Funding for Cognitive Imaging

Lead Research Organisation: Imperial College London
Department Name: Dept of Medicine


Magnetic resonance imaging (MRI) is a powerful research tool, allowing both the anatomy and function of the human brain to be studied without risk to the subject. We are using this technique to study three common diseases: left cerebral hemisphere stroke; mild-to-moderate Alzheimer's disease; and a form of chronic epilepsy, know as temporal lobe epilepsy.

We are investigating patients who have impaired communication (aphasia) after a left cerebral hemisphere stroke. About 25,000 patients each year are left with persistent aphasia after stroke. Treatment for this condition is far from perfect, and in many instances largely of unproven benefit. It mostly takes the form of behavioural retraining, under the supervision of speech and language therapists. A particular difficulty is the problem of predicting the degree of natural recovery that will occur. This is usually maximal by three months, but the amount of recovery can be very variable. The size of the stroke does influence outcome. However, other factors seem to be involved, and what these are are not knowm. Many people believe that recovery is because language 'switches' from the left to the right cerebral hemisphere, but the evidence that this happens is slender. There is much better evidence that recovery is due to wide reorganisation of undamaged parts of the left hemisphere, and our prediction is that if critical pathways remain intact then recovery is much better. Therefore, we are recruiting 50 patients after aphasic stroke, and in the first phase we are assessing their language abilities within the first week of stroke. We are particularly interested in speech production, as patients find an impaired ability to express themselves the most frustrating aspect of their condition. We then reassess them at three months, and so we have a measure of their 'trajectory' of recovery: good, bad or indifferent. We also do detailed MRI, early and late, to assess the boundaries of the stroke, the function of the whole brain as the patients try to speak, and the integrity of pathways (white matter tracts) throughout the brain. We predict that this will give considerable insight into the factors necessary for good natural recovery, and this knowledge will be essential in informing improved future therapeutic strategies.

Alzheimer's disease
The cardinal feature of Alzheimer's disease (AD) is forgetfulness, particularly for verbal information. There are three stages for remembering what one has just been told: registering the information (that is paying attention to the speaker to understand what is being said); encoding the information as a longer lasting memory; and then retrieving that memory when required. These different stages rely on very different networks in the the brain. Most research has concentrated on the encoding and retrieval stages. We are researching the first stage, and specifically the role of attention. In the early stages of AD, language functions are intact, so undertanding speech is not the problem. Rather, we argue, the patients are inattentive. Attention depends on a right cerebral hemisphere system. We are investigating this system during communication, using functional and anatomical MRI sequences. We are also observing whether the drugs used to alleviate some of the symptoms of AD have a particularly strong action on this right hemisphere system.


Some patients with epilepsy have surgery to prevent their attacks. One common operation is to remove the front pary of one lobe of the brain, the temporal lobe. Surprisingly, this has a relatively minor effect on the patient's higher mental functions. We are invetigating the hypothesis that frequent seizure activity throughot the lifetime of the patient has resulted in reorganisation of the temporal lobe networks, thus protecting the patient from a major deficity after the surgery has been completed. This will inform us more about the protective 'plasticity' of the brain.

Technical Summary

The future programme is investigating three disease groups with MRI techniques; left hemisphere stroke patients with aphasia; patients with mild-to-moderate Alzheimer's disease; and patients with chronic temporal lobe epilepsy. All patient groups will undergo anatomical volumetric scanning, diffusion tensor imaging to investigate white matter tract integrity, and functional MRI (fMRI). The scanning data will be analysed using FSL software (FMRIB, Oxford, UK). The functional data will be analysed using both univariate statistics, to observe regional effect size, and multivariate statistics (principally independent component analyses), to assess functional connectivity across the brain.

The stroke project is designed to observe the natural change in language impairment over the three months after stroke, at which time recovery is usually maximal. Early and late scanning sessions will assess lesion size and distribution, the integrity of intrahemispheric language tracts within the left hemisphere, and changes in the distributed connectivity across the brain that relate to better natural recovery. The hypothesis is that these measures will give a reliable set of imaging biomarkers that predict outcome after aphasic stroke.

The project on Alzheimer's disease is specifically investigating the integrity of the right fronto-parietal network for sustained attention. Our hypothesis is that the forgetfulness of AD patients is partly due to a failure to register information, in addition to problems with memory encoding and retrieval. We have piloted a behavioural verbal task that robustly activated the right fronto-parietal system, and this will be applied to AD patients before and after treatment with a disease-modifying agent.

The TLE project is using DTI and fMRI to investigate the reorganisation of semantic memory temporal lobe networks in response to repeated clinical and sub-clinical seizure activity. This addresses issues about anterior temporal lobectomy.

Planned Impact

This is addressed in the section on 'Pathways to Impact'.


10 25 50
publication icon
Braga RM (2013) Echoes of the brain within default mode, association, and heteromodal cortices. in The Journal of neuroscience : the official journal of the Society for Neuroscience

publication icon
Hellyer PJ (2014) The control of global brain dynamics: opposing actions of frontoparietal control and default mode networks on attention. in The Journal of neuroscience : the official journal of the Society for Neuroscience

publication icon
Geranmayeh F (2014) Overlapping networks engaged during spoken language production and its cognitive control. in The Journal of neuroscience : the official journal of the Society for Neuroscience

publication icon
Simmonds AJ (2014) Sensory-motor integration during speech production localizes to both left and right plana temporale. in The Journal of neuroscience : the official journal of the Society for Neuroscience

publication icon
Sliwinska MW (2017) Stimulating Multiple-Demand Cortex Enhances Vocabulary Learning. in The Journal of neuroscience : the official journal of the Society for Neuroscience

publication icon
Simmonds AJ (2014) The response of the anterior striatum during adult human vocal learning. in Journal of neurophysiology