Mapping the principal components of aphasic language recovery onto brain structure and function

Stroke is unfortunately a major global health problem, with a person somewhere in the world having their first stroke every
two seconds. The greatest risk factor for stroke is age, hence it represents a significant healthcare challenge in countries
like the UK with an aging population. Cognitive deficits (problems with memory, attention, and language) are very common
after stroke. Problems with language, known as aphasia, are seen in one in every three stroke cases soon after the stroke,
and persist for more than a year after stroke in one in five cases. At present, we do not understand why some people make
a good recovery from aphasia after stroke while others are left with chronic language problems. Nor do we understand why
some people respond well to speech and language therapy while others fail to benefit. The goal of this project is to
combine advanced neuroimaging with sophisticated neuropsychology to answer these questions.
Ideally, we would like to be able to determine from initial brain scans whether someone is likely to recover well from their
current stroke induced impairments. We need to determine prognosis for stroke aphasic patients accurately so we can then
direct therapeutic resources effectively to those who need them most. We also need to understand what aspects of brain
structure and function predict a person's response to standard interventions, as then we can focus on alternative strategies
where needed. To achieve these goals, we need to develop neural biomarkers that can reliably be used to determine
prognosis over recovery and rehabilitation. Previous attempts to achieve prognosis over the course of recovery have had
limited success as they have not used the most advanced measures of neural integrity not have they focussed on the most
relevant aspects of neural function.
This project will exploit the Anatomical Connectivity Mapping (ACM) method developed and patented by Bioxydyn, a
provider of high value quantitative imaging services and diagnostic tools to the pharmaceutical industry, healthcare and
academia. ACM provides a unique measure of long range connectivity across the whole brain, and therefore allows us to
search for changes in brain structure associated with improved performance over recovery and rehabilitation. Our current
work on a cohort of over 65 chronic stroke aphasic patients has shown increases in right hemisphere connectivity that are
not apparent on any other structural measure of brain integrity. Moreover, our recently published work (Butler, Lambon Ralph, & Woollams, 2014, Brain) has adopted a novel approach to lesion symptom mapping by using Principle
Components Analysis (PCA) to identify key cognitive dimensions and their neural correlates. Current work has shown that
tissue concentration in these areas significantly improves prediction of performance in our chronic stroke aphasic sample.
This project will involve firstly using the ACM information from the neural regions associated with specific language
functions identified using PCA (phonology, semantics, fluency; Halai, Woollams, & Lambon Ralph, in press, Cortex) in our
chronic stroke cohort of more than 65 patients to demonstrate that consideration of long range connectivity improves
prediction of performance. We will also obtain functional imaging data from a subset of 20 of these patients to demonstrate
that activation in the PCA defined regions offers additional predictive power. We will then test out the prognostic power of
these predictive models directly by recruiting 20 participants for testing and scanning in the subacute stage after stroke. We
will use their ACM and functional activation in the PCA derived regions to predict their recovery outcomes at retest after
one year. To the extent this approach proves effective, placement with Bioxydyn will facilitate its translation into clinical
practice for prognosis and to provide baseline recovery trajectories for use in clinical trials.