Improved Approach for Multiple Signal Alignment with Dynamic Time Warping for Application to Diagnosis of Parkinson’s Disease
While there is no cure to Parkinson’s disease, the onset of symptoms can be delayed with more accurate, early diagnosis. To achieve this, one approach uses Capillary Zone Electrophoresis to generate electropherograms to analyze blood samples of patients. However, when aligning electropherogram samples, the order signals are aligned using dynamic time warping (DTW) can cause misalignment. This study implemented and tested two approaches to a DTW algorithm. The first aligned the signal with the lowest overall distance to each of the other signals. The second aligned each signal with the next signal with the relative lowest distance from itself. Overall, the second approach aligned more signals that were closer together, demonstrating a trend toward better signal alignment. This approach could be implemented in wider studies examining the amount of molecules present in the blood stream of Parkinson’s patients.