Figure 1. Heart rate per beat (left axis) and RMSSD over a sliding window of 30 beats (right axis). Dashed: mean heart rate (116 bpm).

Figure 2. RR tachogram — original intervals as recorded, before artefact correction. Band: mean ± 1 SD (518 ± 90 ms); marked: shortest and longest interval. Amber rings: beats flagged as artefacts (54), shown at their recorded value and corrected in the analysis series (Lipponen & Tarvainen 2019, duration kept).

Figure 3. Poincaré plot of successive RR pairs (corrected series). Ellipse with semi-axes SD1 (across the identity line, short-term variability) and SD2 (along it, long-term variability), centred on the mean RR (Brennan et al., 2001). Margins: distributions of RRₙ (top) and RRₙ₊₁ (right) in the grid's bin width.

Figure 4. Poincaré plot over time: the recording split into four consecutive quarters, on a common scale, each with its own SD1/SD2 ellipse (n = intervals of the corrected series ending in the window; segments.csv gives beats_recorded and beats_analysis). Grey: the whole recording. Possible because every beat keeps its position in time.

Figure 5. Power spectral density (Welch's method after cubic-spline resampling at 4 Hz). Shaded: LF 0.04–0.15 Hz and HF 0.15–0.40 Hz with absolute power, normalised units and peak frequency (dashed).

Figure 6. RR interval histogram of the corrected series, bin width 7.8125 ms (1/128 s; Task Force, 1996). Band: mean ± 1 SD. HRV triangular index = number of beats / height of the tallest bin.

Figure 7. RMSSD and SDNN per 10-min segment of the corrected series (RR_corrected_ms × corrected_count in the CSV). With L = 10 min, segment k spans [k·L, (k+1)·L) on the beat-end time axis (Timestamp_ms): a beat belongs to the segment in which it ends. Values above the bars; an incomplete final segment is not shown.

Figure 8. DFA α1 over time (Rogers et al., 2026): 120-s windows every 5 s, smoothness priors λ = 500 (Kubios default), box sizes 4–16 beats; each point at the window end. Grey ticks: windows without a valid value (5 % or more corrected beats, residual artefact, or α1 < 0). Dashed lines at 0.75 and 0.50 mark values discussed in the exercise literature, for orientation only — no thresholds are applied. Values: _dfa_alpha1.csv.
