(MRC DTP) Developing a neuroimaging method for early detection and differential diagnosis of Alzheimer’s disease

As promising new treatments for Alzheimer’s disease (AD) emerge, it is becoming imperative to develop effective methods for early detection and differential diagnosis of AD. AD is only one of many potential sources subjective memory complaints in older adults, which also include late-life depression (LLD), sleep disturbance, as well as other types of dementia (e.g., frontotemporal dementia, FTD). By the time neuropsychological tests indicate a probable-AD diagnosis, the neuropathology may be extensive and irreversible. The proposed project will use neuroimaging methods–electroencephalography (EEG) and functional magnetic resonance imaging (fMRI)–to identify patterns of neural responses that differ between patients with AD, patients with LLD, and healthy older controls, and which could therefore be used to inform differential diagnosis in older adults with subjective memory complaints.

EEG measures of neural responses to words presented in various contexts are abnormal in the early stages of AD. In healthy controls, N400 responses are modulated by semantic relations between words, and P600 responses are modulated by word repetition; however, both of these effects are diminished in AD (e.g., Olichney et al., 2006; Taylor & Olichney, 2007). Further, these abnormal responses are predictive of conversion from mild cognitive impairment (MCI) to AD, even when neuropsychological tests cannot distinguish the two (Olichney et al., 2008). In this project, we will aim to build on these EEG findings by including words with emotional content (which will likely evoke different responses in patients with LLD), and other experimental conditions that may help differentiate between AD and other types of dementia. We will use pattern analysis techniques from machine learning in order to identify the data features that are able to correctly classify data features according to diagnostic categories. We will further investigate the neural sources of these EEG features using fMRI.

The project will have obvious clinical value by producing a protocol that will be both sensitive and specific to AD pathology, which could potentially be used to inform diagnosis, to stage the progression of AD, and to evaluate the effects of pharmacological or other treatments. The combination of EEG and fMRI data may also contribute to a better understanding of how different neural systems are affected in AD and related memory disorders.