Kinds of memory

A lot in animal behaviour seems to indicate the use of memory. Many animals develop an intimate knowledge of their local environment, such as where to find different food sources, where their home and shelters and where danger lies.

Processes of learning and memory storage are tightly connected although they may be differently localised within the brain. The study of memory in animals looks at how information that has been acquired at one time (learned) influences behaviour at a later time and traditionally investigates how information is stored and retrieved.

Pearce (2000) points out that there are three main questions about memory in animals. (1) How long can be the information retained?

(2) What type of information can be retained? (3) How much information can be retained?

Most of research on memory were historically connected with experiments on mammals and, at least, on vertebrates, although many animals with no cortex at all, and some without anything we would call a brain, can be classically conditioned. Moreover, we know now that excellent learning capacities and cognitive skills may be floated about eccentric nervous systems. It is worth noting that many facts concerning memory were first had been obtained from studies on humans and then applied to animals. To some extent the findings of the studies of memory in animals mirror those made in humans. Concepts of intelligence have fuzzy boundaries but one thing is clear to us: there is no straight correlation between powers in memory and intelligence in our species. There are many sorts of strange abilities associated with memory in humans. Idiot savants, a special class of retardates, which includes autistics, are able to multiply two five-figure numbers mentally. Another class of people called eidetikers possess " photographic memory”. The advantage of these capabilities seems enormous only on the surface for our species but not for other ones. Further in this book we will meet non-human idiot savants within classes of Birds and Mammals that take enormous advantage from their strange memory dimension.

Psychologists and memory researchers often divide memory into categories defined by the duration for which the memory is expected to last. The existence of destined categories of memory was first noticed in humans with brain concussions who were not able to recall what happened just before an accident, but could remember what happened earlier.

Sensory memory refers to the fact that, after experiencing a stimulus, information about that stimulus is briefly held in memory in the exact form it was received, until it can be further processed. Typically, sensory memories may last only a few seconds before decaying - or being overwritten by new, incoming information. But, while they last, sensory memories contain detailed information: almost like an internal " copy" of the stimulus, in perfect detail. For example, psychologists have assumed that there is a memory area where incoming visual information is stored as a picture or icon. This is sometimes called iconic memory. While visual information remains in iconic memory, an individual can answer detailed questions, such as what is the third row of numbers in a numerical display. Psychologists have assumed that there is also an echoic memory for auditory information (stored as an echo) and other buffers for information related to the other senses: taste, smell and touch.

Short-term memory refers to memories which last for a few minutes. Unlike sensory memory, which is stored in the exact form it was experienced, short-term memory has received some processing; thus, " A" is stored not as a visual stimulus, but as an abstract concept of the letter " A". Short-term memory is of limited capacity, usually 5-9 items. Beyond this capacity, new information can " bump" out other items from short-term memory. This is one form of forgetting. Objects in short-term memory can be of indefinite complexity: thus short-term memory can hold several numbers, or several words, or several complex concepts simultaneously.

Working memory is sometimes considered a synonym for short-term memory. However, memory researchers often consider this a specialised term, which is conceptualised as an active system for temporarily storing and manipulating information needed in the execution of complex cognitive tasks (e.g., learning, reasoning, and comprehension). Working memory is often described it as the " blackboard of the mind" or " online memory" (Goldman-Rakic, 1992). The term “working memory” is used to emphasise the fact that what is learned in one case may not be relevant on the next. There are two types of components: storage and central executive functions.

Speaking about human working memory, there are two storage systems within the model: the articulatory loop and the visuospatial sketchpad or scratchpad which are seen as relatively passive slave systems primarily responsible for the temporary storage of verbal and visual information (Baddeley, 1986).

Speaking about animal working memory in general, it should be emphasised that working memory is retained only long enough to complete a particular task, after which the information is discarded because it is no longer needed, or because it may interfere with the next task.

Long-term memory (reference memory)is memory that can last for years. It contains everything we know about the world, including semantic and factual information as well as autobiographical experience. In general, long-term memory is organised so that it is easy to reach a stored item by a number of routes.

Hebb (see Chapter 3) first distinguished between immediate memory (short-term memory, or STM) and long-term memory (LTM). He suggested that STM consists of active networks of nerve cells that repeatedly excite one another. These reverberating loops of cells can maintain a stimulus trace in the brain from a few seconds to many minutes. What makes the patterns of cells form into loops is that the cell that originally brings the sensory information into the cortex is hypothesised to receive onto its receiving processes (dendrites) inputs from cells further along in the loop. As the output processes (axon) of these cells stimulate the first cell, the cell then sends back signals to the next cells. This ultimately causes the first cell to be re-stimulated. The excitation in the loop, if fairly weak, finally dies down; the loop then " breaks" and the information is lost from the brain or " forgotten."

The development of LTM, called consolidation, is known now as a process of the transfer from short-term to long-term systems. Apparently, short–term storage may be necessary before long-term storage can occur. During the last decades, a number of studies have revealed that long term memory formation requires a de novo brain protein synthesis. Indeed, pharmacological experiments have shown that administration of inhibitors of protein synthesis around the time of training impairs LTM. In contrast, short- term- memory (STM) is based on transient changes in synaptic morphology (for review, see Welzl and Stork, 2003).

The notion that it takes time to form a stable memory seems to capture a basic property of memory. That is, during consolidation period after training, memories are susceptible to interference by shock, cooling, or pharmacological manipulations and could be obliterated. This consolidation period might ensure that only the really important memories are stabilised (Menzel, 1999): It seems important to keep memory modifiable for some time, to erase it in case of contradictory experience, enhance it in case of affirmative experience, or to modify it according to other, already existing memories.

To analyse memory consolidation on a behavioural level, one can make use of the temporal characteristics of retention. Kamin (1957, 1963) found that retention of an avoidance response in rats follows a non-monotonic time course: Retention is high immediately after training but then decreases to a minimum after 1 hr. Then, retention increases again to reach a stable level after several days. This " Kamin effect" (Denny 1958) was subsequently found in many organisms, including humans.

A number of authors, beginning with Kamin, have suggested that the simplest explanation is to assume the existence of two independent and additive memory systems: one memory system that dominates retention immediately after training but then constantly loses impact and a second one that needs time to consolidate and that is then increasingly responsible for retention. It is the " secondary" rise in retention that is taken as a behavioural indication for memory consolidation. The functional background of such a system might be that animals cannot afford “not to behave" until the stable, second memory is formed. Thus, they perhaps transiently use an independent, potentially less specific, memory system. It has been demonstrated on different organisms such as honeybees, octopuses, goldfishes, mice, and humans that in conditioning procedures retention is often a non-monotonic minimum function (Gerber and Menzel, 2000).

Long-term (reference) memory is retained for longer periods and is used for completing successive tasks. There are several different ways to classify long-term memories according to their content. Cognitive psychologists and neuroscientists have divided memory into two broad classes, or sets of strategies used by the brain to acquire information.

One set of strategies, termed explicit memory (or declarative, or relational), underlies memory for events and the circumstances of their occurrence. In human studies declarative memory is a term for information which is available to conscious recollection and verbal retrieval (i.e., it can be " declared"). In forming and storing explicit memories, associations are done with previous related stimuli or experiences. Therefore, explicit memories can be remembered and recalled, and rely on previous experiences and knowledge.

The other set of strategies, implicit (non-declarative, or procedural, or priming memory) memory, encodes information about perceptual and motor skills, using non-cortical structures and requiring no conscious participation. Implicit memories cannot be looked up or remembered to be used for actions and reasoning. They consist of memories necessary to perform events and tasks, or to produce a specific type of response. Implicit memory is best demonstrated when performance is improved on a mechanical task. This type of memory is shown through activation of the sensory and motor systems needed to perform a certain task.

Humans with amnesia cannot form new conscious, explicit memories, while implicit memory is left intact. Many learning tasks require both memory systems. Explicit memory, like implicit memory, has a short-term phase that does not require protein synthesis and a long-term phase that does.

Declarative memory has been further subdivided into episodic memory, which is autobiographical information, and semantic memory, which is factual information about the world (vocabulary items, knowledge of what things are used for, memory of multiplication tables, etc.). The term “episodic memory” returns us to the epigraph from Marcel Proust. Episodic memory is the ability to " reach back" into the past. More specifically, it is the ability to retrieve egocentric episodes as a temporal chain of life events.

There are some disagreement as to the nature of this division and the relationship between these memory systems, both in terms of function and the neural structures involved. It is generally agreed, however, that episodic memory is concerned with the conscious recall of specific past experience, whereas semantic memory involves the storage and retrieval of factual knowledge about the world (Griffiths et al., 1999). This difference is often referred to in terms of remembering versus knowing: episodic memory is concerned with the remembering specific personal experiences, whereas semantic memory mediates what one knows about the world. As Griffiths et al.(1999) give this, remembering getting soaked in the London rain last Tuesday is an example of episodic memory, but knowing that it often rains in England is an example of semantic memory because it need not be acquired as a result of a personal experience of getting wet.

According to Endel Tulving’s (1972, 1983) classical definition, episodic memory receives and stores information about temporally dated episodes or events, and temporal – spatial relationships among these events. Thus, episodic memory provides information about the “what” and “when” of events (temporally dated experiences) and about “where” they happened (temporal-spatial relations). The different kinds of memory are linked to different levels of " knowing": Non-declarative memory is anoetic (non-knowing), semantic memory is noetic (knowing), while episodic memory is autonoetic (self-knowing). This suggests that episodic memory is critically dependent on the concept of self.

When presenting a new definition of episodic memory, Tulving and Markowitsch (1998) state that it possesses features that no other memory system has and is accompanied by a special kind of consciousness (autonoetic consciousness). This is different from the “noetic” consciousness that is involved with the retrieval of declarative information. Such a distinction is based on the fact that human subjects can distinguish between recalling past personal experience and remembering an impersonal declarative fact. Remembering a specific event requires autonoetic consciousness, whereas knowing a fact is noetic in nature. Tulving and Markowitsch (1998) claim that episodic memory is that really sets human apart. Episodic memory allows human beings to travel back and forth mentally through time. It also gives us an awareness of ourselves and our state of existence within the dimensions of time.

Some recent studies, however, enable us to suggest that animals are capable, at least to some extent, of episodic memory. The fact is that most of the laboratory tests for animal memory that have been used up to date and which will be described in this chapter can be explained in terms other than episodic recall. A different strategy of testing whether or not animals are capable of episodic memory is to adopt an ethological approach, considering cases in nature in which an animal might benefit from the capacity to remember unique episodes that occurred in the past (Griffiths et al., 1999). We will return to this problem and concerned experimental results in Chapter 22.

In general, the terminology concerning different kinds of memory is not completely settled yet. On the one hand, in some fundamental works concepts of explicit and implicit memory have been attributed to snails. On the other hand, sometimes there are no distinct boundaries between concepts of explicit and implicit forms of memory and explicit and implicit forms of knowledge in scientific literature. Besides, distinctions between explicit and implicit knowledge in nonhumans are seldom made yet. This is probably because explicit knowledge is typically defined in terms of consciousness, and most researchers are hesitant to attribute consciousness to nonhuman species. There are, however, no fundamental functional or biochemical differences between the nerve cells and synapses of humans and those of many more “simple” organisms, both vertebrates and invertebrates such as snails and warms.

Many researchers believe that animals are capable of only non-declarative learning and memory, or at least that there is no way to test whether any of their learning is declarative because they cannot tell us what they remember and that animals may know many facts that are important in their lives but do not know what they know (Cheney and Seyfarth, 1990). Recently some insights came from different experimental techniques. These include, in particular, experiments with delayed reactions (see Chapter 10) and language-training experiments (see details in Part IX), in which apes, dolphins and African grey parrots have learned adaptations of human communication system.