The surroundings and circumstances in which an event takes place

Events always occur in some context (from Latin contextus, 'joined together', 'woven together'). The dissociation between a memorized 'target' event and its context is not trivial. This is beautifully illustrated by Locke (1690): '... a young gentleman, who, having learnt to dance, and that to great perfection, there happened to stand an old trunk in the room where he learnt. The idea of this remarkable piece of household stuff had so mixed itself with the turns and steps of all his dances, that though in that chamber he could dance excellently well, yet it was only whilst the trunk was there; nor could he perform well in any other place, unless that or some such other trunk had its due position in the room'. Had the young gentleman been the subject of a controlled experiment on the acquisition of dancing "skill, it would have been fairly easy for the experimenter to discern the target task (dancing) from its context (the trunk). Furthermore, in this case, the context was an innocent part of the environmental surrounds, not expected to influence the target "stimulus in any significant way. This is in contrast to another type of'context', which always affects the significance or meaning of the stimulus, such as the context of a figure in a visual scene, a note in a musical score, or a word in a sentence (Wickens 1987). Having made this distinction, it is still useful to remember Locke's gentleman: what is regarded by the experimenter as an irrelevant context may not so be regarded by the brain of the "subject.1

In the behavioural literature, 'context' typically refers to environmental stimuli that are kept relatively constant in the course of the experiment, and is hence synonymous with apparatus, environment, and background stimuli (e.g. Lubow and Gewirtz 1995). In "real-life, the background often varies significantly from one moment to another, but still is not usually considered the target of learning or "retrieval. The information about the time and place in which the target was acquired is referred to as 'source information'. The context itself could, however, become the experimental variable. This is illustrated in a study by Godden and Baddeley (1975), in which divers were instructed to learn word lists both on land and underwater, and subsequently "recall the lists either on land or underwater. After learning on land, recall was better on land, and after learning underwater, recall was better underwater (Figure 20). It is likely that in this experiment, the context was not only the underwater milieu, but also the physiological state of the subject, generated

Recall environment

Edge of water Underwater

r te





f of






o r










r a e






Fig. 20 A case of context-dependent learning. Godden and Baddeley (1975) instructed groups of members of a university diving club to learn lists of words either while seating 20 feet under water or while seating by the edge of the water. The 'subjects were then requested to recall the lists either underwater or on land. The lists learned underwater were recalled best underwater, whereas the lists learned on land were recalled best on land. The height of each bar represents the success on performance in the task. (Adapted from Godden and Baddeley 1975.)

by the underwater experience. Learning that is dependent on the endogenous state of the subject is called "state-dependent learning. Distinct endogenous states are usually induced, either in real-life or in the laboratory, by chemical substances, but could also be instated by the external context.2 In such case, the distinction between context-dependent learning and state-dependent learning becomes a matter of taste or professional "bias.

In their ability to reinstate context, be it internal or external, not all senses are equal. In his famous Remembrance of things past (which is probably cited many more times than thoroughly read), Marcel Proust (1913) retrieves his childhood memories to the flavour of madeleine cakes soaked in tea. His literary account of the mnemonic power of the chemical senses appears to have received experimental validation: odours are superior to colours in supplying context for target information (Pointer and Bond 1998). Proust's account also points to the flavour of the madeleines as a retrieval "cue, i.e. a stimulus that elicits retrieval by activating an "engram and interacting with it. Retrieval cues are effective to the extent that they help reinstate or recreate processes involved in the original learning (McGeoch 1932; Tulving 1983). They may hence point either to target information, or to source information, or both. An interesting case in which the target information serves as a retrieval cue to bring to mind source information is "flashbulb memory.

Compelling evidence for the functional dissociation of target and source information in the brain comes from the study of human "amnesia. In some types of amnesia, the patient can recall facts or events, but not when and where this information was acquired. This is therefore termed 'source amnesia' (Schacter et al. 1984; Mayes 1988; Lishman 1998). A "classic case of source amnesia was described by Claparede (1911) in a lady with Korsakoff amnesia. Claparede used a trick that probably would not have been accepted nowadays by editors and readers, but still brought him fame. He struck the hand of the patient with a pin hidden between his fingers. Later, when trying again to reach out for the patient's hand, she pulled it back without being able to explain why. In another experiment, Claparede read to her stories, which she was able to recollect while being totally unaware of how these stories were learned. Source amnesia has since been reconfirmed to affect Korsakoff patients, and has also been demonstrated in other amnestic and demented patients (Shimamura and Squire 1987; Brandt et al. 1995; Schnider et al. 1996b). The pathology is considered to result from damage to multiple areas, including the thalamus and frontal "cortex.

In animal studies, the brain organ recurrently implicated in contextual memory is the "hippocampal formation (Penick and Solomon 1991; Holland and Bouton 1999). This is also supported by studies of human subjects (Chun and Phelps 1999). Coming to think about it, this is not surprising. Since context refers to combinations of features contributed by input from multiple modalities, one should expect brain circuits that encode context to process complex polymodal associations, and the hippocampus does just that. Currently, a popular "assay of contextual learning in rodents is a version of "fear conditioning. In the common classical fear conditioning "paradigm, a tone is associated with a shock, and the subject comes to fear the tone. However, as is often the case in seemingly simple "classical conditioning paradigms, there is also learning of the context in which training is performed, e.g. the conditioning chamber. Hippocampal lesions impair the acquired fear of context but not of the tone (Phillips and LeDoux 1992). The effect of hippocampal damage on context learning is, however, detected only when certain protocols or measures of fear conditioning are used, but not others. This implies that some elements of contextual learning can do without a fully functional hippocampus (McNish et al. 1997; Frankland et al. 1998; Maren et al. 1998; Gewirtz et al. 2000).

An notable development in the study of context in recent years is the large increase in the reports of cellular correlates and candidate mechanisms of context (e.g. Zipser et al. 1996; Christensen et al. 2000; Shulz et al. 2000). Such studies identify responses of neuronal populations or single neurons to target stimuli, that are affected by the surroundings and the circumstances. Discussion of "plasticity- and learning-related context has even been "reduced to the "synaptic and molecular level. A striking example is provided by the report that the ability of weak synaptic stimuli to establish "long-term potentiation depends on the recent history (<3h) of the other synapses on the same neuron; this means that stimuli that are ineffective in one context are effective in another (Frey and Morris 1997). But whether this intersynaptic gating of stimuli indeed subserves the effects of context in acquisition or retrieval of memory is still an open question.

Selected associations: Cue, Real-life memory, Retrieval, State-dependent learning

'This applies even to the simplest of learning in *simple systems: 'non*associative' learning, such as 'habituation, could involve an association between the unconditioned stimulus and the context (Marlin and Miller 1981; Rankin 2000).

2Indeed, Locke's trunk-dependent dancer, mentioned above, could also be described as trapped in state-dependency, where the retrieval-permissive state is reinstated in the brain by the presence of the trunk. The formation of the state dependency might have been promoted by the emotions involved in acquiring the dancing skill.

0 0

Post a comment