Time-frequency visualization of alcohol withdrawal tremors.

In this paper, we propose a signal processing method of assessing the severity tremors caused by alcohol withdrawal (AW) syndrome. We have developed an iOS application to calculate the Clinical Institute Withdrawal Assessment (CIWA) score which captures iPod movements using the built-in accelerometer in order to reliably estimate the tremor severity component of the score. We report on the characteristics of AW tremor, the accuracy of electronic assessment of tremor compared to expert clinician assessment, and the potential for using signal processing assessment to differentiate factitious from real tremor in patients seen in the emergency department, as well as in nurses mimicking a tremor. Our preliminary results are based on 84 recordings from 61 subjects (49 patients, 12 nurses). In general we found a linear relationship between energy measured by the accelerometer (in the 4.4-10 Hz range) and the expert rating of tremor severity. Additionally, we demonstrate that 75% of the recordings from patients with actual AW syndrome had a mean peak frequency higher than 7 Hz whereas only 17% of the nurses' factitious tremors were above 7 Hz, suggesting that tremor above 7 Hz could be a potential discriminator of real versus factitious tremors.