Real-Time High-Level Acute Pain Detection Using a Smartphone and a Wrist-Worn Electrodermal Activity Sensor

Figures

• 

  Figure 1. Flow chart of TVSymp and MTVSymp calculation. EDA: electrodermal activity, cvxEDA: convex EDA optimization method, VFCDM: variable frequency complex demodulation.

• 

  Figure 2. (a) Main screen (heat pain was induced around 56 s), (b) device selection in the smartphone application, and (c) the unified model language (UML) sequence diagram for the application.

• 

  Figure 3. Objective real-time pain measuring scheme. The red bars in the signals represent where a stimulus was given. Fs: sampling frequency (2 Hz).

• 

  Figure 4. An example of EDA signal and derived time series: (a) EDA indices calculated using the entire recording, (b) EDA indices calculated using our objective real-time pain measuring scheme based on windowing. Red bars indicate where stimuli were given.

• 

  Figure 5. Thermal grill (left) and an example of data collection (right).

• 

  Figure 6. Grid-search cross-validation scheme with leave-one-subject-out cross-validation (each fold). This was repeated as many as the number of subjects to test all subjects.

• 

  Figure 7. Comparison of Visual analog scale (VAS) between EP and TG stimulation.

• 

  Figure 8. Boxplots for maximum values of TVSymp, MTVSymp, and dPhEDA for both datasets. The circles in the boxplots indicate outliers. Outliers were set if each datum is above Q1 − 1.5 × (Q3 − Q1) or below Q3 + 1.5 × (Q3 − Q1). Q1 and Q3 represent the first and third quartiles, respectively.

• 

  Figure 9. Boxplots for maximum values of TVSymp, MTVSymp, and dPhEDA for EP dataset. The circles in the boxplots indicate outliers. Outliers were set if each datum is above Q1 − 1.5 × (Q3 − Q1) or below Q3 + 1.5 × (Q3 − Q1). Q1 and Q3 represent the first and third quartiles, respectively.

• 

  Figure 10. Boxplots for maximum values of TVSymp, MTVSymp, and dPhEDA for TG dataset. The circles in the boxplots indicate outliers. Outliers were set if each datum is above Q1 − 1.5 × (Q3 − Q1) or below Q3 + 1.5 × (Q3 − Q1). Q1 and Q3 represent the first and third quartiles, respectively.