Skip to contents

Truncate Ascending Segment of Melatonin Profile

Source: R/truncate_ascending_segment.R
truncate_ascending_segment.Rd

This function truncates the ascending segment of a melatonin profile based on a set of rules to ensure valid slopes and segment consistency. It iteratively modifies the ascending column of the input data to satisfy the rules and uses the parallelogram_fit function as a diagnostic measure to validate the final segment. The parallelogram_fit function evaluates the geometric structure of the ascending segment and ensures the segment satisfies predefined geometric constraints, including slope ratios of lateral and diagonal shape segments. This ensures that only the steadily increasing part of the melatonin rise is taken into account when determining the DLMO point and that slower rises, or drops in melatonin levels are not.

Usage

truncate_ascending_segment(profile_data)

Arguments

profile_data

A tibble containing the melatonin profile with the following columns:

  • datetime: A POSIXct column with timestamps.

  • melatonin: Numeric column representing melatonin concentrations.

  • slope: Numeric column with the slope between consecutive melatonin points.

  • ascending: Binary column (1 for ascending segment, 0 otherwise).

Value

A list containing:

  • profile: The updated profile_data tibble with truncated ascending segments.

  • plll: The result of the parallelogram fit diagnostic (from parallelogram_fit).

This function ensures the following rules are satisfied:

  1. The rightmost slope in the ascending segment cannot be less than half the steepest slope in the segment.

  2. The ascending segment must adhere to additional conditions validated by the parallelogram_fit function.

Examples

library(dplyr)
library(lubridate)

# Example data
profile_data <- tibble(
  datetime = seq(ymd_hms("2023-01-01 20:00:00"), by = "15 min", length.out = 12),
  melatonin = c(1.2, 1.4, 1.5, 1.7, 2.0, 2.3, 2.8, 3.5, 4.2, 4.7, 5.0, 5.2),
  slope = c(NA, diff(c(1.2, 1.4, 1.5, 1.7, 2.0, 2.3, 2.8, 3.5, 4.2, 4.7, 5.0, 5.2))),
  ascending = c(0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1)
)

# Truncate the ascending segment
result <- truncate_ascending_segment(profile_data)
print(result$profile)
#> # A tibble: 12 × 4
#>    datetime            melatonin  slope ascending
#>    <dttm>                  <dbl>  <dbl>     <dbl>
#>  1 2023-01-01 20:00:00       1.2 NA             0
#>  2 2023-01-01 20:15:00       1.4  0.2           0
#>  3 2023-01-01 20:30:00       1.5  0.100         0
#>  4 2023-01-01 20:45:00       1.7  0.2           0
#>  5 2023-01-01 21:00:00       2    0.3           1
#>  6 2023-01-01 21:15:00       2.3  0.300         1
#>  7 2023-01-01 21:30:00       2.8  0.5           1
#>  8 2023-01-01 21:45:00       3.5  0.7           1
#>  9 2023-01-01 22:00:00       4.2  0.7           1
#> 10 2023-01-01 22:15:00       4.7  0.5           1
#> 11 2023-01-01 22:30:00       5    0.300         1
#> 12 2023-01-01 22:45:00       5.2  0.200         1
print(result$plll)
#> $pll_datetime_0
#> [1] "2023-01-01 20:42:00 UTC"
#> 
#> $pll_datetime_1
#> [1] "2023-01-01 20:56:41 UTC"
#> 
#> $pll_slope
#> [1] 1.93477
#> 
#> $corners
#> $corners$ll
#> [1] 20.7001  1.7000
#> 
#> $corners$lr
#> [1] 20.94485  1.70000
#> 
#> $corners$ur
#> [1] 22.75385  5.20000
#> 
#> $corners$ul
#> [1] 22.5091  5.2000
#> 
#> 
#> $datetime_ref
#> [1] "2023-01-01 20:30:00 UTC"
#> 
#> $flag
#> [1] FALSE
#>