Brain Advance Access originally published online on June 23, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Brain, Vol. 126, No. 8, 1821-1829,
August 2003
© 2003 Guarantors of Brain
doi: 10.1093/brain/awg187
Spatial coding of semantic information: knowledge of country and city names depends on their geographical proximity
1 Dementia Research Group, Department of Clinical Neurology, Institute of Neurology, University College and 2 Division of Neuroscience and Psychological Medicine, Imperial College, London, UK
Correspondence to: Professor Elizabeth Warrington, Dementia Research Group, The National Hospital for Neurology and Neurosurgery, 8–11 Queen Square, London WC1N 3BG, UK E-mail: e.warrington{at}dementia.ion.ucl.ac.uk
Received December 23, 2002. Revised April 3, 2003. Accepted April 7, 2003.
 |
Summary |
|---|
Top
Summary
IntroductionWe report the case of a patient, A.Z., with a refractory disorder of semantic processing. Using a number of spoken word–written word matching paradigms, her comprehension of the broad proper nouns category of place names was investigated. In addition to being sensitive to temporal factors, A.Z.’s ability to identify country or city names accurately was significantly worse when selecting a target from an array of geographically close rather than distant places. The influence of geographical proximity and location was observed even when identifying proximal and distal English towns, but not with less well-known American states. These close and distant effects are attributed to the build-up of refractoriness in representations with shared neural circuitry. Our findings reveal a previously undocumented level of fine-grain organization in this domain of semantic space. It is argued that information about geographical proximity cannot be encoded in purely verbal or visual terms. Consequently, we propose the existence of a separate module of spatially encoded information within conceptual knowledge.
Keywords: refractoriness; proper nouns; conceptual knowledge; spatial framework
Abbreviations: RSI= response–stimulus interval
|
Introduction |
|---|
Cognitive neuropsychological investigations of conceptual knowledge are based on the assertion that patterns of performance following damage to an organized semantic system enable us to draw implications about the nature of that structure. For example, using a lesion-based methodology, neuropsychologists have demonstrated that semantic knowledge may be divided along a number of axes (e.g. concrete/abstract, noun/verb). Within the noun domain, a series of increasingly fine-grain dissociations have been revealed, from the distinction between man-made artefacts/living things, to that between manipulable/non-manipulable artefacts and animals/fruit and vegetables. These dissociations, many of which would not have been widely predicted before their discovery, offer evidence of partially dissociable neural bases underlying different modules of semantic knowledge. However, it is possible that an even more fine-grain organization exists within even these narrowly defined categories. This prediction underlies any drive towards discovering the nature of a ‘concept’.
Within the context of such a drive for understanding, several criticisms have been levelled at lesion-based methodologies. One limitation of studies reliant on naturally occurring lesions of a vascular origin is that the size and impact of a lesion may mask an even finer level of organization which exists within the impaired cognitive module. In contrast, minimal events which cause highly specific cognitive deficits may be too slight to merit medical and neuropsychological examination. One solution comes from extending our investigations beyond the realm of static impairments of information processing. Indeed, it was the detailed examination of a patient with a dynamic access deficit of semantic knowledge following a large left hemisphere stroke which revealed evidence of fine-grain structure within the category of man-made artefacts (manipulable/non-manipulable; Warrington and McCarthy, 1987
).
A major distinction has been drawn between cognitive deficits arising from damage to central information stores and those attributable to impaired access to otherwise intact stored information. Impairments of the former type are known as ‘storage’ disorders, and produce patterns of cognitive performance which, though impaired, are consistent and item-specific (even when only partial knowledge is retained). For example, individuals with storage disorders of semantic memory consistently identify the same items correctly or incorrectly if tested on several separate occasions (e.g. Coughlan and Warrington, 1981). In contrast, impairments of the latter type, known as ‘access’ disorders, are characterized by response inconsistency when stimuli are presented repeatedly.
One form of access disorder is characterized by two further properties. First, the performance of access patients on word–picture matching tasks often is significantly affected by presentation rate. By introducing a brief interval between each response and the presentation of the subsequent stimulus (the response–stimulus interval; RSI), the response accuracy of access patients can be greatly facilitated. Such patients apparently require longer to ‘recover’ after processing a particular stimulus. Indeed, this temporal influence, termed ‘refractoriness’, has been described as ‘the reduction in the ability to utilize the system for a period of time following activation’ (Warrington and McCarthy, 1983).
Secondly, access patients are strongly affected by the semantic relatedness of competing responses. In word–picture matching tasks, they demonstrate greater difficulty in correctly identifying a target when it is presented in an array of items which are semantically close (e.g. trousers, blouse, dress, jacket) than semantically distant (e.g. teapot, key, hat, glasses). These semantic relatedness effects have been observed in several single-case reports of semantic refractory access disorders (e.g. Warrington and McCarthy, 1983, 1987; Cipolotti and Warrington, 1995; Forde and Humphreys, 1995). The effects are held to occur when refractoriness from one item affects neighbours in a categorically organized store of conceptual knowledge. On the basis of this theory, the occurrence and size of semantic distance effects are considered to reflect the degree of shared neural circuitry supporting individual conceptual representations.
Examinations of semantic distance effects have concentrated mainly on semantic relationships within broad categories of concrete nouns (e.g. Warrington and Cipolotti, 1996). However, category dissociations have been demonstrated between comprehension of proper nouns and common nouns (e.g. Semenza and Zettin, 1988, 1989). The further fractionation of the semantic knowledge of proper nouns has been predicted following the observation of dissociations within the proper noun domain, for instance between patients who show either a selective impairment or selective preservation of country knowledge (e.g. McKenna and Warrington, 1978; Warrington and Clegg, 1993; Incisa della Rocchetta et al., 1998). However, only in one case has the semantic relatedness of proper noun concepts even been contemplated (McNeil et al., 1994).
Here we report a patient whose semantic refractory access disorder was harnessed to examine the semantic relatedness of proper noun concepts. In particular, we demonstrate that access to knowledge of country and city names was significantly affected by the real-world geographical proximity of competing targets.
|
Case report |
|---|
The patient, A.Z., is a 72-year-old retired housewife. In January 2002, she was admitted to Northwick Park Hospital following a left middle cerebral artery stroke. She presented with a very mild right-sided weakness. Following the vascular incident, she also developed a severe expressive dysphasia which was fluent and characterized by semantic and phonemic jargon. There was further evidence of a mild receptive dysphasia. A CT scan conducted at the time revealed damage to a large portion of the left middle cerebral territory, sparing the basal ganglia. The left parietal, temporal and posterior frontal regions were all affected (see Fig. 1).
|
|
Neuropsychological background |
|---|
The investigations reported here were conducted between April and November 2002. A.Z.’s naming abilities were severely compromised. On a category naming test of very high frequency items, she was able to name only six out of 50 items correctly (McKenna and Warrington, 1978
). A.Z. was also almost completely dyslexic and dysgraphic. She was able to read correctly only one out of 99 words from the Patterson and Hodges corpus (Patterson and Hodges, 1992). A.Z.’s repetition skills were less severely affected, but nonetheless significantly impaired. She was able to repeat accurately 39 out of 99 of the same words from the Patterson and Hodges corpus. Auditory lexical decision performance was relatively preserved (71 out of 80, PALPA 5; Kay et al., 1992). On formal tests of single spoken word, written word or picture comprehension, a mild semantic impairment was demonstrated, as shown in Table 1. In particular, A.Z. showed some impairment in her comprehension of both verbal and visual material. With the exception of her poor knowledge of body part names, A.Z. also showed no indication that her comprehension deficit was category-specific. However, it was noted that if spoken word comprehension for an item was tested several times within a short period of time, A.Z.’s responses were inconsistent. It was this observation, typical in a case of access dysphasia, which led to the current investigations.
|
|
Experiment 1: the dynamics of the deficit |
|---|
In our preliminary investigations, A.Z.’s responses were inconsistent when matching spoken words to written words. This is characteristic of an access deficit. However, some access patients have been shown to be sensitive to the length of the interval between response and the subsequent stimulus. In order to assess whether A.Z.’s identification of written words was both inconsistent and sensitive to temporal factors, she performed a matching task at two different rates.
Stimuli
The stimuli consisted of four semantically close arrays, each consisting of the written names of four countries (e.g. Norway, Finland, Denmark, Sweden). A.Z. was presented with a spoken word and requested to point to the corresponding written word in the array. Each array was presented twice, varying the RSI. A.Z. was permitted either a 1 s RSI (fast condition) or a 10 s RSI (slow condition). Under each condition, every item in every array was probed four times in a pseudorandom order. Half the arrays were presented in the fast then slow conditions, and half vice versa.
Results
A.Z. was able to identify country names significantly more accurately when given a 10 s pause rather than a 1 s pause after each response (10 s rate = 54 out of 64, 1 s rate = 38 out of 64; 
2 = 8.70, P < 0.01, df = 1). This result shows that a portion of the system supporting spoken word–written word matching for country names becomes refractory with use.
|
Experiment 2: the locus of the refractory disorder |
|---|
In Experiment 1, A.Z.’s sensitivity to temporal factors indicated that her deficit was refractory in nature. However, although this temporal effect was demonstrated in a task requiring A.Z. to match spoken words to written words, these observations alone are insufficient to determine the processing stage in which refractoriness builds up. Such a pattern of performance may reflect a refractory deficit affecting either the transcoding of phonological (sound form) to orthographic (written form) information or the comprehension of the word concept itself. In Experiment 2, we attempted to evaluate these possible accounts by probing for evidence of semantic mediation in the spoken word–written word matching task.
Stimuli
The stimuli consisted of the names of nine people (politicians), nine places (countries) and nine product brands (shops). These were arranged into nine semantically close arrays of three items in which each item was drawn from the same broad category (e.g. people: Kennedy, Nixon, Reagan). The same stimuli were also arranged into nine semantically distant arrays, in which one item was taken from each of the three broad categories (e.g. Kennedy, Japan, Sainsburys). These arrays were probed in an ABBA design. Every stimulus in every array was probed four times in a pseudorandom order with a 1 s RSI.
Results
A.Z. made significantly more errors in identifying the target item from amongst an array of close than distant items (close = 79/108, distant = 102/108; 2 = 18.0, P < 0.001, df = 1). If A.Z. was performing this task by phonological–orthographic transcoding (i.e. matching the sound form to the word form without recourse to comprehension), one would not predict any effect of the semantic relationship between members of each array. However, the clear semantic distance effect refutes this explanation and indicates that A.Z.’s deficit lies at the level of semantic processing. These results also suggest that the broad categories of proper nouns have semantic representations which are at least partially separable. This is consistent with previous findings of dissociations between knowledge of people and places, and extends published work by predicting that product brand names constitute a further dissociable category.
|
Experiment 3: influence of geographical proximity on knowledge of country names |
|---|
In Experiment 2, we demonstrated that the semantic relatedness of competing responses has a significant effect on response accuracy. On that occasion, classification into semantically close and distant arrays was based on differences between broad categories of proper nouns (people, places, brand names). In Experiment 3, we explored whether semantic distance effects may also be observed within such categories of proper nouns.
Stimuli
For Experiment 3, the stimuli were the same set of country names used in Experiment 1. These were arranged into four semantically close arrays of four items in which each item was located in the same geographical region (e.g. Western Europe: Spain, France, Holland, Germany). The identical stimuli were also arranged into four semantically distant arrays, in which one item was taken from each of the four geographical regions (e.g. Spain, India, Brazil, Norway). Semantic groupings were determined a priori with reference to maps of the regions in question. The names in the arrays did not show any substantial difference in the degree of phonological or orthographic similarity. The arrays were probed in an ABBA design, with each stimulus presented four times using a 1 s RSI.
Results
The percentage correct responses with close and distant arrays are shown in Fig. 2. A.Z. identified target countries significantly more accurately from among unrelated (distant) than related (close) country names (close = 82/128, distant = 103/128, 2 = 7.80, P < 0.01, df = 1). The clear semantic distance effect indicates that country concepts do not have equivalent representations. Geographical proximity was the only factor which determined group membership. As semantic distance effects are held to reflect shared neural representations between concepts, this real-world proximity may be one principle underlying the organization of country concepts in the brain.
|
|
Experiment 4: influence of geographical proximity on knowledge of international and UK city names |
|---|
The findings of Experiment 3 indicated that real-world location may predict the refractoriness which builds up when our semantic refractory patient has to identify place names repetitively. In order to verify this notion, we examined the influence of geographical proximity on A.Z.’s processing of two other place concepts, namely international city names and UK city names.
Stimuli
Two stimulus sets were used. Set 4a consisted of four international cities from each of Italy, America, India and England which were arranged into four geographically close arrays (e.g. Rome, Milan, Venice, Florence) and four geographically distant arrays (e.g. Rome, Miami, Calcutta, Leeds). Set 4b included four cities from each of Scotland, Ireland, Wales and England which were arranged into close (e.g. Edinburgh, Glasgow, Aberdeen, Inverness) and distant (e.g. Edinburgh, Dublin, Swansea, Newcastle) semantic arrays. Set 4a and 4b arrays were presented as separate experiments in an ABBA design.
Results
The percentage correct responses with close and distant arrays are shown in Fig. 2. A.Z. experienced greater difficulties identifying target cities from among the names of other cities from the same country than from an international selection of cities (close = 43/64, distant = 54/64, 2 = 4.26, P < 0.05, df = 1). Similarly, recognition of English, Scottish, Welsh and Irish cities was significantly worse when identifying targets from among the names of cities from the same country than from different countries (close = 38/64, distant = 50/64, 2 = 4.4, P < 0.05, df = 1). These results demonstrate that the semantic distance effect found in Experiment 3 is not specific to country names per se. Semantic level representations of city names may also be strongly influenced by their real-world location relative to one another.
|
Experiment 5: influence of geographical proximity on knowledge of English city names |
|---|
In the experiments involving international cities and even UK cities, the close and distant semantic arrays differed according to whether or not the constituent members were located in the same well-defined countries. The design of a further experiment removed this division by including only the names of English cities.
Stimuli
For this task, semantic groupings were formed purely on the basis of geographical proximity (i.e. cities from the north, south, east and west of England; see Fig. 3A). Sixteen cities were arranged into close (e.g. Brighton, Southampton, Eastbourne, Portsmouth) and distant (e.g. Brighton, Norwich, Manchester, Plymouth) semantic arrays. Arrays were presented in an ABBA design.
|
Results
The percentage correct responses for the close and distant arrays of English cities are shown in Fig. 3C. A.Z.’s performance was significantly worse when the city names were arranged in geographically close than distant arrays (close = 41/64, distant = 53/64,
2 = 4.85, P < 0.05, df = 1). These results demonstrate that the semantic distance effects observed reflect the actual geographical proximity between competing responses and not merely superordinate category membership. Indeed, the cities used in Experiment 5 crossed county boundaries and perhaps are not traditionally recognized as members of a particular region. |
Experiment 6: influence of geographical proximity on knowledge of US state names |
|---|
The significant semantic distance effects recorded in previous experiments would suggest that A.Z. has knowledge of the location of the countries and cities being probed. If this analysis is correct, one would predict that semantic distance effects would not be observed for places which are merely familiar names but about which less is known. In order to test this hypothesis, an American equivalent of the English cities task was devised. On this occasion, the names of US states were probed using arrays generated on the basis of actual geographical position.
Stimuli
A.Z. was requested to identify the names of US states, grouped using geographical principles similar to those used for the English cities (see Fig. 3B). The stimuli consisted of 16 states, four each from the north, east, southeast and south. These names were arranged into four geographically close arrays (e.g. Wisconsin, Michigan, Illinois, Indiana) and four geographically distant arrays (e.g. Wisconsin, Georgia, Oklahoma, Kentucky). Arrays were presented in an ABBA design.
Results
The percentage correct responses for the close and distant arrays of US states are shown in Fig. 3C. For these less familiar place names, A.Z. showed no sensitivity to their real-world geographical proximity (close = 50/64, distant = 50/64; 2 = 0, P = 1, df = 1). It should be noted that A.Z. recognized all the US state names, but said that she had not visited the country and was unsure of their exact locations. These findings support the notion that the concept of a place must be sufficiently well-specified for its real-world location to have a significant influence on accessing that concept.
|
Error analyses |
|---|
In all of the geographical distance experiments, each stimulus was probed four times. An item-specific analysis to determine on which probe each error occurred was conducted. The percentage error rate on probes 1–4 for the countries, international cities, UK cities, English cities and US states experiments (Experiments 3–6) are shown in Fig. 4.
2 tests revealed a significant difference in the number of errors occurring between probes for all experiments (countries: 2 = 9.57, P < 0.05; international cities: 2 = 16.81, P = 0.001; UK cities: 2 = 16.58, P = 0.001; English cities: 2 = 9.45, P < 0.05; US States: 2 = 9.14, P < 0.05; df = 3 for all tests). In all cases, the number of errors increased with each successive probe. This clearly demonstrates the build-up of refractoriness in the semantic system.
|
|
Discussion |
|---|
The focus of this study is the information processing structures within the proper noun category of place name knowledge. We report the case of a patient whose comprehension of proper nouns was shown to be not only inconsistent but also significantly affected by temporal factors (Experiment 1). Thus, in the context of a refractory rather than a storage deficit, the ability of this patient, A.Z., to identify place names was explored. The results described here provide the first detailed evidence of more fine-grain organization within this domain of conceptual knowledge.
Semantic relatedness was manipulated using two array conditions. So-called close semantic arrays contained items from the same category or grouping. Distant semantic arrays were formed by drawing one item from each different category or grouping. Constructing arrays in this manner also provides an ideal method of controlling for variables such as word frequency and age of acquisition, as the items probed in each condition are identical. At the coarsest level, A.Z. had greater difficulty identifying proper nouns when the items in an array were drawn from the same category (e.g. three place names) than from several categories (e.g. one place, one person, one brand name; Experiment 2). This semantic mediation provided evidence that A.Z.’s refractory deficit affected central semantic representations rather than some more peripheral transcoding process.
Close and distant semantic effects in individuals with a refractory deficit imply that the activation of a representation causes a build-up of refractoriness which also has an effect on the subsequent activation of other representations which are partially supported by the same neural structure. On this basis, the presence of a distance effect may be taken to reflect the structure of the cognitive system underlying performance of the task in question. This notion dictates that at the most crude level, a patient will not be able to show a distance effect if they are not sensitive to the closeness or separation of individual test items. In this vein, categorical distance effects are just as informative as categorical deficit effects in shedding light on the organization of conceptual space. The findings reported here lend further weight to the notion that the study of not only storage but also refractory deficits may contribute significantly to our understanding of the brain mechanisms underlying cognition.
A number of close and distant effects were found in the experiments reported here. In Experiment 3, A.Z. was shown to have more difficulty identifying country names accurately when they were arranged in an array with geographically close countries than with geographically distant countries. A serial position analysis revealed that on this task and others, A.Z. made very few if any errors when identifying items for the first time. Only on subsequent probes were errors observed. This suggests that A.Z. knew, and by extension had an intact representation of, each item.
A.Z.’s greater difficulty with identifying place names from among arrays of proximal rather than distal place names was replicated in Experiment 4. Using city names rather than country names, significant close and distant effects were demonstrated for both international cities and UK cities. Given that all the city names in each close array were drawn from the same country, these effects may reflect the organization of geographical place name knowledge according to regional superordinates, in this case country membership. However, the nature of the superordinates supporting both country names and city names is unclear. Are these concepts associated with verbal superordinates, spatial superordinates or more general geographical superordinates? Taking ‘America’ and England’ as examples, it is commonly known that these names refer to unique countries with different populations, cultures and traditions. Central to our understanding though is that these countries have distinct locations, i.e. America is 
3000 miles due west of England. Although the concept of ‘west’ possesses a verbal label, it is meaningless without reference to a spatial framework or map. Thus, it seems likely that country and city names are supported by knowledge which is encoded in both verbal and spatial terms. As they refer to places, albeit on a larger scale, superordinates such as continent membership may also be encoded in similar terms.
Much stronger evidence for the role of spatial information in knowledge of places can be drawn from the results of Experiment 5, in which A.Z. demonstrated a significant distance effect in her identification of English city names. Unlike the experiments probing comprehension of country, international city and UK city names, the close and distant effect observed with English cities cannot be so readily explained in terms of the availability of superordinate information. In identifying English city names, A.Z. had greatest difficulty distinguishing between cities which were regionally close to one another. This was despite the fact that no formal classifications or boundaries, such as the county in which each city was located, could be used to define membership of the close or distant arrays. Indeed, the arrays were formed purely on the basis of their spatial location, i.e. cities from the north, south, east, west. As mentioned previously, these are predominantly spatial concepts, defined only to a lesser degree by common characteristics (e.g. seaside town, industrial city). Moreover, it would appear that the abstract concept of semantic relatedness (‘close’/‘distant’) has a literal realization in Experiment 5.
How does spatial information influence spoken word–written word matching performance? It is widely held that information about concrete and abstract concepts (e.g. objects, emotions) is stored in the form of either a verbal code, a visual code or both. However, it is difficult to see how geographical information could have a strong influence on the comprehension of place names if knowledge of those words was based solely on one of these codes. Thus, in this instance, it would be reasonable to invoke a spatial code or framework. Such coding might take the form of individual topographical representations of well-known places, accessed either directly or by reference to a larger topologically delimited representation.
That place knowledge is based on a distinct spatial code or framework would be entirely consistent with the highly selective preservations of knowledge about countries (e.g. McKenna and Warrington, 1978; Warrington and Clegg, 1993). It is particularly striking that such individuals can demonstrate virtually no verbal knowledge and yet still remain able not only to comprehend but also to produce country names accurately. Taken together with the present findings, these cases indicate that there is something unique about this class of knowledge. Indeed it may be that the type of spatial framework described here represents a specific module within conceptual knowledge.
Some clarification of the mechanism by which such a spatial framework would be developed comes from comparing and contrasting the findings of the English city and US state names experiments. In Experiment 6, no distance effect was observed in A.Z.’s identification of proximal or distal US state names. Her response accuracy was equivalent whether attempting to identify state names from among close or distant arrays, even though the names were organized following the same spatial principles as for the English cities. It seems to be the case that A.Z.’s representations of US state names were not as greatly enriched by a spatial framework as the English city representations. Her near faultless performance on items when probed for the first time indicates the intactness of these place name representations. Also, by her own admission, the US state names were familiar to her but she did not know where in America they were located. In this sense, A.Z. possesses only ‘partial knowledge’ of these place names.
Overall, it appears that when acquiring knowledge about place names, we develop semantic representations of many different places, in the sense that the names become familiar and enter our vocabulary. All such place names are based on a combination of verbal, visual and spatial information. Although it is an empirical question, we would speculate that of these contributions, spatial information has primacy in the acquisition of individual place concepts. However, a detailed spatial framework is not developed for all geographical regions. The availability of such information is largely dependent on education (e.g. book learning of maps) and to a lesser extent personal experience (e.g. perceived distance when travelling between two locations). Consequently, we would predict that a similar refractory disorder in an American patient would produce a converse pattern of performance on the current English and US geographical proximity experiments. In this context, the study of our understanding of proper nouns provides an opportunity to explore the interface between personal and culturally universal semantic knowledge. For those portions of semantic knowledge which are complemented by enhanced spatial information, the spatial framework in turn has consequences for the central organization of conceptual knowledge in the broadest sense. Thus, in the further development and specification of some place concepts, it appears that actual geographical location and proximity are important organizational principles.
In conclusion, the present investigations highlight a number of ways in which studying dynamic refractory deficits in addition to static storage deficits can expand our understanding of the nature of conceptual knowledge. First, theoretical accounts of the categorical organization of conceptual knowledge must be extended to incorporate the divisions within the proper noun fields described here. Secondly, the study of proper names in particular may aid us in elucidating the interaction between personal experience and commonly shared conceptual knowledge. Thirdly, a spatial code may be of primary importance in the development of place concepts.
|
Acknowledgements |
|---|
We wish to thank Dr D. Cohen for allowing us to study a patient under his care, Katie Cohen for her assistance and provision of language assessment information, and Dr J. Stevens for his assessment of the neuroradiological data.
|
References |
|---|
Cipolotti L, Warrington EK. Towards a unitary account of access dysphasia: a single case study. Memory 1995; 3: 309–32.[Web of Science][Medline]
Coughlan AK, Warrington EK. The impairment of verbal semantic memory: a single case study. J Neurol Neurosurg Psychiatry 1981; 44: 1079–83.
Forde E, Humphreys GW. Refractory semantics in global aphasia: on semantic organisation and the access–storage distinction in neuropsychology. Memory 1995; 3: 265–307.[Web of Science][Medline]
Incisadella Rocchetta A, Cipolotti L, Warrington EK. Countries: their selective impairment and selective preservation. Neurocase 1998; 4: 99–109.[CrossRef][Web of Science]
Kay J, Lesser R, Coltheart M. PALPA: psycholinguistic assessments of language processing in aphasia. Hove (UK): Lawrence Erlbaum; 1992.
McKenna P, Warrington EK. Category-specific naming preservation: a single case study. J Neurol Neurosurg Psychiatry 1978; 41: 571–4.
McNeil JE, Cipolotti L, Warrington EK. The accessibility of proper names. Neuropsychologia 1994; 32: 193–208.[CrossRef][Web of Science][Medline]
Patterson K, Hodges JR. Deterioration of word meaning: implications for reading. Neuropsychologia 1992; 30: 1025–40.[CrossRef][Web of Science][Medline]
Semenza C, Zettin M. Generating proper names: a case of selective inability. Cogn Neuropsychol 1988; 5: 711–21.[Web of Science]
Semenza C, Zettin M. Evidence from aphasia for the role of proper names as pure referring expressions. Nature 1989; 342: 678–9.[CrossRef][Medline]
Warrington EK, Cipolotti L. Word comprehension. The distinction between refractory and storage impairments. Brain 1996; 119: 611–625.
Warrington EK, Clegg F. Selective preservation of place names in an aphasic patient: a short report. Memory 1993; 1: 281–288.[Medline]
Warrington EK, McCarthy R. Category specific access dysphasia. Brain 1983; 106: 859–78.
Warrington EK, McCarthy RA. Categories of knowledge. Further fractionations and an attempted integration. Brain 1987; 110: 1273–96.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. M. Mitchell, S. V. Shinkareva, A. Carlson, K.-M. Chang, V. L. Malave, R. A. Mason, and M. A. Just Predicting Human Brain Activity Associated with the Meanings of Nouns Science, May 30, 2008; 320(5880): 1191 - 1195. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Jefferies and M. A. Lambon Ralph Semantic impairment in stroke aphasia versus semantic dementia: a case-series comparison Brain, August 1, 2006; 129(8): 2132 - 2147. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Crutch and E. K. Warrington Abstract and concrete concepts have structurally different representational frameworks Brain, March 1, 2005; 128(3): 615 - 627. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||