Grammatical encoding in aphasia: evidence from a "processing prosthesis".

Agrammatic aphasia is characterized by severely reduced grammatical structure in spoken and written language, often accompanied by apparent insensitivity to grammatical structure in comprehension. Does agrammatism represent loss of linguistic competence or rather performance factors such as memory or resource limitations? A considerable body of evidence supports the latter hypothesis in the domain of comprehension. Here we present the first strong evidence for the performance hypothesis in the domain of production: an augmentative communication system that markedly increases the grammatical structure of agrammatic speech while providing no linguistic information, functioning merely to reduce on-line processing demands.

Effects of morphological complexity on phonological output deficits in fluent and nonfluent aphasia.

This study examined the effects of morphological complexity on aphasic speakers with lexical-phonological output deficits. Subjects were two fluent and two nonfluent aphasic speakers who repeated morphologically simple words at the same level of accuracy and whose errors were virtually all phonological in nature. They were asked to repeat a variety of morphologically complex (i.e., affixed) words. Results were interpreted within our two-stage model of lexical-phonological production (Kohn & Smith, 1994, 1995), which we expand to include a distinction between derivational and inflectional morphology. When comparing overall performance levels between morphologically simple and complex words, only three subjects exhibited more difficulty repeating the morphologically complex targets. Nevertheless, when comparing repetition accuracy between different types of morphologically complex words (e.g., derived vs. inflected), all four subjects displayed similar patterns. These findings suggested that while morphological complexity has different effects on the two stages within the lexical-phonological output system, the relative effects of different morphological structures are constant. At the level of error analysis, patterns of affix errors distinguished the nonfluent from the fluent subjects in ways that are reminiscent of the affix errors associated with agrammatic and paragrammatic speech. This finding raised questions concerning the relationship between morphosyntactic and morphophonological deficits.

Impaired spelling in Alzheimer's disease: a linguistic deficit?

Different kinds of real words and pronounceable pseudowords (PWs) were presented for writing to dictation to patients with the diagnosis of probable Alzheimer's disease (AD) and to age- and education-matched healthy controls. Though spelling less accurately on all tasks, AD patients responded in a manner generally qualitatively similar to controls. Except for a slightly enhanced effect of spelling regularity in real word writing accuracy, AD patients showed the same sensitivity to various lexical, orthographic and phonological variables as controls. Both groups showed no difference in spelling accuracy for words and PWs with regular vs ambiguous spelling patterns, and groups also showed similar orthographic preferences when spelling PWs having several different acceptable pronunciations. Finally, AD patients and controls produced similar types of errors when spelling real words. Dementia severity was related to the overall accuracy, but not to the pattern, of spelling responses. It is suggested that the decline in response accuracy in cognitively demanding writing tasks in patients with more advanced dementia is most likely due to semantic impairment and impairments of nonlinguistic functions of attention, executive control and praxis, rather than to a disturbance within language specific processes.

Cognitive mechanisms for processing nonwords: evidence from Alzheimer's disease.

Repetition and reading of various types of pronounceable nonwords (pseudowords) was examined in patients with probable Alzheimer's disease (AD) and healthy elderly controls. Overall accuracy of performance was lower in AD patients compared to controls, but the two groups showed qualitatively similar response patterns when reading different kinds of pseudowords aloud and when repeating pseudowords composed of familiar phonological forms, analogous to those in real English words. AD patients diverged in performance from controls, however, when repeating pseudowords composed of phonologically unusual forms. These results support two conclusions: (1) Aspects of phonological processing may become disrupted in AD patients in association with increasing dementia severity, while orthographic processing remains comparatively less impaired. (2) The results are consistent with the view that the processing of pseudowords is achieved through the same system as real words, and further show that the influence of prior language experience on the processing of novel linguistic forms occurs primarily at the level of phonological, rather than orthographic processing.

Repetition of single words and nonwords in Alzheimer's disease.

Repetition of single words and pronounceable nonwords (pseudowords) was assessed in Alzheimer's Disease (AD) patients to evaluate how lexical phonological processing might be accomplished when semantic and conceptual knowledge is impaired. AD patients performed significantly worse than healthy elderly controls on all repetition tasks. However, repetition abilities and dementia severity were not correlated, and AD patients produced the same distribution of error types as controls. Furthermore, despite their semantic problems, AD patients, like controls, showed a significant advantage for repeating real words compared to pseudowords, even when repeating low frequency phonologically complex words whose meaning is not likely to have been retained. The results support the postulated existence of a lexical phonological system that is used to repeat both known and novel words and that processes linguistic information independent of its meaning.

Differential recovery from impairment to the phonological lexicon.

Word production was examined in four aphasics diagnosed acutely with neologistic jargon and who displayed impairment to the lexical stage of phonological production (Kohn & Smith, 1993, 1994a). To investigate the major source of their nonword errors (i.e., neologisms, phonemic paraphasias), single word production was tested at three different times over the first 6 months postonset, with one subject receiving additional testing at 14 months postonset. Two subjects showed signs of recovery to the phonological output system with respect to: (1) improved word production scores, (2) increased frequency of phonemic paraphasias versus neologisms, and (3) increased production of target phonemes. These subjects also displayed above-chance production of target phonemes and no significant tendency to perseverate phonemes across picture-naming trials. It was argued that this pattern reflects a resolving disturbance in retrieving entries from the phonological lexicon. The other two subjects showed no improvement in word production. They also consistently produced target phonemes at chance levels and had a tendency to perseverate phonemes across picture-naming trials. This static pattern of performance was considered to reflect loss of information from the phonological lexicon. The neuroanatomical damage sustained by each case was consistent with these two recovery patterns.

Consonant harmony as a compensatory mechanism in fluent aphasic speech.

This study addresses how fluent aphasics construct complete phonological representations, given the premise that their phonological speech errors result from faulty information about stored lexical representations. We explored whether consonant harmony, a common rule-governed process of feature copying, operates as a compensatory device for completing phonological representations in fluent aphasia. This was examined in a corpus of phonemic paraphasias (n = 543) produced by 8 fluent aphasics during picture naming. Consonant substitutions due to a single feature change (n = 143) were analyzed for the properties of consonant harmony predicted by the phonological principles embodied in a Universal Markedness version of Underspecification Theory (e.g., Chomsky and Halle, 1968). Results indicated that harmony constrained the feature substitution errors involving the feature class of voice (e.g., calendar-->/[symbol: see text]/), but not place of articulation (e.g., igloo-->/idlu/); substitutions due to an error in manner were rare. These findings were used to argue that for English-speaking fluent aphasics a consonant harmony rule for the feature voice is incorporated into a compensatory output mechanism that is used to complete faulty lexical-phonological representations.

Lexical-phonological processing of functors: evidence from fluent aphasia.

The purpose of this study was to investigate how functors are processed at the lexical-phonological level. Four fluent aphasics with impaired ability to fully activate entries in the phonological lexicon, as revealed by their noun production, were asked to repeat and read aloud individual functors. Similarities between their functor and noun production (e.g., level of performance, production of phonological errors) indicated that functors are stored along with content words in the phonological lexicon, and are stored with a similar abstract structure. We also considered whether the sentence-planning system can facilitate the phonological encoding of functors presented in isolation.

The relationship between oral spelling and phonological breakdown in a conduction aphasic.

Unsolicited oral spelling has been observed in conduction aphasics during word-production difficulty (Kohn, 1985), but has never been examined systematically. In the present study, the possible role of oral spelling in the word-production process of a conduction aphasic (JM) is explored. Oral spelling occurred much more frequently during picture naming than during either oral reading or repetition, and it seemed to be operating as a strategy to lessen the phonemic output deficit often observed in this syndrome.

The remediation of conduction aphasia via sentence repetition: a case study.

Repetition is viewed typically as a primary deficit in conduction aphasia. Consequently, repetition is considered to be a target of treatment, as opposed to an approach to treatment for this syndrome. However, our experience with a conduction aphasic who had greater speech fluency in repetition than in conversation encouraged us to develop a treatment programme based on sentence repetition. The treatment programme has three components: (1) a pre-/post-test of sentence repetition; (2) the treatment component, in which a new set of sentences is introduced each week in outpatient therapy for daily repetition; (3) analysis of verbal picture description prior to and following treatment as an independent language measure. The patient CM participated in 2 months of sentence repetition treatment, after which he displayed improvement in both sentence repetition and picture description. The reasons for such improvement and the general utility of the treatment programme are discussed.

Impaired activation of the phonological lexicon: effects upon oral reading.

The role of the phonological lexicon in oral reading is examined in a patient with a small focal left hemisphere lesion. Impaired access to the patient's phonological lexicon is suggested by a number of findings, including the production of phonemic errors across a variety of tasks; increasing difficulty in word production with increasing word length; and difficulty on tests of homophone and rhyme judgments. Two competing models of reading are tested: the nonlexical ("rules") and the lexical ("no-rules") models. The rules model predicts that a disturbance in the phonological lexicon will result in surface alexia; the no-rules model predicts phonological alexia. Results indicate that the patient's reading is most similar to phonological alexia, providing support for the no-rules model. The applicability of the no-rules model to other forms of acquired alexia is explored.

Verb finding in aphasia.

Word finding for nouns and verbs was examined in a heterogeneous group of aphasics (N = 9) by comparing the ability to generate synonyms and sentences for the same set of 20 nouns and 20 verbs. Synonym Generation performance resembled that of an age-matched group of normal control subjects (n = 9): In both groups, some subjects produced comparable numbers of synonyms for nouns and verbs while other subjects produced significantly fewer synonyms for verbs. Essentially the same two patterns were displayed on Sentence Generation using the frequency of "empty" nouns (e.g., 'it', 'man') and "empty" verbs (e.g., 'is', 'do') as an index of word-finding difficulty: In both groups, some subjects produced comparable numbers of empty nouns and verbs, while other subjects produced significantly more empty verbs. However, the Sentence Generation performance of one aphasic subject stood out overall by her tendency to avoid empty verbs and produce incomplete sentences. This pattern of performance was interpreted as a breakdown in an early stage of sentence planning that may be directly related to her diagnosis of transcortical motor aphasia.

Picture-naming in aphasia.

The distribution of picture-naming errors for Broca's aphasics (n = 9), Wernicke's aphasics (n = 9), conduction aphasics (n = 9), frontal anomics (n = 7), and posterior anomics (n = 9) was examined to determine the diagnostic power of error types in picture-naming. Negated responses were associated with Broca's aphasia, whole-part errors ("hose" for nozzle) were associated with frontal anomia, and poor phonemic cuing was associated with Wernicke's aphasia. In addition, the relative distribution of the three most prominent naming errors-phonemic errors, semantic errors, and multiword circumlocutions-tended to distinguish the two anomic subgroups from the other aphasia subgroups. Anomic aphasics produced the fewest phonemic errors and the most multiword circumlocutions; this pattern suggests minimal word-production difficulty in anomic aphasia relative to the other aphasia syndromes. Despite such group differences, the overall picture indicates that there is considerable similarity among aphasia syndromes in terms of picture-naming behavior.

The nature of the phonological disorder in conduction aphasia.

Sequences of attempts to name pictured objects were used to examine phonological dysfunction in three diagnostic subgroups of aphasia. A prevalence of "phonologically-oriented sequences" (i.e., those sequences that contained only attempts with a phonological resemblance to the target word) was found to be a diagnostic criterion for conduction aphasia. When compared to a group of Broca's (n = 7) and Wernicke's (n = 5) aphasics, all the conduction aphasics (n = 6) produced proportionately more of such phonologically-oriented sequences on a picture naming test. An examination of the phonologically-oriented sequences produced by the conduction aphasics indicates that speech production in conduction aphasia involves dysfunction at an early stage of sound-encoding. The theoretical implications of this view are discussed.