Parkinson's disease (PD) is the second most common neurodegenerative disorder, after Alzheimer's disease, affecting approximately 1% of the population aged 55 or older. Major symptoms of PD include tremor, bradykinesia and freezing of gait. The precise diagnosis of PD, at its early stages, remains a challenge for modern clinicians. The difficulty to differentiate PD from other neurodegenerative disorders is high due to the similarity of symptoms with other disorders. Although clinical diagnosis primarily rests on the presence of typical clinical manifestations such as bradykinesia, tremor and other cardinal motor features, PD is associated with a plethora of non-motor symptoms adding to its overall disability. Recent research into early diagnosis of PD has taken advantage of technological advancements in machine learning. In the stream of applying machine learning for data analysis, several studies have been shown to accurately diagnose PD patients using striatal binding ratio (SBR) values. Furthermore, with deep learning techniques, several neuroimaging modalities, such as single photon emission computed tomography (SPECT) and positron emission tomography (PET), have aided early and differential diagnosis of PD. In this chapter, we will shed some light upon relevant machine learning algorithms, their limitations and advantages, and provide new insights on how machine learning techniques can be potentially and clinically utilised in contributing to the ongoing PD diagnostic research and practice.
Applications of machine learning techniques in the diagnosis of Parkinson's disease: promises and challenges, Page 1 of 2
< Previous page Next page > /docserver/preview/fulltext/books/he/pbhe024e/PBHE024E_ch23-1.gif /docserver/preview/fulltext/books/he/pbhe024e/PBHE024E_ch23-2.gif