Between-subject correlation of heart rate variability predicts movie preferences

Fig 1. HRV data collection process in study 1. Schematic diagram of the experimental procedure: Participants watched four movies (A, B, C, and D) in sequence while HRV data was collected and facial expressions were recorded.

Fig 2. Normalized HRV graphs of four of the participants in the five-hour-long study 1. The strongest synchrony was marked by the longest plateau (enclosed by the green rectangle) that was produced by the “calm down” period. The graphs also show that synchronized upward spikes (red arrows) and downward spikes (blue arrows) exist.

Fig 3. The scale group moving average (over every 100 data points) of heart rate variability over time. Raw pulse-pulse intervals (PPI) values were linearly interpolated and fit with univariate splines of degree three before absolute derivatives were taken. The vertical axis was scaled between zero and one using Min-Max scaling. The vertical dashed lines mark the four sections corresponding to the four movies watched.

Fig 4. Audience preferences for comparisons 2A and 2B. The bar chart shows the respective number of participants who preferred each video cut. The four movies are represented by different shades. A chi-square goodness of fit test was performed to examine whether participants prefer the “most aroused” cut over (2A) the random cut or (2B) the “least aroused” cut. Both relations were significant. Participants were more likely to prefer the “most aroused” cut over the random cut, χ2(1, N = 24) = 10.7, p = .001. Participants were also more likely to prefer the “most aroused” cut over the “least aroused” cut, χ2(1, N = 16) = 4.00, p = .046.

Fig 5. Audience preferences for comparisons 3A, C, and D (a) by study and (b) by movie. The pie chart shows the proportions of participants’ preferences. The sector of HRV based cuts is further broken down to individual bars showing (a) the respective number of participants who preferred each HRV based cut, including (3A) the “most aroused” cut, (3C) the “least aroused” cut, and (3D) the “most synchronous” cut; or (b) the respective number of participants based on the movie watched.

Fig 6. Audience preferences for comparison 3E by movie. The bar chart shows the respective number of participants who preferred the “most aroused” cut based on female HRV and the “most aroused” cut based on male HRV, based on the movie watched.