研究介紹: 手指拍打(開合) 測驗為動作障礙學會修訂之巴金森症狀衡量表中的其中一個測驗。量表中的各項測驗分數皆是由醫生根據自身的經驗評斷。因此有許多研究致力於利用感測器和行動裝置輔助量化及數位化診斷的過程。雖然使用行動裝置輔助的研究大多都將手指開合(開合) 納入為研究中其中一個測驗，但大部分的研究皆使用修改過的手指測驗(交替拍打測驗)，而此動作與量表中的手指拍打(開合) 測驗有很大的差異。也因為此動作的差異，量表中的分數和交替拍打測驗的結果之連結性變得較不直觀。
研究目標: 我們的目標為提供一個可以紀錄病患在平板上做標準手指拍打(開合) 測驗的連續資料並給出一個客觀且具專一性的分數並用以評估病患的當前狀況，且此一分數和量表的分數存在直接的對應關係。此測驗可以在任何地方進行，無須醫師或護理師在場，如此即可讓病況的長期追蹤更方便。
研究方法: 實驗中病患須將拇指與食指在平板螢幕上開合，由平板收集病患的動作資料。由平板收集到的資料須先經過前處理來將病患拇指與食指的資料分開。接著我們參考量表中對動作的描述定義出相對應的特徵，並使用兩種機器學習模型，支援向量機(SVM) 和隨機森林，在搭配竭盡式特徵選取法來進行資料分析與二元分類模型建立。模型的表現好壞是由兩種評分方式來評斷，分別為AUROC 和Mattew's correlation coefficient。
研究結果與結論: 在控制組vs.帕金森組情境中，最佳的模型表現為AUROC 0.836 (p = 0.0008) 和MCC 0.512 (p = 0.0008)，且最具差異性的特徵為峰值高度標準差和平均速度。而在低UPDRS 組vs. 高UPDRS 組情境中，最佳的模型表現為AUROC 0.973 (p = 0.0005) 和MCC 0.797 (p = 0.0001)，且最具差異性的特徵為峰值高度減少量、峰值高度標準差和最大速度。實驗的結果說明了我們提出的方法有潛力能判別出受試者是否患有帕金森氏症，且也能判別出他的UPDRS 分數為低分或高分。

Background: The Finger tapping test (FTT) is one of the motor function tests in the Movement Disorders Society-sponsored Revision of the Unified Parkinson's Disease Rating Scale (MDS-UPDRS). The score of each test is given by the physicians and depends on their experiences. Therefore, many studies have focused on automating and digitalizing motor function assessment with the help of sensors and mobile devices. Although many studies include FTT, most of them used the Alternating finger tapping test, which is a very different movement compared to the standard UPDRS FTT that only is recorded as discrete tap events. Aim: Our goal is to record the continuous movement of the standard FTT on a tablet to give an objective and consistent score reflecting the subjects' current status, which is directly comparable to the familiar MDS-UPDRS score for FTT. The test can take place anywhere without the need for a physician, which enables long-term frequent tracking of disease progression.
Methods: The subject performs the FFT on the tablet laying on a table by sliding the thumb and index finger against the screen. The data is preprocessed to distinguish the data points of the thumb and index finger from other touches, like the wrist or another finger. Features, designed to capture the traits mentioned in the FTT part of MDS-UPDRS, are constructed based on the trajectory of the thumb and index fingers. Support Vector Machine and Random Forest along with an exhaustive feature selection are used to build models for binary classification. The performance of the models is measured using AUROC and Matthew's correlation coefficient.
Results and conclusion: In scenario healthy control vs. Parkinson's disease (PD), the best model performance is AUROC 0.836 (p = 0.0008) and MCC 0.512 (p = 0.0008), and the distinctive features are Peak amplitude standard deviation and Average speed. In scenario low vs. high UPDRS score, the best model performance is AUROC 0.973 (p = 0.0005) and MCC 0.797 (p = 0.0001), and the distinctive features are Peak amplitude decrement, Peak amplitude standard deviation, and Maximum speed. In conclusion, the proposed method has the potential to distinguish if a subject has PD and if the subject has a low or high UPDRS score.

Arora, S., Venkataraman, V., Zhan, A., Donohue, S., Biglan, K. M., Dorsey, E. R., and Little, M. A. (2015). Detecting and monitoring the symptoms of Parkinson's disease using smartphones: A pilot study. Parkinsonism & related disorders, 21(6):650–653.

Arroyo-Gallego, T., Ledesma-Carbayo, M. J., Sánchez-Ferro, A., Butterworth, I., Mendoza, C. S., Matarazzo, M., Montero, P., López-Blanco, R., Puertas-Martin, V., Trincado, R., et al. (2017). Detection of motor impairment in Parkinson's disease via mobile touchscreen typing. IEEE Transactions on Biomedical Engineering, 64(9):1994–2002.

Ashyani, A., Lin, C.-L., Roman, E., Yeh, T., Kuo, T., Tsai, W.-F., Lin, Y., Tu, R., Su, A., Wang, C.-C., Tan, C.-H., and Nordling, T. E. M. (2022). Digitization of UPDRS upper limb motor examinations towards automated quantification of symptoms of Parkinson's disease [unpublished manuscript]. Manuscript in preparation.

Cortes, C. and Vapnik, V. (1995). Support-vector networks. Machine learning, 20(3):273–297.

De Lau, L. M., and Breteler, M. M. (2006). Epidemiology of Parkinson's disease. The Lancet Neurology, 5(6):525–535.

Ho, T. K. (1995). Random decision forests. In Proceedings of 3rd international conference on document analysis and recognition, volume 1, pages 278–282. IEEE.

Kassavetis, P., Saifee, T. A., Roussos, G., Drougkas, L., Kojovic, M., Rothwell, J. C., Edwards, M. J., and Bhatia, K. P. (2016). Developing a tool for remote digital assessment of Parkinson's disease. Movement disorders clinical practice, 3(1):59–64.

Lee, U., Kang, S., Choi, J., Kim, Y., and Ma, H.-I. (2016a). Mobile application of finger
tapping task assessment for early diagnosis of Parkinson's disease. Electronics Letters, 52(24):1976–1978.

Lee, W., Evans, A., and Williams, D. R. (2016b). Validation of a smartphone application measuring motor function in Parkinson's disease. Journal of Parkinson's disease, 6(2): 371–382.

Lin, Y., Yu, H., Wan, F., and Xu, T. (2017). Research on classification of chinese text data based on svm. In IOP Conference Series: Materials Science and Engineering, volume 231, page 012067. IOP Publishing.

Lipsmeier, F., Taylor, K. I., Kilchenmann, T., Wolf, D., Scotland, A., Schjodt-Eriksen, J.,
Cheng, W.-Y., Fernandez-Garcia, I., Siebourg-Polster, J., Jin, L., et al. (2018). Evaluation of smartphone-based testing to generate exploratory outcome measures in a phase 1 Parkinson's disease clinical trial. Movement Disorders, 33(8):1287–1297.

Mitsi, G., Mendoza, E. U., Wissel, B. D., Barbopoulou, E., Dwivedi, A. K., Tsoulos, I.,
Stavrakoudis, A., Espay, A. J., and Papapetropoulos, S. (2017). Biometric digital health technology for measuring motor function in Parkinson's disease: results from a feasibility and patient satisfaction study. Frontiers in neurology, 8:273.

Sahyoun, A., Chehab, K., Al-Madani, O., Aloul, F., and Sagahyroon, A. (2016). Parknosis: Diagnosing Parkinson's disease using mobile phones. In 2016 IEEE 18th International Conference on e-Health Networking, Applications and Services (Healthcom), pages 1–6. IEEE.

Shichkina, Y., Stanevich, E., and Irishina, Y. (2020). Assessment of the status of patients with Parkinson's disease using neural networks and mobile phone sensors. Diagnostics, 10(4):214.

Stamate, C., Magoulas, G. D., Küppers, S., Nomikou, E., Daskalopoulos, I., Jha, A., Pons, J., Rothwell, J., Luchini, M. U., Moussouri, T., et al. (2018). The cloudupdrs app: A medical device for the clinical assessment of Parkinson's disease. Pervasive and mobile computing, 43:146–166.

Suominen, H., Manocha, M., Desborough, J., Parkinson, A., and Apthorp, D. (2021). Finger tapping measures for Parkinson's disease: Preliminary evaluation of an android application for data collection in Australia. In Nurses and Midwives in the Digital Age, pages 475–480. IOS Press.

Wissel, B. D., Mitsi, G., Dwivedi, A. K., Papapetropoulos, S., Larkin, S., Castellanos, J.
R. L., Shanks, E., Duker, A. P., Rodriguez-Porcel, F., Vaughan, J. E., et al. (2017). Tablet-based application for objective measurement of motor fluctuations in Parkinson's disease. Digital biomarkers, 1(2):126–135.

Zhan, A., Mohan, S., Tarolli, C., Schneider, R. B., Adams, J. L., Sharma, S., Elson, M. J.,
Spear, K. L., Glidden, A. M., Little, M. A., et al. (2018). Using smartphones and machine learning to quantify Parkinson's disease severity: the mobile Parkinson disease score. JAMA neurology, 75(7):876–880.