Transdermal optical imaging revealed different spatiotemporal patterns of facial cardiovascular activities

Figures

• 

  Figure 1. Schematic overview of the acquirement of the transdermal face image. (a) The traveling and re-mitting of light through different skin tissues. (b) An example of transdermal face image.

• 

  Figure 2. Schematic overview of group ICA. (a) The video images of 11 participants are concatenated head-to-tail and the concatenated time courses that include the transdermal facial blood flow are extracted from 10 regions of interests, respectively. (b) The group ICA is actually an extended version of Equation 3. Here, each row of the left column shows the concatenated time courses of 11 participants originally extracted from each of the 10 facial ROIs, which is indicated by the corresponding row of Matrix X in Equation 3. Each independent component decomposed from the original concatenated time courses (i.e., the left column) have two features, namely the temporal dynamics and the spatial distribution. Each row of the right column shows the concatenated temporal dynamics of each independent component across 11 participants, which is indicated by corresponding row of Matrix S in Equation 3. Each sub-figure of the middle column shows the spatial distribution of each independent component, which is indicated by corresponding column of matrix W−1 in Equation 3. It should be noted that, for each independent component, all the time points of its concatenated temporal dynamics have the same…

• 

  Figure 3. The results of group ICA of an example participant. (a) The individual temporal dynamics and spatial distribution of each independent component for this participant. It should be noted that the individual temporal dynamics are segmented from the concatenated temporal dynamics of respective independent components. The color map presents the spatial distribution of strength with which the temporal dynamics of one independent component is projected into different ROIs of a face. For an independent component, the more yellow the ROI of its spatial distribution, the stronger strength with which its temporal dynamics is projected into this ROI. (b) The magnitude spectrum of the individual temporal dynamics. (c) The phase spectrum of the individual temporal dynamics. (d) The time course of ECG signal (left, collected at 200 Hz, resampled to 20 Hz) and its magnitude spectrum (middle), and phase spectrum (right). The shade region indicates the frequency band of 0.7~2.5 Hz.

• 

  Figure 4. The linear-arranged spatial distributions of five independent (Normalized within independent component). The color map presents the spatial distribution of strength with which the temporal dynamics of one independent component is projected into different ROIs of a face. For an independent component, the more yellow the ROI of its spatial distribution, the stronger strength with which its temporal dynamics is projected into this ROI.