آموزش مجدد لیبل کلامی در دمانس معنایی

Semantic dementia is a degenerative disorder of temporal neocortex characterised by loss of word and object concepts. There is limited evidence that temporary relearning of lost vocabulary may be possible, attributed to sparing of hippocampal structures. However, learning is variable across patients and factors underlying learning success are poorly understood. The study investigated relearning of object names in two severely anomic semantic dementia patients. Following memory models that assume that hippocampal memories require some neocortical representation to underpin them it was predicted that relearning would be influenced by patients’ residual semantic information about stimuli. Experiment 1 confirmed that residual knowledge influenced learning success. On the assumption that neocortical knowledge encompasses concepts of space and time, as well as words and objects, it was predicted that learning would be affected by the availability of contextual (temporo–spatial) information. Experiment 2 demonstrated effective learning of object names, attributed to the patient’s use of temporal order and spatial position knowledge. Retention of object names over months was linked to the patient’s capacity for autobiographical experiential (temporo–spatial contextual) association. The findings indicate that relearning of lost vocabulary is possible in semantic dementia, indicating a role of the medial temporal lobes in the acquisition of semantic information. Effective learning does not imply reinstatement of lost concepts, but, it is argued, does involve some reacquisition of meaning. The findings challenge the traditional semantic–episodic memory dichotomy and are consistent with a “levels of meaning” account of semantic memory.

Semantic dementia is the descriptive term applied to the progressive loss of semantic knowledge that occurs in association with focal degeneration of the temporal neocortex [14] and [28]. A salient characteristic is an inability to name and to understand the meaning of words [14], [28] and [41]. Impaired naming ability is frequently so profound that patients score at floor level on standard picture naming tasks. The semantic disorder is not, however, confined to a single sensory modality [3] and [33]. Patients have difficulty recognising faces and objects, may no longer understand the significance of non-verbal auditory sounds such as the ringing of a telephone or doorbell and may fail to identify smells, tastes or tactile stimuli. Non-semantic cognitive skills, by contrast, remain well preserved. Language is phonologically and syntactically correct and patients speak fluently and without effort. Patients carry out normally perceptual matching, spatial judgement and constructional tasks and show no difficulty in the detection and discrimination of tastes, smells and tactile stimuli that they do not recognise. A striking clinical feature is patients’ well preserved current day-to-day memory. Patients have no difficulty recalling autobiographical events, remember appointments, carry out activities at the appropriate time of day unprompted and find their way around their locality without becoming lost, enabling a degree of functional independence that would be inconceivable in patients with classical amnesia. There is evidence, both from neuroimaging [4], [5], [13] and [22] and post-mortem pathological examination [33] and [34] that inferior and middle temporal gyri are predominantly affected, highlighting the importance of these temporal neocortical structures for the representation of conceptual knowledge. Medial temporal structures are typically relatively intact and it is the preservation of hippocampal structures to which patients’ capacity for current day-to-day memorising is attributed. There is, nevertheless, growing awareness that the conceptual separation between memory systems, namely a temporal neocortical system for the representation of semantic facts and a medial temporal system for the formation of episodic memories, is an oversimplification. Despite impressive functioning in their daily lives patients with semantic dementia typically perform rather poorly on traditional episodic memory tests, indicating that impairments in semantic memory also influence episodic memory performance. Moreover, there is anecdotal evidence that patients with semantic dementia can learn new factual information and reacquire lost semantic information at least to some degree. In our experience, a patient in the process of selling her house had no difficulty learning the name of her solicitor and of prospective purchasers. Several patients have successfully learnt the names of new acquaintances including hospital personnel and of newly prescribed medications. The patient KE who failed to recognise a newly purchased electric toaster had no difficulty learning its function and was able to consistently use it appropriately [29]. The inference is that the hippocampal system may play an important role, not only in event memory, but also in the acquisition of semantic facts. Such a view is supported by complementary findings that patients with classical amnesia, in whom the hippocampal system is damaged, are impaired in their acquisition of new semantic information [8], [40] and [45]. Nevertheless, in semantic dementia acquired information is unstable and depends on constant rehearsal. Our patient who could name her solicitor and house purchasers with ease during the period of her house sale no longer recognised those names a few weeks after her house sale was complete. The assumption is that although the hippocampal system supports new learning an intact neocortical system may be necessary for the stable, long-term representation of semantic knowledge. If the hippocampal system supports semantic learning then it ought to be possible to demonstrate fact learning too in an experimental test setting, albeit reliant on frequent rehearsal. However, to date formal evidence for successful fact learning has been limited. Graham et al. [11] demonstrated effective learning in their patient DM. Over a 15 months period DM showed improved performance on tests of naming and verbal fluency, which were attributed to his home drill. In a controlled study of category fluency, DM produced significantly more exemplars for categories that he practised at home than for non-practised categories. This beneficial effect of practice rapidly declined once he ceased his daily drill. DM is, however, unusual. Graham et al. found no benefit from practice in other patients with semantic dementia. The authors contrast, in particular, DM’s performance with that of the patient AM, who showed no evidence of improvement from self-initiated study, despite his apparently high level of motivation and the comparable time spent on his daily exercises. The factors that underlie fact learning in semantic dementia are, thus, far from clear. Why should DM show relearning of category exemplars but not AM? Why, moreover, should learning of semantic information be demonstrated so rarely in patients with semantic dementia in a formal test setting whereas convincing, albeit anecdotal, examples of effective learning are regularly demonstrated in patients’ daily life? Graham et al. posit several possible factors that might contribute to performance differences between DM and AM. Firstly, the two patients adopted different learning strategies. DM used descriptions, drawings and photographs of the concepts corresponding to the irretrievable names to assist vocabulary learning and words and pictures were organised predominantly by category. Thus, DM was continually activating both semantic and phonological representations of vocabulary that he was attempting to learn. By contrast, AM concentrated principally on word lists organised by initial letter, so that his practice was purely phonological. In view of the established memory advantage of semantic over phonological processing [16] DM’s superior learning of vocabulary may be attributable to his more efficient learning strategy. Secondly, the two patients differed considerably in terms of the magnitude of their anomia. AM performed virtually at floor level on a simple test of picture naming, whereas DM showed a relatively mild anomia. Thirdly, the patients may have differed in terms of the degree to which their anomia reflected genuine underlying semantic impairment. AM showed clear semantic loss, as evidenced by his poor performance on the pyramids and palm trees test [15], whereas DM’s scores on the same test were virtually normal. Graham et al. regarded the factors relating to semantic severity as of lesser interest than the strategic factor: they argued that patients such as AM might simply be so severely affected that little can be done to remedy the situation. There are nevertheless theoretically important reasons why issues of severity might be relevant. DM’s deficit is very mild. Although his naming skills are impaired it is likely that he has some remaining conceptual information pertaining to the objects whose names he is attempting to learn. It is probable that he recognised the pictures and photographs that he used as aide-memoires. In learning category exemplars he may have understood the meaning of category labels under which he grouped verbal terms. He may even have comprehended some of the words that he had difficulty retrieving. Thus, he had residual semantic information upon which to ‘tag’ the verbal labels that he was attempting to learn. By contrast, most patients with semantic dementia have a severe semantic loss, which cuts across input modalities: patients show no understanding of the words, of category labels and of pictures corresponding to words that they cannot retrieve. They have no meaningful tags on which to link the phonological form of the word. Thus, although a learning strategy that activates semantic representations might certainly be relevant in accounting for DM’s superior naming performance, it might be precisely because of his relatively mild disorder and hence the availability of residual semantic information that he is in a position to derive such benefit. A requirement for successful learning may be a meaningful substrate upon which the information to be learnt can be tagged. If that is so, then it is plausible that DM is unrepresentative of the semantic dementia population. He might be capable of relearning words only because his semantic impairment is so mild that he has sufficient conceptual knowledge to support new learning. A similar demonstration of relearning might not be possible in the majority of semantic dementia patients, whose conceptual understanding of the material to be learned is totally lost. Such an argument would be consistent with computational models of memory [19], [23] and [24] that posit an interdependence between the functions of temporal neocortex and medial temporal lobes. In establishing new memories, it has been assumed [23] and [24] that incoming sensory information activates representational nodes in association cortices, which in turn activates links with the hippocampus. That is, hippocampal memories require the presence of appropriate neocortical (semantic) representations. The inference would be that relearning of category exemplars in a patient such as DM, although mediated by hippocampal structures, is nevertheless dependent upon some knowledge substrate, represented by temporal neocortex, to support that new memory. A patient like AM who has little or no residual semantic knowledge about the information to be learned has no neocortical substrate from which hippocampal links can be forged. The prediction, then, would be that the capacity to relearn factual information in patients with semantic dementia should depend on the availability of residual semantic knowledge pertaining to the information to be learned. One aim of the present study was to test that prediction. Severity of semantic disorder per se is, however, unlikely to be the sole predictor of new learning capability. Clinical observation suggests that some new fact learning pertaining to patients’ current daily lives, such as the names of people or of newly prescribed medicines, is possible even in patients whose semantic impairment is very profound. Any account of new learning in semantic dementia needs to explain, for example, how our own patient KB who no longer understands such common vocabulary as pen, cow and glove nevertheless has had no difficulty learning the name of a new medication “ibuprofen” and asks for it by name in the pharmacy. Does such learning in patients with profound semantic disorder contradict or can it be accommodated within a model that supposes that (hippocampal) learning requires some (neocortical) semantic representation to underpin it? Semantic knowledge has traditionally been construed in terms of meaning of words and objects. Yet, conceptual knowledge about the world comprises substantially more than that. It includes knowledge of space and time. To reach for an object requires understanding of the spatial relationship between one’s own body and the external object. Daily activities depend on knowledge of temporal order: water flows only after the tap is turned on; the doorbell sounds after and not before it has been pressed. A daily routine, incorporating such activities as preparing breakfast, shopping, walking the dog, preparing an evening meal and getting ready for bed involve a conception of the continuity of time. Spatial and temporal concepts are strongly dependent upon neocortical regions such as the parietal and frontal lobes that are largely spared in semantic dementia. Consistent with this sparing, semantic dementia patients carry out normally complex spatial tasks and typically keep track of day, dates and time of day and indeed frequently ‘clockwatch’, carrying out activities at identical times each day [27] and [33]. Thus, although word and object-based semantic knowledge is profoundly impaired, temporal and spatial concepts are relatively preserved. Episodic memories, as originally defined by Tulving [36], are memories that are tied to a specific temporal and spatial context. In anatomical terms hippocampal structures would be viewed as drawing together information from distant neocortical sites: object information, represented by temporal neocortex, is linked with temporal and spatial information, represented by other brain regions. We presume that it is this linking of (weak) word/object information with (strong) spatial and temporal information that provides the basis for patients’ relative preservation of autobiographical memories. We have put forward a parallel argument to account for the finding that patients’ residual semantic knowledge appears to be influenced by its relevance to personal experience. Patients recognised names of personal acquaintances (e.g. Margaret Simmons) better than names of famous people (e.g. Margaret Thatcher). They recognised as places the names of places they had visited (e.g. Dordogne) better than the names of places that they had not visited (e.g. Portugal). One patient was shown to recognise personal objects (e.g. the patient’s comb) better than alternative exemplars of the same objects (e.g. the examiner’s comb) [29]. Another patient’s understanding of vocabulary [30] was found to be restricted to the way in which she used that vocabulary in her daily life. She understood “licence” to refer to her own driving licence but had no broader understanding of the term. We have observed too that famous names experienced directly through the medium of televised events (e.g. Jill Dando) are better recalled than high profile names, not linked to a personal experience (e.g. Winston Churchill). Parallel findings have been documented by others [44]. Knowledge seems to be personalised and tied to specific exemplars of a patient’s recent experience. To account for these findings, we supposed [31] and [32] that when the neocortical system is damaged as in semantic dementia existing corticocortical connections are gradually eroded and the rich associations that make up a normal concept are lost. However, personally experienced objects (or names) retain “meaning” for the patient in the sense that concepts of time and space, which are relatively preserved, provide the conceptual reference to which the object can be linked. A patient may no longer, for example, have the rich network of associations that under normal circumstances constitutes a person’s abstracted concept of dog. Nevertheless, they may have personalised knowledge of a particular animal, experienced in a particular place and at a particular time (e.g. the thing in the house on the opposite side of the road that makes a noise when the postman calls). The inference is that patients with semantic dementia learn new semantic facts in their daily lives, including the names of people, by virtue of the fact that they have a meaningful reference (temporal and spatial) to which labels can be attached. Clearly, daily life experiences provide a richer source of spatial and temporal contextual information than does the artificial situation of a laboratory-based learning task. Nevertheless, even the relatively impoverished context of, for example, a list learning task provides some limited spatial and temporal information, such as position on the page and sequential order. We would propose then, in line with current computational models [19], [23] and [24] that hippocampal memories require the presence of appropriate neocortical (semantic) representations. However, we would argue that semantic information extends beyond word and object meanings and includes knowledge of space and time and that in semantic dementia it is the preservation of spatial and temporal concepts that provides a (cortical) knowledge base, against which new learning of otherwise meaningless information might occur. Thus, we hypothesise that two factors will contribute to a semantic dementia patient’s capacity for relearning: (1) the degree of residual information the patient has about a word or object and (2) the availability of temporal and spatial contextual information. A patient with mild semantic impairment such as DM may have sufficient residual knowledge of the material to be learned to support learning irrespective of learning context. More severely affected patients may be more dependent upon the provision of experiential, temporal and spatial contextual information to support learning of what is otherwise meaningless material. The purpose of the present study was to examine the proposal outlined above that both residual word/object information and context information influence relearning performance in semantic dementia. The study examines vocabulary learning in two patients with semantic dementia using picture naming tasks. The first experiment tests the hypothesis that success of learning of object names by simple rehearsal is influenced by the degree of residual semantic knowledge the patient has about the object concept. It was predicted that learning would be superior when the patient had some residual knowledge, as defined by the ability to recognise the picture or comprehend the object name. Name learning was anticipated to be poorer when the stimulus was totally devoid of meaning, when the patient showed no recognition either of the picture or of the object name. The second experiment examines learning of object names in a patient with no residual knowledge about the stimuli to be learned. It was predicted that learning would be superior if the “meaningless” material could be linked to external contextual information and the patient’s experiential memories. The study helps to resolve the issue of whether new learning of semantic facts is possible in patients with profound semantic loss and to delineate factors that underlie learning success. The findings challenge traditional conceptions of the distinction between episodic and semantic memory and suggest an alternative conceptualisation of the nature of semantic knowledge.