Retrieval of verbs from the lexicon in people with aphasia: an information-theoretic approach.

Word finding difficulties are a common and very frustrating symptom of aphasia.
Research over the last thirty years has investigated which categories of words
are more difficult or easier for people with different types of aphasia. For
people with Broca*s aphasia verbs (in addition to words from functional
categories, such as determiners and pronouns) are particularly difficult
(Berndt et al, 1997; Kim and Thompson, 2000). Recent research shows that this
difficulty with verbs may not be restricted to patients with Broca*s aphasia:
patients with Wernicke*s patients have been shown to have similar difficulties.
When verbs are produced by patients with aphasia, they often lack inflection
(e.g. he walk, instead of he walks). Interestingly, not all inflected verb
forms are equally difficult to produce. What is still unclear is why some verb
forms are more difficult than others. Some researchers claim that in aphasia
grammatical knowledge about how to form inflected verb forms is lost due to
damage to the brain (Friedmann and Grodzinsky, 1997), whilst others suggest
grammatical knowledge is intact but is difficult to access and activate due to
reduced processing capacity in these patients. The lost knowledge hypothesis
predicts that if a patient has for example difficulties forming regular past
tense forms, they will aways have this difficulty. Variability within patients
should not be found. Furthermore, in this framework this difficulty should
occur regardless of the type of verb that the patient wants to use. All verbs
should be equally difficult. The processing difficulty hypothesis on the other
hand, allows for variability within and between patients. It suggests that
efficient information transmission in the brain is disrupted by tissue damage.
Depending on context, both linguistic and non-linguistic, a patient might
sometimes succeed in forming past tense and sometimes fail. Thus far there has
been very little research on which (linguistic) factors may enhance or decrease
performance. Finding out which factors impede on effective information
transmission in the brain, both in language and other cognitive domains has
crucial implications for diagnosis of aphasia and aphasia therapy.
As discussed below, this research project aims to explore information
processing in the brain. Linguistic factors that increase complexity of verb
forms will be explored in order to explain variability of verb retrieval within
patients. In addition, a visual task will be used to explore the possibility
that the tissue damage in the brain has lead to more general reductions in
effective information transmission. Exact quantitative data on how information
transmission in the brain (in language as well as other domains) is affected by
tissue damage will contribute to a better understanding of information
processing in the healthy brain, and is expected to assist in identifying
effective compensatory mechanisms / interventions in brain damaged patients.
The linguistic factors that will be investigated in this project have been
shown to play an important role in lexical retrieval in healthy adults. This
research has shown that the complexities of words and word families are
important factors in the processing time required to retrieve lexical items
(including inflected forms) from the mental lexicon (Moscoso del Prado Martin,
Kostic and Baayen, 2004). This line of research uses statistical measures from
information theory to calculate complexity of specific word forms. For each
inflected verb form a complexity measure can be created. The higher the value
of the complexity measure, the longer it takes for normal adults to process the
verb form. In addition measures for the competition of the family of words can
be calculated and its influence measured on processing time. In healthy adults
these differences in processing time are very small and go unnoticed in
conversation. The capacity of the normal brain is by far sufficient to process
a verb of any complexity measure without errors. However, we know that in the
damaged brain verb retrieval is problematic. The processing capacity of
(linguistic) material is reduced. It is therefore very interesting to find out
if those verb forms with higher complexity values are the verb forms that are
more difficult (and sometimes impossible) for these patients to produce. In
addition, the influence of the verb*s surrounding neighbours, or its family, in
the mental lexicon may have changed as transmission of information is reduced.

This research project aims to examine possible reduced processing capacity in
people with aphasia. Unlike previous research, this project will be the first
to provide quantitative measures of *complexity* of linguistic material. This
provides a unique opportunity to investigate information processing in the
damaged brain. Clear predictions can be made and if confirmed by our
experiments, strong evidence can be provided for a reduced processing capacity
account of aphasia. In addition to investigating linguistic factors, this
project will also probe information processing in the visual domain. Reduced
information processing for linguistic material in the damaged brain is likely
to be found in the aphasic speakers, as they are know to have language
difficulties. However, it is unclear whether subtle difficulties in information
processing may also be evident in other cognitive modalities.

The goal of this research project is to find out to what extent people with
aphasia suffer from reduced processing capacity. We will use new insights from
healthy language processing providing quantitative measures of "complexity" to
investigate to what extend complexity of verbs and their families influence the
reaction and error rates in verb retrieval in aphasia. If processingcapacity is
reduced in people with aphasia, we would expect that verb forms with a high
complexity (high information load) will be retrieved more slowly (or not at
all) from the lexicon. Furthermore, we aim to explore whether this reduced
processing is limited to language material. We will include a taks of visual
cognition to search for dissociations or parallels in language and visual
cognition.

Participants will take part in 3 experiments. Experiment 1 involves an auditory
lexical decision task in which the participant will be asked to indicated
whether an auditory stimulus is a real word or not. In experiment 2
participants will be asked to name a series of action pictures. For experiment
3 participants will be asked to indicate whether a predetermined goal is
present or absent in a visual display.
For each task reaction times are error rates will be noted.

For a detailed description of the research design, please refer to page 13 in
the protocol.

Participants will be asked to perform 3 behavioural tasks. This will take 2
hours spread over 2 sessions. THere will be ample opportunity for breaks. The
set up is mobile and participants can therefore be tested at home if this has
their preference. THe burden for the participants is therefore very limited.
As two of the task involve language these tasks may lead to some frustration
for the participants with aphasia. There will be plenty of opportunity to
discuss this and of course participants are free to leave the research at any
moment.