Truman State University's Applied Psychophysiology Lab compared the Institute of HeartMath's new 500-sample/second Inner Balance Coherence Plus PPG sensor with Thought Technology's "gold standard" EKG Flex/Pro (T9306M) sensor.
The Inner Balance Coherence Plus connects to a computer USB-A port or a smartphone using Bluetooth.
We obtained usable data from 26 participants. We simultaneously monitored their ECG using a lower torso placement and PPG on the right ear while sitting without feedback or breathing instructions for 10 minutes. Photograph © BioSource Software.
Caption: A lower torso placement recommended
by Dr. Erik Peper reduces movement artifacts.
We acquired data using the new HeartMath app and BioGraph Infiniti program, which we ran on separate computers. We uploaded both sets of interbeat intervals (IBIs) to Kubios and artifacted the data using their strong setting.
Conclusions
The measurements showed acceptable concurrent validity for 4 of 5 variables with Spearman’s values from 0.99 to 0.97 and mean differences from 0-8%. The PPG sensor readings accounted for 94-98% of the variability in the ECG values for all 5 variables. Moreover, whereas maximum percentage differences have not been codified for these metrics, differences within ± 5% are generally satisfactory for variables like HR.
Limitations
Our Truman undergraduates ranged from 18-21 years and were generally healthy. We cannot generalize these findings to a more diverse population.
Bottom Line
The $249 Inner Balance Coherence Plus PPG sensor achieved acceptable accuracy for clinic and home training.
Granular Analysis
We performed the Shapiro-Wilk test of normality for all pairs. ARLF, FFTLF, and the RMSSD showed normality deviations, making Spearman’s rho and the Wilcoxon Signed-Rank the appropriate statistics. AR stands for an autoregression frequency analysis. FFT stands for Fast-Fourier Transformation frequency analysis. ARLF power Spearman’s rho = 0.98, p < 0.001
Pearson’s =  0.97, p < 0.001 FFTLF power Spearman’s rho = 0.98, p < 0.001
Pearson’s = 0.98, p < 0.001 HR Spearman’s rho = 0.99, p < 0.001
Pearson’s = 1.00, p < 0.001 RMSSD Spearman’s rho = 0.99, p < 0.001
Pearson’s = 0.75, p < 0.001 SDNN Spearman’s rho = 0.99, p < 0.001 Pearson’s = 0.99, p < 0.001 We performed a Wilcoxon test since the paired samples t-test assumptions were not met. The two sensor/placement combinations slightly differed on ARLF power, HR, and the RMSSD. ARLF power
ECG ARLF power, MÂ = 1652.21, SDÂ = 1255.38
PPG ARLF power, M = 1596.35, SD = 1240.82 z = 2.349, p = .018, SE Rank-Biserial Correlation (effect size) = 0.221 difference = 3% FFTLF power ECG FFTLF power, M = 1666.05, SD = 1237.71
PPG FFTLF power, M = 1607.11, SDÂ = 1242.97
z = 1.33, p = .19 (NS)
SE Rank-Biserial Correlation (effect size) = 0.221
difference = 4%
HR ECG HR, MÂ =77.32, SDÂ = 9.19
PPG HR, MÂ = 78.14, SD = 9.31
z = 4.46, p = < .001 SE Rank-Biserial Correlation (effect size) = 0.221 difference = 1% RMSSD ECG RMSSD, M = 39.55, SD = 17.40 PPG RMSSD, M = 42.74, SD = 20.01 z = -2.02, p = .045 SE Rank-Biserial Correlation (effect size) = 0.221 difference = 8% SDNN ECG SDNN, M = 49.46, SD = 14.56
PPG SDNN, M = 49.60, SD = 13.91
z = -0.17, p = .88 SE Rank-Biserial Correlation (effect size) = 0.221
difference = 0% Glossary
low-frequency (LF) band: a HRV frequency range of 0.04-0.15 Hz that may represent the influence of PNS and baroreflex activity (when breathing at the RF). paired-samples t-test: a statistical technique used to compare the means of two related groups to determine if there is a statistically significant difference between these means. Pearson's r: a measure of the linear correlation between two variables, quantifying the strength and direction of their relationship. RMSSD: the square root of the mean squared difference of adjacent NN intervals. SDNN:Â the standard deviation of the normal (NN) sinus-initiated IBI measured in milliseconds. SE Rank-Biserial Correlation: a statistical measure used to assess the strength and direction of association that exists between one dichotomous and one ordinal variable, often used in the context of non-parametric tests. Shapiro-Wilk test of normality: a statistical test used to assess whether a sample comes from a normally distributed population. Spearman's rho: a non-parametric measure of rank correlation, assessing how well the relationship between two variables can be described using a monotonic function. It evaluates the strength and direction of association between two ranked variables.
Wilcoxon Signed-Rank Test: a nonparametric test that compares two related samples to assess whether their population mean ranks differ.
Learn More
Comments