|
|
Project 2: Interactive Processes in Language: Lexical Processing
The proposed research concerns the cognitive functions and underlying neural structures involved in word-level processing tasks, such as reading aloud or forming the past tense of verbs, and how this processing can be impaired by brain disease or injury. The communication deficits that result from impairments to these processes are among the most crippling disabilities and therefore deserving of increased research effort. Complex patterns of performance in normal and impaired lexical processing have led other researchers to postulate numerous sub-systems specialized both for different language tasks and for different word classes. By contrast, our working hypothesis is that the data are better explained in terms of graded activations and interactions of distributed representations within three basic forms of word knowledge: semantics (word meaning), phonology (word sounds), and orthography (word spellings).
The specific hypotheses guiding the planned research are that a) linguistic knowledge is shaped by the language user.s experience of the distributional characteristics of the relevant vocabulary; b) the efficiency of lexical processing depends on the frequency of a given word and the consistency of its input-output mapping with those of other words; c) all items are processed by the same basic mechanisms rather than by discrete procedures specialized for different subsets of items; and d) differences in performance on subsets of items are attributable to differential demands they place on semantic versus phonological components of language processing, which are independently disruptable by brain damage.
The research will comprise a) behavioral experiments on word-production tasks with normal speakers and neurological patients who have language deficits as a result of a degenerative brain disease or a stroke; b) structural brain imaging studies (MRI) in patients, to determine the location of their lesions; c) functional brain imaging studies (PET or fMRI) in both normal participants and patients to determine the location of regional cerebral activation when people produce words; d) comparisons of data from sections (a), (b) and (c) in English-speaking participants with results from parallel studies in other languages; and e) computational connectionist modeling of these tasks to simulate accumulating data and make predictions for further experiments.
Key Personnel:
| Name | Organization | Role on Project |
| Matthew A. Lambon-Ralph | Univ. of Manchester | Principal Investigator |
| Karalyn Patterson | MRC Cognition and Brain Sciences | Co-PI |
| Cathy J. Price | Univ. College London | Co-PI |
| David C. Plaut | Carnegie Mellon/CNBC | Co-PI |
| Mark S. Seidenberg | Univ. of Wisconsin (Madison) | Consultant |
|
