Shukla Pitamber, Basu Ishita, Tuninetti Daniela
Annu Int Conf IEEE Eng Med Biol Soc. 2014;2014:2605-8. doi: 10.1109/EMBC.2014.6944156.
Deep Brain Stimulation (DBS) is a surgical procedure to treat some progressive neurological movement disorders, such as Essential Tremor (ET), in an advanced stage. Current FDA-approved DBS systems operate open-loop, i.e., their parameters are unchanged over time. This work develops a Decision Tree (DT) based algorithm that, by using non-invasively measured surface EMG and accelerometer signals as inputs during DBS-OFF periods, classifies the ET patient's state and then predicts when tremor is about to reappear, at which point DBS is turned ON again for a fixed amount of time. The proposed algorithm achieves an overall accuracy of 93.3% and sensitivity of 97.4%, along with 2.9% false alarm rate. Also, the ratio between predicted tremor delay and the actual detected tremor delay is about 0.93, indicating that tremor prediction is very close to the instant where tremor actually reappeared.
深部脑刺激（DBS）是一种外科手术，用于治疗某些晚期进行性神经运动障碍，如特发性震颤（ET）。目前美国食品药品监督管理局（FDA）批准的DBS系统采用开环运行，即其参数随时间不变。这项工作开发了一种基于决策树（DT）的算法，该算法在DBS关闭期间，通过使用非侵入性测量的表面肌电图和加速度计信号作为输入，对ET患者的状态进行分类，然后预测震颤何时即将再次出现，此时再次开启DBS一段固定时间。所提出的算法总体准确率达到93.3%，灵敏度为97.4%，误报率为2.9%。此外，预测震颤延迟与实际检测到的震颤延迟之间的比率约为0.93，这表明震颤预测非常接近震颤实际再次出现的时刻。
