Für neue Autoren:
kostenlos, einfach und schnell
Für bereits registrierte Autoren
Veröffentlichen auch Sie Ihre Arbeiten - es ist ganz einfach!Mehr Infos
Magisterarbeit, 2005, 113 Seiten
2 Brief Overview
2.1 The Aphasias
2.1.1 Amnestic Aphasia
2.1.2 Broca’s Aphasia
2.1.3 Wernicke’s Aphasia
2.1.4 Global Aphasia
2.2 Dementia of the Alzheimer’s Type (DAT)
2.2.1 Communication Deficits in DAT
3 Word Retrieval: A Theoretical Account
3.1 Multiple versus Unitary Semantics
3.1.1 The Dual-Coding Hypothesis
3.1.2 The Organized Unitary Content Hypothesis (OUCH)
4 Naming in Alzheimer’s Disease
4.1 Category-Specific Deficits
4.1.1 The Sensory-Functional Theory
4.1.2 The Domain-Specific Model
4.1.3 The Connectionist Account
4.1.4 Neural Substrates
4.1.5 Neuropsychological Findings
4.2 The Role of Uncontrolled Stimulus Variables
4.3 Tip-of-the-Tongue States: Impaired Access or Degradation of Semantic Knowledge?
4.3.1 Famous Faces
4.4 Priming in DAT
4.4.1 Theoretical Perspectives
4.4.3 Automatic versus Attentional Priming
4.4.4 A Matter of Inhibitory Control? Further Network Suppositions
4.5 Visual Capacity in DAT
4.5.1 Visual versus Semantic Deficits
4.5.2 The Role of Stimulus Quality
4.5.3 Bottom-Up and Top-Down Mechanisms
5 Naming in Aphasia
5.1 Error Patterns
5.1.1 Aphasia versus DAT: A Comparative Naming Study
5.1.2 On the Validity of Category Effects
5.1.3 Semantic Hierarchies
5.2 Priming and Cueing Revisited
5.2.1 Lexical Competitors and Phonetic Structures
5.2.2 Summation Priming: A Lexical Decision Paradigm
5.2.3 Facilitatory Effects on Word Retrieval
6 Summary and Outlook
7 Index of Figures
The highest activities of consciousness have their origins in physical occurrences of the brain just as the loveliest melodies are not too sublime to be expressed by notes.
-- William Somerset Maugham, A Writer's Notebook, 1902.
Language is one of the pillars of human consciousness, the main implement whereby individuals articulate thoughts and convey them to others. Impairments of the brain that affect this requisite part of the human intellect―stroke, head injuries, progressive conditions or developmental disorders―not only disrupt a person’s ability to carry out daily activities, but can also cause severe damage to the emotional state of a person and put a tremendous strain on caregivers.
Pathologia physiologiam illustrat: For more than a century, it has been primarily by evidence from speech errors that neuroscientists have tried to understand the mechanisms of how the brain learns, stores, and processes language. Owing to the brain’s structural and functional complexity, many tasks remain unresolved. These include localizing language areas within the two hemispheres, mapping language-related functions, understanding how they relate to each other, and in what way their damage contributes to different degrees and varieties of disabilities. More specifically, the main problem remains the lack of a one-to-one correspondence between specific mental processes and cortical regions.
Modern neuroscientific measurement techniques that aim at extracting quantitative information about physiological functions from image-based data may offer great insights into localized brain activity for specific cognitive tasks. For example, fMRI ( functional magnetic resonance imaging ) noninvasively highlights potentially significant language areas by monitoring regional changes in blood oxygenation, while PET (positron emission tomography ) can measure abnormalities in the patient’s cerebral glucose metabolism.
However, the results are highly variable from one study to the next, as there often is no single cause or pathologic mechanism, which particularly holds true for degenerative diseases.
Hence, it is premature to think of functional imaging as a reliable tool to accurate diagnostic analysis as of yet, let alone grant it the power to pinpoint the internal structure and functions of the neural subsystems involved in normal language processing. A major advance in our understanding of the cortical organization of language at the systems level has not been achieved so far (Wise, 2003, p. 97).
It is still primarily by a number of traditional pen-and-paper neuropsychological tests that attention, perception, memory, and speech production are assessed and conclusions are being drawn about the inner mechanisms of language processing in healthy individuals in general and their altered performances under pathological circumstances in particular and the other way around. As Nickels put it, “The use of neuropsychological data rests on the assumption that patterns of language breakdown reflect the structure of the normal language system” (Nickels, 1997, p. 7).
The goal of this Magisterarbeit is to provide a critical review of the present observations in peer-reviewed literature on two well-studied but still controverted morbid conditions respectively: aphasia, a syndrome most commonly caused by a cerebrovascular event, and Alzheimer’s disease, or dementia of the Alzheimer’s type (DAT), a widely distributed, degenerative cortical disease among seniors with an unknown etiology. However diverse the symptoms of both conditions may be, they share some language impairments in the area of spoken-word production. One of the impairments, impaired single word retrieval in naming tasks, will be the focus of this investigation.
This comparison is supposed to shed some light on the likelihood or unlikelihood of these two syndromes regarding lexical access. My ultimate goal is to grasp the inner nature of what essentially determines lexical access in aphasics and patients with DAT. Marshalling evidence from priming and other testings, I will engage in a discussion of the internal structure of each (assumed) component in the language system, the kinds of representations produced at each stage, and the interaction among components and subcomponents.
The structure of this thesis is as follows:
Chapter two briefly outlines the nature of neurophysiological damage and language disturbances in aphasics and patients with DAT.
Chapter three is fitted in as an account of theoretical issues, briefly elucidating the complexity of the naming process and the boundaries of theory.
The following sections constitute the main part of this thesis, focusing on a number of patterns of impaired speech output, first in Alzheimer’s disease (section 4) and finally in aphasia (section 5), on the basis of different results of examinations. The goal always is to draw conclusions of their underlying causative processing deficits, and incorporate the variables affecting the production of these errors.
Furthermore, I will be concerned with the quantitative and qualitative differences in naming performances. For instance, aphasics show improvement when provided with verbal or orthographic cues, while Alzheimer’s patients hardly do so at all. Does this finding provide sufficient evidence for the assumption that the concepts in their minds simply fall victim to the decay of brain cells and tissues (atrophies)? Different studies report on Alzheimer’s patients showing improvement in their word finding after an intensification of visual input. Are there other, nonlinguistic factors, such as visual, perceptual, and attentional selective impairments that lead to such results?
These are a few of the questions I intend to tackle in the following analyses, which are strictly theoretical in nature, giving an account of my investigations in the relevant literature.
Gandhi’s saying, “Honest disagreement is often a good sign of progress,” will finally lead me to expound my own conclusions regarding the main themes in the last sections of this thesis.
The space given obviously does not allow one to exhaustively present all controversial issues currently underway, so no claim is made as to the completeness of these overviews.
I will (for the most part) limit the choice of studies to the latest, and from my point of view to the most interesting available. Controversial empirical and theoretical issues will arise along the way, many of which will remain unsatisfactorily resolved.
Classical aphasia is an acquired, sudden language disorder caused by damage to the left cerebral hemisphere. Most aphasias originate from apoplectic disturbances (strokes). Other causes are head traumata, cerebral tumors, and infections (Tesak, 1997).
The term aphasia used here comports with the typology established by the European science community. The term refers strictly to the clinical syndrome of language disturbances after focal lesions. This narrow definition is in contrast to the American literature, in which aphasia refers to language disorders produced either by a focal lesion or by widely multifocal or diffuse lesions. This conceptual discrepancy dates back to the time in which Carl Wernicke’s and Alois Alzheimer’s insights about language in aphasia and dementia shaped the views of the research community of today (for more details see Mathews et al., 1994).
Aphasia causes problems with both receptive and expressive functions and displays impairments with all language modalities (speaking, writing, reading, and comprehension), at all linguistic levels (phonology, morphology, syntax, and semantics), and in all linguistic units (phoneme, syllable, word, and sentence). About 400,000 people in Germany suffer from aphasia; 70,000 are stroke patients (Bundesverband Aphasie e.V.).
There are four classical varieties of aphasia formerly believed to correspond to the very discrete location of brain injury. Today, ample evidence can be provided that the lesions causing aphasia are more complex and multifocal and many classical anatomical assertions have been abandoned in favor of new ideas.
Nonetheless, I would like to outline briefly the four classical aphasic syndromes, not only to pay tribute to the scholars of the past whose accomplishments cannot be denied, but also to give an overview of the symptomatic characteristics of these syndromes which have persisted despite all controversies.
Amnestic aphasia is the lightest form of the four main types of aphasia, but it can be localized with the least dependability of all. It is said to be probably caused by damage to the posterior temporal lobe or to the lower area of the parietal lobe, as well as the angular gyrus, or the temporoparietal border areas (Tesak, p. 45).
The main symptom associated with amnestic aphasia is the difficulty in finding words, particularly nouns and verbs. While patients are able to carry on a short conversation fluently and without any comprehension deficits, they are left with a persistent inability to supply the words when asked to name things. To cover up their problem, they tend to make extensive use of circumlocutions. However, their speech remains vague and contains little information. Semantic paraphasias with little divergence from the target are frequent (Huber et al., 1983, p. 8).
The French physician Pierre-Paul Broca (1824–1880) first described the syndrome now bearing his name. It was assumed to be caused by a lesion affecting both gyrus and sulcus of the frontal lobe. CT studies from the late 1970s and 1980s showed that destruction of much more than Broca’s area must occur for this type of aphasia (Kirshner, 2000, p. 125), which exhibits great deficits on the expressive side of language. Their output seems laborious, as it is hesitant with long pauses and lacks syntactical structuring; the prosody is disrupted. Phonematic paraphasias display frequently. One major symptom of Broca’s aphasia is agrammatism (or telegram style), a reduction of utterances to a few words, usually to verbs and nouns, while function words are generally excluded. The naming of objects after visual presentation is poor (Huber et al., p. 11). The aural comprehension level is less impaired. Broca’s patients are well aware of their difficulties and often feel very self-conscious.
The Wernicke area, named after the German neurologist and psychiatrist Carl Wernicke (1848–1904), is located in the left posterior section of the superior temporal gyrus (Tesak, p. 45). The syndrome thought to be related to the injury of this area is characterized by a chief impairment in spoken-word comprehension (particularly at the semantic level) as well as extremely fluent speech yet without much informative purpose. By the release of verbal torrents (logorrhea), the patient often gets caught in neologistic jargon, devoid of any content, and freely uses verb tenses, clauses, and subordinates. Paraphasias are common. Like Broca’s aphasics, individuals diagnosed with Wernicke’s aphasia have difficulty in naming tasks. Contrary to Broca’s patients, they usually lack awareness of their disorder (anosognosia).
Global aphasia, the severest and most stable type of all aphasias, usually correlates with large lesions in the whole perisylvian region in the dominant hemisphere, reaching as deep as the underlying white matter and involving the middle cerebral artery (Tesak, p. 45). Communication becomes almost impossible, as all aspects of language are affected (some more than others). Patients are deprived of their ability to understand spoken and written language as well as to express themselves adequately. Complete sentences are unimaginable. Instead, speech is compressed and telegraphic in style. In fact, merely single words and short, recurring utterances and automatic phrases are left in the patients’ inventory. Articulation is distorted, writing and reading abilities are almost completely absent. Naming is severely disturbed. Sadness over their personal tragedy may result in depression, but their level of awareness often has them respond well to therapy. Nevertheless, therapeutic measures hardly have an effect due to the severity of the disturbances.
In addition to these classical syndromes, the nomenclature contains nonperisylvian aphasias, including transcortical sensory, transcortical motor, and transcortical mixed aphasias, which will not be covered in this investigation.
As stated above, the clinical syndromes cannot be assumed to be anatomically or functionally homogenous entities. Poeppel and Hickok (2004) put it this way:
“Instead of thinking of aphasia as a largely fixed set of symptoms that could be explained straightforwardly by damage to a single computational system, we appreciate that clinical aphasic syndromes are comprised of variable clusters of symptoms” (p. 4).
Today, we are aware of a much more complex architecture of brain lesions and language deficits. A variety of studies proved that other areas outside the classical regions are implicated in language processing.
Moreover, it became apparent that the number of instances where the observed lesion is at odds with the aphasic syndrome, a fact which is sufficient to undermine confidence in the localizing of language disorders. Furthermore, cognitive neuropsychology has levelled criticisms at the classical taxonomy by claiming that there is no invariant pattern shared by all members of one particular type of aphasia, but instead the assignment to one case or the other is based on some kind of family resemblance (Basso, 2000, p. 16).
The following paragraph introduces the second neural impairment under investigation, dementia of the Alzheimer’s type, in a few sentences, before I finally engage in the elucidation of the specific deficit of word retrieval.
Methinks I should know you, and know this man;
Yet I am doubtful; for I am mainly ignorant
What place this is; and all the skill I have
Remembers not these garments; nor I know not
Where I did lodge last night.
-- William Shakespeare, King Lear.
Life expectancy in basically all developed Western societies rose rapidly in the past century, and it is still rising, thanks to improvements in medical care, public health, and nutrition. With rising life expectancy there has arisen a more frequent occurrence of the most common degenerative vascular disease of aging: Alzheimer’s disease or dementia of the Alzheimer’s type (DAT).
This form of dementia was named after the German neurologist Alois Alzheimer, who, in 1906, reported on changes in the brain tissue of a woman who had passed away after a peculiar mental illness.
He found anomalous clumps of protein (now called amyloid plaques) and bundles of fibers, today known as neurofibrillary tangles (Bauer, 1994, p. 2). These plaques and tangles in the brain, along with a diffuse loss of volume, particularly around the hippocampus area (Wick et al., 2000, p. 1572), are considered the hallmarks of DAT. There are several hypotheses about the causes of DAT. The main risk factor is old age, and with it the decline of the immune system. Among persons 65 or older, 9% are diagnosed with DAT. Among seniors 95 or older, the percentage rises to 43%. (Hock, 2000, p. 14). Another contributor could be an increased amount of mercury in the brain. The latest studies, however, proved this to be rather unlikely (Bauer, p. 49). The genetic hypothesis suggests that there is a higher risk for other family members if Alzheimer’s disease strikes someone at a younger age. Nonetheless, only about 7% of all Alzheimer’s cases are attributed to genetics, which makes it a rather weak genetic disorder (Hüll, 2000, p. 25). In Germany and in other industrialized countries, the overall prevalence rate is about 6% to 9% of the elderly population. (Deutsche Alzheimer Gesellschaft). Since no definite causes have been identified as of now, the pathognomonic disease is thought to be of multifactorial etiology.
The condition itself is insidious in onset and develops in a steady manner over a period of 7 years on average (Hock, p. 14). In contrast to aphasia, where spontaneous remission might occur, DAT is an acknowledged irreversible disorder. It is well-known for its steady, stern attack on memory. In the beginning, usually only recent memory is impaired, but as the disease progresses, long-term memory is affected, too. Memory loss, however, is not the only impairment patients afflicted with DAT suffer from. Symptoms extend to cognitive deficits in language, object recognition, and executive functioning. Behavioral symptoms—such as depression, psychosis, agitation, and aimless wandering—are further common characteristics of this neurological disease. The Global Deterioration Scale (GDS) by Barry Reisberg provides a measurement for the manifestations of the progressive cognitive decline. This scale, based on functional ability, is broken into seven stages. Stages 1 through 3 are considered predementia stages. Stages 4 to 7 are the dementia stages, 4 being mild dementia with moderate decline, and 7 severe dementia, over which course all verbal abilities and basic physical functions are lost (Reisberg, Geriatric Resources).
Rate and severity of the decline become severer with each stage of the disease, but they can vary substantially from one individual to the other. Alzheimer’s dementia is diagnosed through extensive neuropsychological and neurophysiological testing, which excludes other neurological, systematic, or psychiatric underlying factors potentially responsible for the deficit. This is complicated by the observation that the deficits observed in early DAT coincide with those observed in normal aging (Belleville et al., 1996, p. 195). Thus, it is difficult to determine whether old age contributes to the development of dementia or whether dementia influences normal aging. This chicken-and-egg conundrum can be solved by the most probable idea of an interaction between the two.
What remains to be elucidated, however, is the question of where normal aging ends, and where dementia of the Alzheimer’s type begins. In addition, an undisputable diagnosis requires postmortem examinations, because of the lack of biological markers. Patients referred to in this paper, although not explicitly indicated, will therefore be thought of as patients with probable dementia of the Alzheimer’s type.
The language of Alzheimer’s patients in mild to moderate stages is globally noninformative, characterized by error prone lexical-semantic processing as compared to a relative sparing of syntax, phonological aspects, and diction (Shimada et al., 1998, p. 57). As stated in Kirshner et al. (1984), “the instrumentalities of language may remain preserved, though no longer in the service of cognition” (p. 28).
Patients experience difficulty following conversation and verbalizing thoughts, and they often digress from the topic. They exhibit a shortened attention span and take more time for all language processing. Naming and word-finding deficits are commonly seen from the very onset. All these symptoms pervasively increase over the course of the disease and can eventually culminate in muteness in the final stages.
A definition of confrontation naming and its cognitive processes as provided in chapter 3 will provide a good theoretical basis for the following empirical data obtained from several studies.
Referring to the world with names is a fundamental operation in the use of language and seems simple and quite unremarkable at first glance. A closer look at the processes underlying lexical access, however, inevitably falsifies this view and reveals an amazing complexity. Numerous theoretical approaches have led to various experimental strategies, yet these have failed to provide universally acknowledged conclusions reconciling with present-day empirical data and thus stand as proof that understanding naming is still an ongoing challenge.
Single-picture-naming tasks (called confrontation naming) with healthy as well as brain-damaged subjects have become very popular, since these tasks are simple yet calling for all major processes of language production. Real language processing, even at the single word level, usually requires (and is likely dependent on) the processing of multiple associations between lexical, semantic, and phonological representations (Milberg et al., 2003, p. 32). On the other hand, this type of language processing certainly is an oversimplification, because many processes regarding grammatical encoding are not typically engaged. However, this is what makes it a lot easier to work with. In addition, the target word that the patient is looking for is supposedly known to the examiner without any ambiguity, which is very different from the spontaneous speech situation. Nonetheless, as Deloche et al. (1996) point out, there is no absolute certainty as regards which particular lexical entry the patient is searching for, since pictures elicit more than one name (p. 106). For the sake of being unequivocal, I will focus on findings obtained from experimental single-word designs (such as confrontation naming, priming tasks, and cueing), not least because we will soon learn that Alzheimer’s disease and aphasia are both highly variable diseases even within trials―that is, response inconsistencies have been observed. Thus, a comparison of the two syndromes in the context of spontaneous speech would become even vaguer than it already is for single-word tasks under laboratory conditions.
An image stimulus—be it a photograph, a simple line drawing, or a pictogram—symbolically mediates a concrete object through physical similarity. That is, recognizing a picture comprises essentially the same cognitive processes as perceiving the object itself (Glaser, 1992, p. 62). I will return to the issue of qualitative differences between objects and pictures in more detail later on.
Recognizing a stimulus item to be named is the first of three broad steps in a successful word-retrieval task. The rapid process of recognition (~62–100 ms per item; Johnson et al., 1996, p. 116) is more or less sequentially followed by the activation of the corresponding (perceptual, semantic, lexical, and phonological) representations within the cognitive system, and finally by the activation of an overt phonological response (Davidoff and De Bleser, 1994, p. 1). This is about the point where the general consensus ends. From here, there are more questions than answers. Which and how many systems are exactly involved in naming? How are its internal processes shaped in our minds? How are the representations or concepts of the world stored in our minds (if they are “stored” at all and not generated according to demand), and what mechanisms underlie their activation? In addition, do we access them according to the modality of input (e.g., spoken language, vision, touch)? As outlined above, the sheer amount of studies embarking on the issue of naming in general and on the above questions in particular have led to numerous promising theories and models. Nonetheless, there exists no generally agreed upon hypothesis applicable to all available data. There are two reasons for this. Either the theory/model itself is not coherent (e.g., exhibiting flaws; too unspecific or not specific enough) and thus contradicts the data, or the way the data have been collected is questionable as it contradicts expectations based on the theory.
I would like to insert a digressive section investigating the dissociative arguments of how sensory information might be coded. I will present two cognitive models, each falling into one of the two major “theory camps” working on coherent accounts as to how sensory information is coded and thus attempt to answer the latter question. The first hypothesis suggests multiple modality-specific subsystems which store and process meanings for all concepts according to their modality (auditory, visual, tactile) and type of information, while the latter postulates an amodal, unitary semantic system. This issue has a long history in research, but no consensus has yet emerged.
Neuropsychological evidence for separate visual and verbal semantic systems was first provided by Warrington (1975) in a study of two visual agnosic patients, one of whom appeared to be able to recognize a visually presented object, but not its name, while the other was able to identify the name of an object but not its visual representation (Bright et al., 2004). Further cases of related dissociations in picture- vs. word-based semantic judgments have supported the notion of neurally distinct, modality-specific semantic representations, one of which is presented in the following section.
In Paivio’s dual coding model (1971), cognition is assumed to be based on two independent, but interconnected, mental representations―one verbal and the other one visual or symbolic. The visual subsystem (in the right cerebral hemisphere) specializes in the representation and processing of non-linguistic, perceptual properties, termed imagens, while the verbal subsystem (in the left hemisphere) stores discrete linguistic word representations as logogens, including auditory and visually presented words. Processing in the visual subsystem is believed to be holistic, while in the verbal type of representation, each entity is described individually in a consecutive manner. The representations are linked within each subsystem through associative connections derived from prior experience.
For example, imagens corresponding to concrete objects are connected in a one-to-many fashion to logogens for their names and vice versa. These connections vary in strength and number according to prior experience with the particular objects and their names.
In the dual-coding view, picture-naming unfolds as follows: The pictorial stimulus initiates activity within the whole nonverbal system in proportion to the structural similarities of the representations to the target representation until the recognition threshold for a particular imagen is exceeded. The stimulus picture is identified. Referential connections then forward activation from the imagen to associated logogens (names) in the verbal representational subsystem. Once a particular logogen receives sufficient activation to exceed its threshold, the production of that specific name is initiated. (Johnson, p. 115).
Paivio’s concept is supported by the so-called modality-effect, which has been found in various neuropsychological and neurophysiological examinations. A study presented by Saffran in 2003, for example, compared the spontaneous responses generated by pictures and word stimuli alike. Their data revealed that pictures and words arouse different associations: Pictures elicited more verbs than words. Words were more likely to call fourth common verbal associates (e.g., lion à tiger; bride à groom), while perceptual attributes of the stimulus were often produced in response to pictures. Saffran suggests that these findings are consistent with a model of semantic organization according to which information is distributed across different sensory, motor and verbal domains, reflecting the manner in which the knowledge was obtained and experienced with the concept or object. For example, manipulable objects like “scissors” are more likely to elicit verb associates (like “cut”), because knowledge of these objects is acquired by means of sensory-motor interaction with the object (p. 1,544).
The diagram on the next page demonstrates Paivio’s model visually. In order to minimize confusion, I have decided to continue using the term verbal stimuli for visually presented words, although we should bear in mind that, as noted before, these stimuli are actually of a visual nature.
The Dual Coding Model (adopted from Rieber)
Abbildung in dieser Leseprobe nicht enthalten
Crottaz-Herbette et al. (2004) recently examined the neural bases of verbal working memory by using fMRI. A quantitative comparison of the modality effects in naming responses to visual and auditory stimuli (a comparison that has, paradoxically, received little attention in the literature) revealed important differences in their internal representations. On the one hand, the superior and middle temporal (auditory) cortex showed significant deactivation during visual working memory; that is, activation was greater during the control than the working memory condition. The lingual, fusiform and inferior temporal (visual) cortex, on the other hand, displayed significant deactivation during auditory verbal working memory (p. 348). These neurological results suggest that different modalities are represented in different brain regions and thus complement the conclusions drawn from the notion of modality-specific conceptual representations and access to them.
One weakness of the modality- or material-specific semantics assumption is to attest unambiguously that the observed dissociation is, in fact, located at the level of conceptual representation, rather than within the presemantic representations or processes necessary for access to the conceptual system (Bright et al., p. 419). Unless each task unequivocally taps into conceptual representations, modality-specific effects may emerge, because of impaired presemantic representations (e.g., visual-structural rather than visual-semantic; cf. chapter on vision, section 4.5).
The distributed representations account is in contrast to an abstract amodal notion, which proposes that all processing routes converge on a single set of conceptual representations common to all input modalities. Such a reductionist account is introduced in the following section.
Basic to Caramazza et al.’s (1990) Organized Unitary Content Hypothesis (OUCH) is a common conceptual system stored in an extensive left-hemisphere network that is accessed irrespective of modality. Some perceptually salient features of visual stimuli, however, are thought to have privileged access to their representations, as they are more important than others in defining the meaning of the stimuli. For example, outstanding parts of an object, such as a computer (e.g., keyboard, screen) directly access corresponding semantic properties (knowledge about keyboards and screens in general). Visually or aurally presented words, on the other hand, will have to activate the semantic lexicon first, which in turn activates the properties that define their meaning. According to this account, privileged accessibility and bias towards particular subcomponents of a semantic representation, are used to explain how modality-specific semantic effects can arise from damage to a unitary conceptual system. (Bright et al., p. 419).
One major drawback of the OUCH is mentioned by its creator himself. The prevalence of category-specific deficits (e.g., for animals) in the brain cannot be unequivocally explained by this hypothesis, as it lacks sufficient specificity (Caramazza and Shelton, p. 19). I will not go into further detail at this point. Another shortcoming of the OUCH was detected in an ERP (event-related potential) study by Kounios and Osman (2002). They express doubt about the validity of the unitary semantics view and report major differences in stimulus processing in visual imagery and verbal-associative function tasks, namely as early as 100 milliseconds after stimulus onset, thereby demonstrating differential input processing for these two groups and thus predicting different neural pathways for inputting information, which is incompatible with a unitary account predicting one single pathway (p. 12).
Until now, the overall picture that has emerged from a vast amount of neuroimaging and neuropsychological literature remains unclear. Although modality-specific accounts seem to be favored, extensive research needs to be done in order to illuminate the nature and location in the brain of an as yet unknown number of semantic stores (Kounios and Osman, p. 18) and to provide unmistakable evidence concerning the factor time, order, and dependability of all involved subsystems. Therefore, modality-specificity will not be discussed further, also because this subject is far beyond the scope of this thesis.
The following main chapters embark on a discussion of the breakdown of word retrieval and production that are common and distressful features in both classical aphasia and dementia of the Alzheimer’s type, becoming particularly apparent in picture- and object-naming tasks. I will for the first part focus on different patterns of these profound word-finding difficulties reported in Alzheimer’s disease studies, all the while trying to offer an explanation of their relevance to broader organization of mechanisms encompassed by word retrieval, before the characteristics of word-retrieval disturbances in aphasia are discussed in more detail.
In comparison to healthy individuals, patients afflicted with dementia of the Alzheimer’s type name fewer objects correctly. This naming deficit, which is called anomia, has been observed in early stages of the disease. Nevertheless, it is not considered a characteristically symptomatic feature of Alzheimer’s disease, since it is also a feature of other neurological syndromes (Schecker and Schmidtke).
In general, the erroneous responses of Alzheimer’s patients frequently relate to the target word in meaning, typically either as a contrast coordinate (e.g., orange for apple) or a superordinate (such as fruit for apple), often including visual proximity (Wallesch and Hundsalz, 1994, p. 599). Kim and Thompson (2004) postulate a similar level of impairment for both the noun and verb class in Alzheimer’s patients. They suggest a bottom-up breakdown in the verb lexicon, paralleling the pattern of loss of semantic information in the noun class. Degradation thus progresses from the specific features at the bottom of taxonomic hierarchies to the basic level, and possibly the superordinate level, at the top of the hierarchies (p. 15).
Phonemic errors are usually few in Alzheimer’s disease, but speakers may make use of interloper words (Laiacona et al ., 1998, p. 410). Circumlocutions describing the target item and nonresponses (such as “I don’t know”) show at significant proportions (Williamson et al., 1998, p. 601). A word of caution regarding nonresponses comes from Barbarotto et al. (1998), who state that the types of errors may change for each stimulus and each patient as the disease progresses (pp. 403–404). In addition, picture-naming errors in a laboratory environment have to be deliberately distinguished from those occurring in day-to-day conversation. DAT patients have been found to exhibit more nonresponses in natural conversation at an earlier stage of the disease, which was accounted for by their relative ease in filling discourse gaps with pauses or passe-partout-words. On the other hand, in a picture-naming task, patients are (typically) granted all the time needed for the lexical search. Nonresponses, therefore, occur here most frequently in later stages of the disease, that is, when the linguistic skills generally have weakened considerably (p. 404).
Another interesting aspect of the word-retrieval disorders in DAT is a possible category-specific impairment, which the burgeoning literature has called attention to lately. Such specific deficits shall be informative in regard to the functional architecture of concept representation in the brain.
Deficits for living things have often been associated with herpes simplex encephalitis, a rather rare neurological disorder characterized by inflammation of the brain (Tyler and Moss, 2001, p. 244). Since dementia of the Alzheimer’s type results in relatively diffuse damage in comparison to herpes encephalitis, it seems reasonable to expect that specific impairments would not occur here (Gonnerman et al., 1997, p. 258). However, quite a few surveys conducted research on naming in the Alzheimer’s type of dementia and have reported on prevailing impairments of living things (e.g., animals) as opposed to nonliving things (e.g., tools) within the category of concrete entities on a range of semantic tasks, including picture-naming, word-to-picture matching, or generation of definitions. The impairment does not always comprise a whole domain, but it can be restricted to a specific category within the living-things domain.
See figure 2 for an example of a patient’s picture-naming performance by object category.
Naming performance by object category; data from Caramazza and Mahon (2003, p. 356)
Abbildung in dieser Leseprobe nicht enthalten
The elucidation of category and concept structures in semantic memory is a hotly debated issue that as of yet has not yielded an adequate explanation for the detailed pattern of semantic deficits observed. The next sections discuss three notions based on different ideas regarding the differentiation of semantic categories.
The sensory-functional theory (initially suggested by Warrington and Shallice, 1984) describes non-living things by their function and living things more by their sensory properties. Damage to the sensory system thus can affect living things as well as equally weighted nonliving categories such as musical instruments or food, which both rely upon sensory features (the smell, taste, and visual information of food, and the typical sound that an instrument produces) (Gainotti, 2000, p. 552). A major threat to this theory is the existence of more diffuse impairments of categories that are not within sensory or functional proximity. A further limitation of the sensory-functional account is the underestimation of functional properties in living things. As Tyler and Moss state, animals have important biological functions, such as running, breathing and eating and are also central to the functional knowledge of the living domain (p. 245). Gainotti underlines this view, admitting that the pattern of categorical impairment often does not respect the living/nonliving distinction. In some cases the effect is narrower; in others, it is wider (p. 551).
As discussed above, Caramazza and Shelton make a strong claim for neural segregation in their domain-specific account. The main hypothesis of this account is that evolutionary pressures have resulted in specialized, innate neural circuits involved in semantic processing of the living and nonliving categories (see also Leube et al., 2001). Thus, in their organized unitary contents hypothesis, there is no reliance on modality specificity. Rather, it is assumed that items belonging to natural-kind classes share a comparatively higher number of attributes, and that strongly correlated properties are represented in adjacent substrates (Caramazza and Shelton, p. 8).
Category-specific deficits in this account involve only those categories that have proved important in the evolution of the human brain. Domains include animals, plant life, and possibly artifacts, for obvious reasons. For example, animals are potential predators as well as food, and plants may be a sort of medicine (p. 20). It can be vital to be able to recognize animals quickly as well as to distinguish among plants for their alimentary, and medicinal values or potential jeopardy to one’s health. Thus, the more evolutionarily important a category, the more “hardwired” the intercorrelated properties could be in the brain (Leube et al., p. 425).
Neither the sensory/functional theory by Warrington and Shallice nor the clear-cut domain-specific neural segregation theory of Caramazza and Shelton adequately take into account the nature and structure of items within their categories. As Tyler and Moss argue, some kinds of properties may be more robust than others, i.e., those that are true of many items within a category are generally more preserved than those that are more specific, and properties that are densely correlated with each other are better preserved than those that are more weakly interrelated. Furthermore, brain damage usually does not selectively impair certain categories of knowledge or property in an all-or-none manner (p. 245).
An alternative approach that takes these issues into account is presented in connectionist models of conceptual knowledge.
In the connectionist account, category and domain structure are presumed to be based on similarity, that is, the degree to which semantic properties overlap (Giffard, 2002, p. 2,045). For instance, the perceptual feature has fur is easily attributable to the domain of living things in general and the category of animals in particular. But the species of birds is not meant here. Further attributes, such as barks or guards house, greatly increase the likelihood that the entity aimed at is DOG.
The principle of this model is that representations are categorized and distinguished from each other by the overlapping and distinctiveness of numerous attributes respectively, leading to “identifiable clusters in semantic space” (Tyler and Moss, p. 246). Semantically related items usually share many features and therefore are likely to be found in a common cluster. Devlin et al. (1998) support this view and explain category-specific disorders by assuming damage to the semantic system affecting random features by reducing the degree to which they are activated (p. 80). A feature such as having stripes is key information when referring to TIGER, distinguishing it from, say, LION. Once this feature is damaged or lost, TIGER becomes most likely misnamed. Conversely, the loss of a less informative or distinctive feature such as has eyes certainly has few behavioral consequences.
The connectionist theory seems to be the most convincing approach to category deficits, as it not only accounts for selective category-specific deficits, but also acknowledges the fuzziness of complex patterns observed in many patients. However, it should be noted that all three models covary. Concepts in the same category tend to share many features and are sometimes similar in their sensory-functional characteristics (McRae and Boisvert, 1998, p. 567). An all-or-none-distribution of semantic representations certainly is unrealistic. Tyler and Moss say that none of the major models of conceptual knowledge can currently account for all of the neuropsychological data, and that the connectionist model is not yet sufficiently well-formulated (p. 248).
The following two studies give an idea of how difficult it is to assign to the category-specific deficits in Alzheimer’s a reasonable explanation of cause and effect from both neurophysiological and neuropsychological perspectives.
A recently conducted PET activation study by Rinne et al. (2003) observed the neural correlates of semantic decision making in patients with early Alzheimer’s disease as compared to age-matched controls. In the main trial, nine DAT patients and eight control subjects were presented with three familiar written words on a computer screen. One of them was shown above the other two.
The assignment was to quickly press a two-choice reaction time button after deciding which word they thought was semantically closest to the target word. After that, the procedure was repeated with words for animals (p. 90). The examiners kept their cognitive tasks very simple to ensure equal levels of success between patients and controls. This method facilitated a direct comparison of brain activation in equally successful semantic decision making.
The task yielded increased regional blood flow in DAT subjects contrasting normal controls, who displayed limited increases. Moreover, larger and partly different regions of the brain were activated in the DAT patients.
Right side view (left) and back view (right), similar to Rinne et al. (p. 93)
Abbildung in dieser Leseprobe nicht enthalten
Figure 3 displays the increases in regional blood flow rendered on a single subject brain model during decision making with animal names (top) and artifact names (below). Task induced activations in patients are depicted with red, activations in control subjects with green. Orange colors represent overlapping areas (as indicated by arrows). Rinne et al. postulate decreased cognitive capacity early in the Alzheimer’s course, which leads to more effort in the task compared to intact individuals and also explains longer reaction times of the patients. Broader and partly different brain areas are thought to be used to compensate for the cognitive weakness in the patients, a mechanism that is reinforced by extensive inferior dorsolateral activation. The authors assume that the activation of these regions could even be related to phonological processing, that is, the patients could have employed inner speech as support during their task performance (p. 94).
One rather unexpected result of the present survey is that the Alzheimer’s patients are reported to exhibit no significant difference in processing animals as compared to artifacts, in fact, the task-related changes for both categories were more similar in the patients than in the controls, who showed specific activation close to the motor cortex exclusively for the artifact nonliving (p. 93). Rinne et al. therefore assume that the compensatory mechanisms in patients were category-independent.
Unfortunately, the examiners provide no further details on the animate/inanimate distinction in their Alzheimer’s subjects. Nonetheless, we have seen that albeit early phase patients and controls were both equally successful in accomplishing the tasks, they made use of different regional and magnitude aspects.
It might be supposed that a possible category-effect could not yet be observed in the present study because of the simplicity of the task or because such an effect only shows at more advanced stages of the disease. Another account would be that there actually exists no semantically induced category-effect at all, neither in early nor in late stages of the disease, and previous studies falsely reported statistical results uncorrected for factors such as frequency or familiarity that facilitate or impede naming/categorization respectively for both animate and inanimate categories independent of their semantic properties. I will return to this issue shortly.
Yet a further explanation is to apply the observed large-scale distribution of cerebrovascular activation not to compensatory mechanisms but to a profound alteration of the semantic system itself possibly making use of more extensive and regionally different brain areas. This is a rather bold assumption that remains to be further examined.
In an earlier study comparing the semantic networks of 33 DAT patients across all stages of the disease and 22 control subjects, Chan et al. (1997) propose quantitative and qualitative changes in the organization of semantic knowledge corresponding to disease severity. As far as the procedure of the testing trials was concerned, the subjects were each presented with a triadic comparison task first and a picture matching task second (p. 242).
The results were then applied to a map representing the organization of semantic knowledge, divided into domesticity, size, and predation. See a sample in the adopted figure below.
 Note that the term concept is still a hotly debated one: Platonists propose that concepts exist independently of minds, whereas naturalists argue that concepts are causal relations between the mind and the world. Following psychological theories, I assume that a concept is knowledge about a particular category (e.g., birds, eating, happiness). Thus, knowledge about birds represents the bodies, behaviors and origins of the respective entities (Barsalou, 2003, p. 84).
2 However, some studies doubt the integrity of the phonological processing and phonetic articulation until the late stages of the disease, as their findings suggested otherwise (e.g., Croot et al., 2000, p. 297).
 I will return to this issue repeatedly.
 Visual agnosia patients are unable to recognize familiar object. Most cases are brought about through cerebral vascular accidents or traumatic brain injury, typically inhibiting sufficient amounts of oxygen from reaching vital body tissues (Farah, 1990).
 Bright et al. (2004) and others (Caramazza and Shelton, 1998, Plaut, 2002) express discontent with most influential investigations, in which visually presented words were taken as representing a sensory modality. Strictly speaking, they conflatingly combine content (e.g., visual/perceptual vs. functional), context (e.g., visual vs. spoken input) and format (e.g., visual/pictorial vs. symbolic/propositional). That is, unlike visual vs. auditory representation, the visual representation of pictures vs. that of words does not constitute a comparison of sensory systems of input (auditory, visual, and tactile), but of objects (or pictures of them) vs. printed words (Bright et al., p. 417).
 Paivio used the concept of the logogen developed by Morton (1969).
 That is, a picture may arouse associated images, such as object or category names, or subsequent word associations (e.g., names of related objects and properties).
 Baddeley and Hitch developed this term to describe the short-term memory system, which is involved in the temporary processing and storage of verbal information (Gathercole and Baddeley, 2003).
 Neighboring phonological forms.
 Or even for each trial session with the same patient.
 Musical instruments are an especially complex case. Within the framework of the sensory-functional theory, musical instruments are considered more similar to living things, that is, classical musical instruments are not as easily distinguishable by their function since all are used for producing sounds. However, their function and perceptual aspects are strictly intermingled as in nonliving categories, since shape, size and sound type are largely independent (Barbarotto et al., 2001, p. 406).
 I consider it noteworthy that such a small amount of data derived from nine DAT patients will not arrive at statistical significance. The results presented here certainly need further verification by larger corpora because of the overall fair variability in dementia of the Alzheimer’s type. As Barbarotto et al. put it: “Naming in Alzheimer’s disease is more a probabilistic event. The same latent status of naming may give rise to different responses at different times” (p. 403).
 The subjects faced three animal names simultaneously and had to indicate the two that were most alike.
 Twelve black-and-white drawings of animals were presented that they then had to match to cardboards with the names of the animals just encountered.
Diplomarbeit, 73 Seiten
Diplomarbeit, 123 Seiten
Diplomarbeit, 188 Seiten
Diplomarbeit, 115 Seiten
Diplomarbeit, 193 Seiten
Masterarbeit, 79 Seiten
Doktorarbeit / Dissertation, 556 Seiten
Diplomarbeit, 92 Seiten
Diplomarbeit, 71 Seiten
Diplomarbeit, 86 Seiten
Diplomarbeit, 153 Seiten