Assessment of Cutaneous Microcirculation Autoregulation with Phase and Linear Based Eulerian Video Magnification

Abstract

IntroductionThe identification of low frequency variations in blood pressure recordings date back to the mid‐1800's. The mechanism for the low frequency oscillation is associated with intrinsic myogenic activity of smooth muscle cells in resistance arteries. Therefore, local metabolic demand modulates blood flow distribution. Non‐invasive assessment of the microcirculation in dermatology, diabetes, skin‐graft, and burn‐injury patients could facilitate assessment of therapeutic interventions. Eulerian Video Magnification (EVM) software was developed by the Massachusetts Institute of Technology (MIT) to magnify subtle changes in motion and color in video recordings. Using this technology, magnification of a video recording may reveal local changes in the cutaneous human microcirculation.HypothesisWe hypothesize that EVM can be used to assess oscillations in cutaneous blood flow velocity.MethodsIRB approval was obtained to acquire physiological data from healthy graduate students (N = 10). Tissue Doppler was used to measure local changes in blood flow velocity on the forehead. Heart rate was assessed with an ECG and subjects breathed in time to a metronome (15 breaths/min) to normalize the effect of ventilation. Videos were processed using MIT's EVM Matlab code to magnify motion (Phase), or color & motion (Linear). All videos were magnified in the frequency range of interest (0.067–1.67 Hz; 4–100 cycles/min) to enhance human physiological frequencies. Gain was set to 50%; and a Gaussian filter was set to level 4. Phase magnification used a pyramid filtering scheme with octaves (1, ½, ¼) to specify the number of times the video is filtered; while linear magnification utilized chrominance attenuation to reduce color changes created by the magnification process. Red, green, and blue color spectrums extracted from magnified videos were analyzed with a Fast Fourier Transform (FFT) to identify magnified frequencies and power.ResultsPhase magnification (motion) and corresponding FFTs suggest strong correlations to heart rate and respiration rate but not tissue Doppler blood flow. On the other hand, linear magnification (color) and accompanying FFTs were strongly correlated to heart rate (r = 0.987; p = 0.001), respiration rate (r = 0.741; p = 0.014), and local tissue Doppler recordings (r = 0.945; p = 0.001).ConclusionThese preliminary data demonstrate that a video recording of a human subject can be magnified with EVM to assess heart rate, respiration rate, and regional blood flow velocity oscillations in the cutaneous microcirculation. EVM may be a useful non‐invasive modality to evaluate changes in microcirculation autoregulation in human subjects.Support or Funding InformationMidwestern University, Chicago College of Osteopathic Medicine, College of Health Sciences