show the same difficulty in the retrieval of information.

There is a controversy as to whether there is a qualitative difference between STM and LTM.

An alternative model is the level of processing approach, as it is evident that durability of memory depends upon the depth of rehearsal of material. Processing through more than one subsystem will enhance the consolidation of memory.

For example:

Memory using sight only is less solid than using sight and sound or sound, sight and meaning. Deeper processing produces longer-term retention. A variant of the level of processing is the depth of analysis (that is the amount of meaningfulness extracted from the stimulus). Semantic encodings are associated with better recollection. But as retrieval is the only indicator of the depth of processing, circular argument results.

Consolidation:

Once information is stored in LTM, it must be left undisturbed for a few minutes. This may be speeded by caffeine. A major disruption (ECT or head injury) induces retrograde amnesia.

State-dependent learning: Recall increases in the setting where learning took place. It may be verbal context, visual setting or possibly drug state.

Retrieval:

Retrieval is the pulling of information out of storage from LTM to STM. The recall is one component but it is not everything. For example, in the case of the “tip-of-tongue” phenomenon, we know that something is there but we can’t specify it, though we may have a loose idea of its organizing characteristic (e.g. “it starts with T”). Other components are knowing that one knows and knowing that it is close to recall, plus knowing that what is recalled is right. Recall appears to be organized according to applied strategies such as semantic clustering. Recognition alone indicates storage but it is incomplete retrieval, in fact, it bypasses retrieval.

In learning word lists, mnemonic devices such as forming new associations to each word appear to enhance learning and retrieval.

Forgetting :

It is the process of rapid loss of most recently acquired material initially (the forgetting curve has a sharp initial gradient). It is explained by two hypotheses:

1- Interference theory:

Forgetting is determined by activity between learning and recall. New information learned in the interim period impair recall (and therefore presumably consolidation) i.e., retroactive inhibition. For example, a subject learns subject A and then learn subject B then tries to recall subject A will have a moderate score, while if he learns subject A then recall subject A only, his recall score will be much higher. Previous learning is likely to impair rather than facilitate subsequent learning; this is called proactive inhibition. On the other hand, if a subject learns A then B then C his score on recall will be A more than B and still more than C, while if he just learns C his score on C will be higher than in the previous case. Obviously, learning skills can affect this but in simple tasks such as word lists, proactive inhibition is easily demonstrated.

Primacy Effect : Things that are learnt first are retained better in the LTM.

Recency effect:

Last words learned are remembered better if tested immediately after the presentation because they are retained in STM.

Decay theory:

Memories fade with time. It is impossible to demonstrate this but it has an intuitive appeal. There is some suggestion that it is a problem of retrieval not decay, as some skills, like a foreign language, can be recovered under hypnosis or in delirium. Benzodiazepines if taken after learning a word list improves its subsequent recall, perhaps by partly suppressing registration of new information.

Repression as motivated forgetting is difficult to demonstrate but intuitively plausible.

Reconstructive memory:

It is possible to demonstrate that eye-witness accounts are distorted by bias questioning. Serial reproductions of narrative show evidence of shortening and seeking a more coherent account with the omission of detail and idiosyncratic aspects (episodic memory shows effort after meaning).

Perception

Perception is the interpretation of sensory stimuli. It is an active process that creates a clear and meaningful world.

Selection:

Raw sensory data are not available to consciousness. Sense organs reduce raw data by the selection of the data available to them (e.g., only visible light within the electromagnetic spectrum will be perceived)

Further processing by e.g. habituation at the sense organ level reduces central input. Central perceptual mechanisms assemble sensations to create a mental representation of the world, discarding even more (95% or more) of data in the process.

Principles of Gestalt school:-

Perception is preferentially a perception of organized wholes. We perceive the “Good” form (high redundancy). we can predict the whole figure from fragment). The whole is greater than the sum of its parts.

Law of closure:

an incomplete circle is seen as a whole circle.

Law of continuity:

A sequence of dots is seen as a line

Law of proximity:

Adjacent items are grouped together.

Law of similarity:

like items are grouped together.

Figure ground differentiation:

We may focus on the ground if it is uniform, or focus on the figure if it is particular or familiar

Depth perception:

Depth perception use of multiple cues:

1- Binocular vision: disparity of retinal images produce depth.

2- Binocular convergence

3- Monocular accommodation

4- Interposition of objects

5- Texture gradient

6- Linear perspective

7- Elevation (distant objects seems higher)

8- Relative size

9- Relative brightness: brighter objects seems nearer.

10- Aerial perspective: Blue haze gives the impression of distance

11- Motion Parallax

Perceptual Constancy: –

Experience of a consistent world despite incomplete, ambiguous, confusing sensory information :

Size Constancy:

An approaching object produces an increasingly large retinal image but does not appear to grow. Objects retain their size regardless of distance (constancy scaling).

Shape constancy :

e.g. window perceived as a rectangle despite non-rectangular and varying retinal image.

Back to Top