Fit Recovery Trajectory to Longitudinal Activation Data

Fits parametric recovery curves (exponential, sigmoid, or linear) to longitudinal muscle activation data from an MSKLongitudinalTracker.

Usage

mskRecoveryTrajectoryFit(
  tracker,
  model = c("exponential", "sigmoid", "linear"),
  muscle_subset = NULL,
  partial_pool = FALSE
)

Arguments

tracker

An MSKLongitudinalTracker object.

model

Character, recovery curve model: "exponential" (default), "sigmoid", or "linear".

muscle_subset

Optional character vector of muscle names to fit (NULL = all muscles).

partial_pool

Logical; when TRUE and there is more than one muscle, an exponential fit is estimated by partial pooling across muscles via a population NLME (PhysioClinStats::recoveryTrajectoryLME, if installed), rather than one independent NLS per muscle. Defaults to FALSE (the independent-NLS behaviour), which is also the fallback if the NLME is unavailable or fails to converge.

Value

An S3 object of class "MSKRecoveryTrajectory" with:

fits

list per muscle with coefficients, residuals, R-squared, AIC

predicted

matrix of predicted values (muscles x timepoints)

model_type

character

recovery_rate

numeric vector (slope at midpoint for each muscle)

time_to_90pct

estimated time to 90 percent of asymptotic recovery

Examples

C <- matrix(c(1,1,0,0,0, 0,1,1,0,0, 0,0,1,1,0, 0,0,0,1,1, 1,0,0,0,1, 0,1,0,1,0),
            nrow = 6, byrow = TRUE,
            dimnames = list(paste0("bone", 1:6),
              c("Biceps Brachii", "Deltoid", "Trapezius",
                "Triceps Brachii", "Brachialis")))
hg <- MSKHypergraph(C)
mk <- function(s) { set.seed(s); m <- matrix(abs(rnorm(300 * 4)), 300, 4)
  colnames(m) <- c("Biceps Brachii", "Deltoid", "Trapezius", "Triceps Brachii"); m }
tracker <- mskLongitudinalTracker(list(T0 = mk(1), T1 = mk(2), T2 = mk(3)), hg = hg)
traj <- mskRecoveryTrajectoryFit(tracker, model = "linear")
traj
#> MSK Recovery Trajectory
#> =======================
#> Model: linear 
#> Muscles: 4 
#> R-squared range: 0.039 - 0.894