FPCA-based method to select optimal sampling schedules that capture between-subject variability in longitudinal studies