Correct RR Intervals Using Cubic Spline Interpolation

Source: R/preprocessing.R

Applies cubic spline interpolation to detect and correct artifacts in RR intervals. This method follows the Kubios methodology and balances automation with accuracy.

correct_rr_cubic_spline(
  rr_intervals,
  threshold_percent = 25,
  hr_adaptive = TRUE,
  max_correction_rate = 5
)

Arguments

rr_intervals

Numeric vector of RR intervals in milliseconds

threshold_percent

Percentage threshold for artifact detection. Default is 25% following Kubios methodology

hr_adaptive

Logical indicating whether to use HR-adaptive thresholds. Default is TRUE

max_correction_rate

Maximum percentage of beats that can be corrected. Default is 5% to avoid over-smoothing

Value

A list containing:

  • corrected_rr: Numeric vector of corrected RR intervals

  • artifact_mask: Logical vector indicating which beats were corrected

  • n_corrected: Number of intervals that were corrected

Details

This function implements the Kubios-style cubic spline correction:

  • Uses HR-adaptive thresholds (20-30% or 0.20-0.30s ranges)

  • Limits corrections to <5% of beats to avoid over-smoothing

  • Applies cubic spline interpolation using R's spline() function

  • Automatically detects artifacts based on percentage deviation

The method balances automated detection with preservation of physiological variability, making it widely adopted in HRV research.

References

Based on Kubios HRV methodology for cubic spline artifact correction.

Examples

if (FALSE) { # \dontrun{
rr_data <- c(800, 850, 400, 830, 810, 1600, 820)
result <- correct_rr_cubic_spline(rr_data, threshold_percent = 25)
corrected_data <- result$corrected_rr
} # }