Digging Deep: Levels of Processing Memories

When considering LTM you have to ask, how does the information get there? What processing allows you to remember information for a long time?

To help understand memory processes, Fergus Craik (a British cognitive psychologist) and Robert Lockhart (a Canadian psychologist) devised a novel approach in 1972. Most research up to that point had examined memory as a series of stages or components without looking at how things moved between stages. Craik and Lockhart specifically did and proposed the levels of processing framework of memory elaboration.

We call it a framework rather than a theory, because it’s a descriptive model and doesn’t make predictions about what exactly deep processing is. Nevertheless, it’s a beautifully simple model that has stood the test of time and can be used to help you learn things.

With the levels of processing framework, the deeper people process something, the more likely they are to remember it. Deep processing is also known as elaborative processing, because it involves thinking around to-be-remembered information (in contrast to shallow, surface-level processing).

To highlight the difference, Craik and Lockhart asked participants to learn a series of words in three different ways:

Structural coding: State whether the word was in italics or not.
Phonetic coding: Consider whether it rhymed with another word or not.
Semantic coding: Evaluate whether it can fit into a sentence or not.

They then gave the participants a memory test and found that people remembered the words processed semantically better than those processed phonetically and structurally. These results suggest that participants engaged in elaborative rehearsal when processing the word in a semantic manner, compared to maintenance rehearsal when a word is simply repeated.

Later, researchers found that when you process things in a personally meaningful way (you relate it to yourself, such as remembering a topic in a class because it’s something you’ve experienced yourself), you’re more likely to remember them than if they’re processed semantically, suggesting a scale representing the depth of processing (check out Figure 9-1).

© John Wiley & Sons, Inc.

Figure 9-1: Various depths of processing.

If you’re trying to remember something, process it as deeply as possible. Instead of simply repeating it, make it semantically meaningful or personally relevant to you. This technique helps when you’re trying to recall it.

Evidence suggests that these different depths of processing are processed by different parts of the brain. Shallow processing is often highly perceptual and is linked with brain areas associated with vision (the occipital lobe at the back of the head). Phonetic processing correlates with activity in the auditory cortex. Deeper, semantic processing involves more brain areas, including in the frontal cortex and the temporal lobes. Even more brain areas are active when you make something personally relevant (such as, the bilateral caudate in the basal ganglia – a brain area right in the middle).

Classifying Long-Term Memories

Psychologists commonly distinguish between the following two types of long-term memories:

Declarative (or explicit) memory: Memory that you can declare or talk about. It’s conscious memory of knowledge about yourself, your own life or all facts you’ve learnt over your life.
Non-declarative (or implicit) memory: All other forms of memory that you can have. You can’t talk about it (for example, describing how you kick a ball is difficult) and so it’s considered unconscious.

‘Let me tell you all about it!’ Declarative memory

When psychologists talk about memory, they normally mean declarative memory, which they split into two types:

Episodic memory: When someone asks you a question about something you’ve done
Semantic memory: When someone asks you a fact about the world

In both cases, you’re aware that you know something (even if you forget from time to time; see Chapter 11!). This division was first described by Endel Tulving, an Estonian-Canadian psychologist.

Remembering life events: Episodic memory

Episodic memory covers memory for events you’ve experienced. Tulving claimed that episodic memories have the following characteristics:

Developed recently (in evolutionary terms)
Develops later in life; that is, infants don’t have it
Earliest to be affected by ageing
Easiest to be damaged by brain injury

You can differentiate episodic memory from autobiographical memory, which contains detailed information of events and experiences that have personal meaning. Autobiographical memory lasts a very long time whereas episodic memory is much more trivial and shorter-lasting.

One of the key pieces of evidence for the existence of episodic memory comes from neuropsychological studies on brain-damaged patients (check out Chapter 1). Hugo Spiers, a British neuroscientist, and colleagues reviewed 147 cases of amnesia (long-term memory loss) and found that episodic memory was damaged in all cases, but semantic memory was less often damaged.

People with no episodic memory can’t remember things that they’ve done during the day, such as TV programmes they watch or conversations they have. But if they have normal semantic memory they can still function (children with no episodic memory can go to school, for example, and learn).

Typically, only damage to the hippocampus leads to episodic memory loss. The hippocampus is a small, sea-horse shaped structure of the brain in line with the ear. The area of the brain including the hippocampus is called the medial temporal lobe.

Remembering facts: Semantic memory

Semantic memory is your store of all the information that you’ve obtained throughout your life. The structure of semantic memory is open to much debate and discussion – so much so that we devote a whole chapter to it (Chapter 10). Here, we present some basic information about semantic memory and relate it to episodic memory.

One of the key differences between semantic and episodic memory is that semantic memory is much more conceptual: the knowledge is stored as concepts and not linked to any experience of obtaining them. For example, you may know that the Hag˘ia Sophia is in Istanbul, but the fact that you watched a TV programme about it yesterday is an episodic memory and that you visited it with your friend last January is an autobiographic memory.

Although some patients can have no episodic memory but an okay semantic memory, many people with amnesia have problems with both episodic and semantic memory. When the brain injury is bigger, including parts of the brain around the hippocampus (the perirhinal and entorhinal cortices), semantic memories are affected.

If these two types of memory really are different, some patients should show the opposite pattern (a double dissociation; see Chapter 1) of an okay episodic memory but no semantic memory. They do: sufferers of semantic dementia can’t recall semantic information but generally have a working episodic memory (see Chapter 21 for a case study).

Semantic memories are usually stronger than episodic memories, because you’ve recalled them many times, whereas you don’t recall episodic memories that frequently. Some episodic memories that you do recall a lot, such as childhood stories you tell people, are more similar to semantic memories and are often unaffected by hippocampal damage.

‘But I don’t know how I do it!’ Non-declarative memory

Non-declarative memory is implicit: it’s not available to your consciousness. You may well be able to remember your first swimming lesson (an episodic memory; see the preceding section) but if you try to describe how you swim, you wouldn’t be able to describe verbally the exact process in detail (such as precisely which muscle to move exactly when).

Psychologists distinguish four types of non-declarative memory:

Procedural memory: A memory for how to do things
Priming memory: Repetition of information and where a recent event or thing influences your behaviour
Associative learning: Conditioned behaviours, where you learn to link events or objects as being related
Non-associative learning: Habits, where you learn behaviours and knowledge through experience

Remembering doing things: Procedural memory

You use procedural memory almost every single second of your life: in other words, you’re almost always doing something that you’ve learnt. Procedural memory exists for every motor skill you have, such as writing, language learning, walking and playing sports.

One of the most important ways of investigating procedural memory is to look at how new skills can be learnt. Psychologists often train people on rather useless novel skills to study these processes. Examples include mirror tracing (learning to write by looking in a mirror – try it, it’s very difficult), mirror reading and artificial grammar learning (see Chapter 14).

The first few times that people start learning these tasks, they probably rely on episodic and semantic memory (that is, declarative memory). For example, when you start to learn to drive a car, at first every single task is very difficult (checking mirrors, turning the wheel, controlling the gears, using the indicators and pressing the brake – just to make a right turn!).

But with practice, all these abilities are changed into a single procedural task, ‘turning’. This process is similar to the chunking process we describe in Chapter 8: you chunk all the tasks into one. At this point, what you’re doing can no longer be easily described and is therefore a procedural memory.

Evidence from neuropsychological studies of brain-injured patients with episodic and semantic memory problems shows that they have no problem with any of the skills they used to have; that is, their procedural memories.

Clive Wearing, a famous conductor and musician, developed amnesia as a result of a virus to the brain. Despite losing much of his episodic and semantic memory, he can still remember how to play music and conduct. In fact, he can learn new music and new skills (such as mirror writing), even though he has no memory of learning them. He’s surprised every time he mirror writes because he believes that he has never done it before.

Further studies show that even severely amnesic patients can learn skills necessary for the real world, such as using new machinery and driving. Having an episodic memory of the learning experience is helpful (because you know that you’ve learnt it), but it’s not vital.

Psychologists think that procedural memory is controlled by the striatum, a brain area just on top of the hippocampus. They can’t conduct studies with patients with lost procedural memory but intact episodic and semantic memory, however, because those patients wouldn’t be able to do anything and would be paralysed (though with no physical damage to the body).

Priming memory

Priming is a rapid form of unconscious learning. Both the following types mean that the repeated presentation of something affects the processing of it the second time:

Perceptual priming: If you perceptual prime something, when you see it again, you process it faster.
Conceptual priming: The same as perceptual priming, except that the two things need to be related conceptually rather than looking the same.

For example, presenting the word ‘baby’ makes people faster at recognising the word ‘baby’ (perceptual priming) and the word ‘cot’ (conceptual priming).

A lot of evidence shows that patients with amnesia have intact priming. In fact, the famous amnesic HM (see Chapter 21) had priming abilities of the same level as people without amnesia. Conceptual priming doesn’t occur in patients with semantic dementia (again, check out Chapter 21), because they’ve lost their semantic memories, which are typically in the form of concepts (see the earlier ‘Remembering facts: Semantic memory’ section). But perceptual priming does exist in patients with semantic dementia.

Priming is known to activate the brain areas in the cortex associated with the thing being primed. So priming the word ‘baby’ causes all the brain regions associated with babies to become active for a short period of time. This extra bit of activity makes it relatively easy for additional information to activate them fully. Therefore, detecting these things for a few seconds is easier (known as perceptual fluency).

Conditioning memory: Associative learning

Some forms of learning exist even without conscious memory. Studies show that even patients with amnesia can be conditioned to learn a new behaviour:

Associative learning: Where you learn to link two things together (such as the colour and shape of an orange and its flavour, or the library with the process of learning!). It’s the basis of all simple learning – even babies can learn this way!
Conditioning: Where you learn a behaviour through training and behaviour modification or by associating a stimulus with a behaviour (see the nearby sidebar ‘Psychology as behaviour’).

Children who don’t have episodic memory can be conditioned to fear things: the most famous case is that of the 1-year old Little Albert. Whenever poor Albert played with a pet rat, John Watson, an American psychologist, made Albert afraid by making a loud noise behind him. Very soon Albert developed a fear of white rats (or indeed anything white and fluffy, including Santa Claus’s beard). Later in his life, he retained this fear but had no episodic memories of the fear training.

Psychology as behaviour

In 1919, American psychologist John Watson created the approach known as behaviourism (see Chapter 1), which is concerned only with observable behaviour and all about the link between a stimulus and a response. Two forms of learning exist according to behaviourists:

Classical conditioning: Russian physiologist Ivan Pavlov found that dogs learnt to associate two stimuli together – the sound of a trolley bringing food and the food itself. The dogs salivated to the food, but after learning the pairing, they salivated to the trolley noise.
Operant conditioning: Particular behaviours are rewarded to increase the chances of the behaviour happening again (say, by giving food) or punished to reduce the behaviour (say, with a slap). So behaviours can be explained in simple stimulus–response pairings. For example, the simplest way to teach a child to tidy her room is to reward her after she tidies a bit (but don’t do it too often; otherwise she may learn to clean obsessively!).

These forms of learning are thought to be associated with brain regions including the cerebellum (for operant conditioning) and the amygdala (for classical fear conditioning).

Learning habits: Non-associative learning

Learning behaviour can occur through simple processes such as habituation and sensitisation, both types of unconscious forms of learning:

Habituation: When responses to a particular stimulus are reduced after prolonged exposure to that stimulus. For example, when you first put on your clothes, you’re aware of the material rubbing on your body, but after a short while you’re no longer aware of the sensation.

Habituation is seen in American developmental psychologist Robert Fantz’s interesting studies on children’s perceptions. He put infants into a viewing chamber and presented them with a stimulus (such as a checkerboard pattern). At first, the infants looked at the stimulus and then they got bored with it (habituated to it). If, after he replaced it with a new stimulus, infants showed renewed interest, the suggestion is that they can tell the difference between the two images.
Sensitisation: Similar to habituation, but where repeated exposure to something causes participants to become overly responsive to it.

This occurs for especially irritating things. For example, the sound of your partner snoring is extremely irritating (especially when you’re trying to sleep). Instead of getting used to this sound, it gets more and more annoying because you’re becoming sensitive to the sound.

The information regarding the stimuli to which people become habituated or sensitised is stored in their memories. When they’re presented with these stimuli, they simply react to them due to their brains’ reflex pathways.

Storing and Recalling Long-Term Memories

Information needs to be stored appropriately to be in your LTM. In the earlier section ‘Digging Deep: Levels of Processing Memory’, we indicate that different processes help to store memories. But the levels of processing framework doesn’t tell psychologists how memories are stored.

Also, after memories are stored in memory, you need to be able to use them (otherwise storing them is pointless). Memory may be like a warehouse stored full of knowledge, but that knowledge needs to be used from time to time. This process is called retrieval.

In this section, we investigate the active processes involved in storing information in and retrieving it from your memory. Of course, these processes concern recalling memories of things that happened in the past, but we also cover another type of memory: remembering for the future.

Consolidating memories

The process of storing memories in the brain is consolidation. Consolidation modifies the coded information from your perceptual system and binds (see Chapter 8) and combines all the features of what you’re learning together. In other words, it fixes it into memory.

Consolidation takes a great deal of time, which means that the process of the successful storage of memories can be enhanced or made worse:

Enhancing consolidation:
- Drugs that stimulate the central nervous system, administered after learning, enhance learning. Other naturally-occurring stimulants can also increase consolidation, including endorphins released following exercise.
- Sleeping after learning enhances the consolidation of procedural memories and, to a lesser extent, semantic memories (see the earlier section ‘Remembering facts: Semantic memory’ for more on these memories).
Reducing consolidation:
- Electric shocks to the brain (up to 14 hours after learning)
- A lack of oxygen to the brain (caused by suffocation)
- Certain drugs (propranolol to treat post-traumatic stress disorder)

Consolidation has two stages:

Stabilisation of the cells in the memory centres of the brain. This process can take minutes or hours to complete.
Reorganisation of the parts of the brain that store all your knowledge. This process can take days, months or years to complete.

The reorganisation stage occurs in the hippocampus. As a result, the hippocampus is one of the most important structures in the brain for memory. It reorganises memories by activating all features of the memory at the same time and binds them together. The hippocampus can bind new information with existing information in memory by activating both at the same time. We illustrate this process in Figure 9-2.

© John Wiley & Sons, Inc.

Figure 9-2: The process of consolidation in the hippocampus.

Retrieving memories

Your memory wouldn’t be very good if all it could do was store information. It would be like a library without any visitors: quiet and dull. No, you need to retrieve information from memory as well. Interestingly, this area of cognitive psychology hasn’t received as much attention by researchers as the storing of information.

We discuss three types of retrieval:

Active/conscious retrieval: When you try consciously to bring the information to mind.
Non-deliberate/unintentional retrieval: When things just spring into your mind.
Recognition: Knowing that you’ve encountered something before.

Cognitive psychologists often investigate these three types using slightly different techniques (here we give examples of words, but these forms of retrieval are for all types of material):

Free recall to assess active, conscious retrieval: Participants are asked to recall as many items as they can from a list of presented words.
Cued recall to assess non-deliberate retrieval: Participants are given prompts to aid their remembering: for example, they’re provided the preceding word in the word list and asked to recall the next word.
Recognition: Participants are presented with a series of words and simply asked to identify whether they’ve seen a particular word before.

Conscious retrieval

Retrieval brings memories back into an active state and reinstates aspects of the past. In fact, retrieval changes a memory so that it goes through the process of consolidation (see the preceding section) again, meaning that it can be damaged and affected in the same way as if it were new learning.

Some psychologists call this process reconsolidation, because you reinstate the pattern of brain activation associated with when you learnt the information during retrieval. The only difference in the pattern of brain activation is that it starts from the cortex (where the memories are stored), moving to the hippocampus (where the memories are consolidated) and then to the sensory cortices (where the item was perceived), whereas during encoding, this pattern is in reverse.

Retrieval doesn’t produce an identical pattern of brain activation to that seen during learning, because learning requires the involvement of a few more processes and because some information hasn’t been stored properly. It may also represent memories changing while they’re in your head.

Active retrieval is primarily directed by the frontal lobes in the brain. These areas control attention, thinking and deliberate behaviours. Patients with damage to the frontal lobes often have trouble retrieving information. Even when they can retrieve factual information, they can’t remember how or when they learnt the fact itself (called source amnesia).

Unintentional retrieval

Sometimes you can’t help but remember things, often when you’re anxious or worried about something. Your mind wanders and you recall many events. All very annoying, and it can sometimes stop you sleeping.

This irritating tendency is down to the fact that retrieval of memories is dependent on cues: hints or clues that help lead to the memory. They can come from internal or external sources:

Context: Memories are better retrieved when you’re in the same place as where they were learnt.
State: Memories are better retrieved when you’re in the same physical condition as when you learnt the information. For example, researchers showed that if you learn a list of words after having smoked marijuana, you’re better at recalling those words if you’ve just smoked marijuana.
Mnemonics: If during learning, you link information to something you’ve already stored in your memory, you’re better able to recall it.

Cues help retrieval because when you learn information, you don’t simply encode the information. Instead, you’re binding all the features of the information or event together in your memory: you’re trying to form links between every aspect of the information to be learnt. So when you store the information in your brain, it’s stored according to the information, the environmental context and the state you were in when you learnt it.

For successful retrieval, you try to access the stored memory, which requires you to find a route to the memory: the more cues to that memory, the easier the retrieval. The more you’re able to reinstate the context of learning, the easier you find retrieving the information (called the encoding specificity principle; see Chapter 11 for more details).

Therefore, you sometimes retrieve information when you don’t want to because the cues to that memory have become active. These cues then activate the memories attached to them. When you’re anxious, all memories associated with anxiety become active and keep popping into your head.

Recognition

Recognition is the process of seeing something that you’ve seen before and knowing that you know it. For example, you may bump into someone when walking to school or work one day. She says ‘hello’ and you politely respond, knowing that you know the person but not remembering her name. This leads to that awkward conversation where you’re trying to find out who she is without her realising that you have no idea of her name!

Recognition is central to memory because it’s the unconscious process of establishing whether something is new to you (and needs to be treated with caution) or familiar. It differs from recollection, where you have a distinct episodic memory (which we discuss in the earlier section ‘Remembering life events: Episodic memory’) for something. Unlike recollection, recognition isn’t based on retrieval cues, just on that sense of knowing something.

Cognitive psychologists can test episodic memory using the remember/know procedure. People are given a memory test in which they learn some words and later are shown more words and asked whether they saw the word before or not. If they did, they’re asked whether they actually remember it (with an episodic memory attached) or if it’s just familiar (a know response).

Recognition is different from recollection:

Recognition is faster and less affected by context and state, because these cues to memory aren’t needed.
Failure to recognise something is usually because the item isn’t stored in memory, whereas failure to remember something is due to a lack of an appropriate retrieval cue or because the item isn’t stored in memory.

Travelling back to the future: Prospective memory

One type of memory is vital to your survival: prospective memory, which is remembering things for the future. Prospective memory concerns how people remember to do things, such as meeting that really hot guy at the cafe tomorrow afternoon. Therefore, it still involves retrieving memories but faces additional problems to recalling past events.

Cognitive psychologists have identified two types of prospective memory:

Time-based: Remembering to do something at a given time
Event-based: Remembering to do something when a particular event occurs

Event-based prospective memory is more reliable than the time-based type, because the event can serve as a cue to remember.

Prospective memory has five stages:

Forming the intention to do something.
Monitoring the environment for a cue to the event (another event or a time cue).
Detecting the cue and retrieving the intention.
Recalling the intention.
Executing the intention.

These stages contain all the possible chances for error as memories for the past, as well as the added component of a possible long retention interval between forming an intention to do something and when this intention is executed. This long delay can cause the intention to fade from memory or be interfered with (see Chapter 11 for all about forgetting).

Prospective memory has other differences to memory for the past as well:

It’s often about knowing when to do something rather than knowing about something.
It’s more reliant on internal rather than external cues and therefore more vulnerable (external cues are usually more useful aids to memory).

The vulnerability of prospective memory is highlighted in many tragic cases of plane crashes. Although they’re extraordinarily rare, when they do occur failures of the pilot’s prospective memory is often the cause. Prospective errors are more likely to occur when someone is interrupted from a task they have planned to do. Just think about how many times you plan to do something when you get to work, but you forget. The reason may be because you get distracted meeting someone and this distraction disrupts the cue.

Looking at When Memory Goes Wrong

People can’t remember everything that happens to them. They often forget things such as when they were married or how old they are (though the latter may be deliberate in our cases!). We cover relatively minor disturbances in memory in Chapter 11, but more serious cases of memory loss also exist, as we describe in this section.

Two main forms of memory loss are associated with brain injury:

Anterograde amnesia: People can’t form new memories.
Retrograde amnesia: People lose their existing memories.

These types of memory loss can be caused by head injury, severe alcoholism or viruses that damage parts of the medial temporal lobe. These cases typically damage episodic memories (memories of your own lifetime) and not semantic memory (memory for facts). Check out the earlier ‘ “Let me tell you all about it!” ’ Declarative memory’ section for more.

In this section we detail these two types of amnesia, providing examples and showing what cognitive psychologists have learnt from them.

Failing to form new memories

People with anterograde amnesia usually have their non-declarative memories intact (see ‘ “But I don’t know how I do it!” ’ Non-declarative memory’ earlier in this chapter), so they can still learn new skills.

Think of the film Memento, where Guy Pearce’s character has lost his ability to form new memories. His experiences quite accurately depict anterograde amnesia. He lives his life from one brief moment to the next, sometimes being unaware how he arrived somewhere and what he was doing.

Clive Wearing’s pattern of anterograde amnesia is typical. Wearing (whom we introduce earlier in ‘Remembering doing things: Procedural memory’) describes his life as being in a constant state of feeling as if he has just woken up. With no continuity to his wakefulness, his only awareness is of what’s in his attention at one particular moment. As soon as he’s distracted, the effect is like being woken up. He can meet new people and become friendly with them, but at the next meeting he doesn’t remember them.

Losing stored memories

Many films and TV programmes depict retrograde amnesia, in which characters are hit over the head and lose all memories of themselves (like The Bourne Identity).

If you’re asked to recall ten events from your life where you were riding a bike, chances are that you’d remember events from throughout your life, emphasising recent ones but also including old events. But if someone with retrograde amnesia is asked, a completely different pattern emerges. The person would recall memories from much further back in her life and no recent events. This temporal gradient indicates that older memories are stored more strongly than more recent events (sometimes called Ribot’s law).

Someone with retrograde amnesia could navigate around the place where she grew up as a child with no problem. But when asked to navigate around where she currently lives, she’d struggle.

Two main explanations suggest why this pattern of memory loss may exist:

Consolidation and reconsolidation of memories strengthens memories: Every time a memory is reactivated, it becomes stronger, with more retrieval cues, and less susceptible to damage later on.
Semanticisation of episodic memories occurs: Older memories are less emotional and ‘episodic’ than recent memories. In fact, older memories resemble semantic memories (which aren’t damaged in amnesia). This suggests that older episodic memories have turned into semantic memories (we define these types of memories in the earlier ‘ “Let me tell you all about it!” ’ Declarative memory’ section).

Retrograde amnesia is caused by damage to the hippocampus and surrounding brain structures: the more damage to these areas, the more severe the memory loss. Another common cause of retrograde amnesia is Korsakoff’s syndrome, which is caused by a deficiency in the vitamin thiamine (B1), usually due to chronic alcohol consumption.

You Don’t Remember Our Wedding Day? Long-Term Memory