Patient Subgroups Based on Modeled Recovery Trajectories.

Patient subgroups based on whether they showed exponential trajectories converging to an impairment z-score<−2.5 (subgroup ExpC), exponential trajectories converging to an impairment z-score of 0±2.5 (subgroup Exp) or linear trajectories (subgroup Lin). Parameter values are standardized to scores of healthy volunteers; lower values indicate a higher degree of impairment. Model parameters are: I = Intercept (initial deficit), β = exponent (rate of recovery), c = constant (plateau reached at the end of the observation period). Abbreviations: AICc = Akaike Information Criterion with bias correction for small sample sizes; No. = patient number, SSE = model sum of squared errors; w = normalized Akaike weights, i.e. conditional probability of chosen model given the data and the candidate set of models.
*Weighted mean and corresponding 95% CI, i.e. average of individual parameters multiplied with corresponding Akaike weight.