Parkinson's Disease (PD) is a neurological disorder that affects the central nervous system causing degeneration in a person's speech and motor skills. A diagnosis of Parkinson disease requires the presence of bradykinesia, as well as tremor, rigidity, and/or postural instability. One of the most commonly used scales for Parkinson patients is the Unified Parkinson's Disease Rating Scale (UPDRS). In most PD patients, symptoms are initially unilateral and are caused by a lack of dopamine production predominantly in a specific part of the brain called the substantia nigra. The clinical manifestations do not follow a linear progression with the scales, and subjects present with a variety of combination of symptoms. Thus it is desirable to have a large database of patients and healthy controls to evaluate how different disease manifestation is reflected in resting state functional connectivity. The goal is to be able to characterize the patients and measure their evolution and response to treatment in a quantitative manner.
This summer I worked on a dataset of 31 patients in the 'off' condition, 16 in the 'on' condition and 26 gender and age matched healthy volunteers. After acquiring both an anatomical and time series image, most of the work involved preprocessing the data, so that later analysis could be done across groups. The first step in the preprocessing was to remove unwanted material from the images, such as the skull in the anatomical and signal spikes in the time series data. Next, the skull stripped anatomical data had to be aligned to the filtered time series data. In order to allow for a better registration into a standard space, first a transformation was applied to the anatomical data, which has more of a similar contrast to the standard space, and then the same transformation was applied to the time series. Once aligned, FreeSurfer segmentation of the anatomical was used to create masks of unwanted brain areas, such as the ventricles and white matter. These masks along with respiratory, cardiac, and motion signals were regressed out.
All the datasets are now organized in a program (3dGroupInCorr) that allows for exploring connectivity and contrast within 2 groups and also for covariate regression; thus clinical data can be included in the analysis of the functional connectivity networks.
We are now ready to explore different networks and ask clinically relevant questions. The example presented in the poster shows a seed placed in the left putamen of the healthy volunteers, PD off, and PD on. In comparing the network connectivity between groups a difference in functional connectivity from the left putamen to other regions of the brain can be seen. By first looking at the comparison between healthy volunteers and PD off questions, regions of the brain that are inherent to Parkinson Disease can be identified. Comparing these same regions to PDon and PDoff, further analysis of these networks can then be looked at to determine effectiveness of medication or if there exists a correlation between functional connectivity and, currently the only measurement of disease severity, UPDRS scores.
Last updated March 20, 2013