Among Parkinson’s disease patients, up to 70 percent eventually experience dementia as their disease progresses. Recent studies have reported that the average time from onset of Parkinson’s to developing dementia is approximately 10 years. Thus, an early diagnosis of Parkinson’s Disease Dementia (PDD) can effectively enhance the outcome of their treatment, while slowing the progress of the condition. Furthermore, an early diagnosis can also raise public awareness of PDD, as well as its negative impact on the affected individuals.
Recent studies demonstrate that the onset of serious cognitive decline occurs alongside substantial cholinergic denervation. Imaging of the cholinergic basal forebrain (cBF) volume reductions in PDD offers a relatively low cost, widely available and non-invasive procedure for revealing the tissue changes associated with cognitive symptom progression and then using its in vivo measurement as a predictive cognitive biomarker. The current research focused on longitudinal evaluation of cognitive status in participants while it is generally believed that, in mild cognitive impairment, there are spatial and temporal relationships between pathological changes and the cBF alteration. MRI spatial and temporal imaging information can accurately detect regional cBF degeneration in PDD, suggesting its relationship with early related cognitive impairments.
In this proposed research, a novel Deep Convolutional Neural Network (CNN) architecture, which simulates the natural neuromorphic multi-layer network, will be utilised to automatically and adaptively learn a hierarchical representation of cBF patterns, based on both spatial and temporal features. This would be subsequently applied to the detection of the relevant regions of interest, while obtaining more detailed information on the cBF structure. This would also lead to the identification of any change in its volume, alongside the cognitive degeneration in PDD, which would be crucial in assessing the suggestion that volumetric measurement of the cBF can predict early PDD.
|Effective start/end date||1/03/22 → 28/02/23|
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.