Diabetes alters cardiorespiratory dynamics: insights from short-term recurrence quantification analysis of pulse-respiration quotient

FIGURE 1. Schematic representation of the electrode configuration for electrocardiogram (ECG) acquisition and the positioning of the thoracic respiratory sensor. The ECG was recorded using a three-electrode bipolar lead system, with the negative electrode placed on the right infraclavicular region, the positive electrode on the left infraclavicular region, and the reference electrode positioned on the left lateral chest at the level of the fifth rib. The thoracic respiratory sensor belt was positioned around the chest to monitor respiratory activity (RESP).

FIGURE 2. Assessment of the autocorrelation function for each Pulse-Respiration Quotient (PRQ) time series (thin lines) and averaged values per group (thick lines). The panels compare control participants (upper row) and individuals with type 2 diabetes mellitus (T2DM, lower row) across four experimental conditions: supine rest (S), orthostatic challenge (O), paced breathing (R), and the Valsalva maneuver (V). The autocorrelation function quantifies the temporal dependencies in the PRQ time series, providing insight into the underlying dynamics of autonomic regulation in response to different physiological challenges.

FIGURE 3. Assessment of the false nearest neighbors’ method for each Pulse-Respiration Quotient (PRQ) time series (thin lines) and averaged values per group (thick lines). The panels compare control participants (upper row) and individuals with type 2 diabetes mellitus (T2DM, lower row) across four experimental conditions: supine rest (S), orthostatic challenge (O), paced breathing (R), and the Valsalva maneuver (V). The false nearest neighbors’ method estimates the optimal embedding dimension for nonlinear analysis, helping to determine the minimal dimensionality required to reconstruct the system’s phase space and assess its underlying complexity.

FIGURE 4. Comparison of recurrence quantification analysis (RQA) indices and linear indices for Pulse-Respiration Quotient (PRQ) time series across different experimental stages: supine rest (S), Orthostatic challenge (O), Paced respiration (R), and Valsalva maneuver (V) for control (blue) and diabetic (red) groups. Panels represent the following indices: (a) Trapping time (TT), (b) Entropy (ENT), (c) Recurrence time of first type (T1), (d) Recurrence time of second type (T2), (e) Mean of PRQ (mPRQ), and (f) Standard deviation of PRQ (SDPRQ). Data are expressed as mean ± SEM. Significant differences are indicated as follows: *p < 0.05 between Control vs T2DM; p < 0.05 for #(S vs R), ¬(S vs O),-(S vs V), &(O vs R), +(O vs V), and ^(R vs V). The color of the symbol represents the group: blue for the control group and red for the diabetic group.

FIGURE 5. Comparison of recurrence quantification analysis (RQA) indices for Pulse-Respiration Quotient (PRQ) time series across different experimental stages: supine rest (S), Orthostatic challenge (O), Paced respiration (R), and Valsalva maneuver (V) for control (blue) and diabetic (red) groups. Panels represent the following RQA indices: (a) Recurrence Rate (RR), (b) Determinism (DET), (c) Mean diagonal line length (L), (d) Longest diagonal line (LMAX), (e) Laminarity (LAM), and (f) Maximum vertical line length (VMAX). Data are expressed as mean ± SEM. Significant differences are indicated as follows: **p < 0.01 between Control vs T2DM; p < 0.05 for #(S vs R), and ^(R vs V). The symbol’s color represents the group: blue for the Control group and red for the T2DM group.

FIGURE 6. Representative example of Pulse-Respiration Quotient (PRQ) time series and corresponding recurrence plots across four experimental stages—supine rest (S), orthostatic challenge (O), paced respiration (R), and Valsalva maneuver (V)—for a control participant (blue) and a participant with type 2 diabetes mellitus (T2DM, red). In this 4 × 4 panel figure, each row corresponds to a stage (S, O, R, V) while the first and second columns display the PRQ time series for the control and T2DM groups, respectively, and the third and fourth columns show the corresponding recurrence plots. Panels (a), (b), (e), (f), (i), (j), (m), and (n) correspond to PRQ time series, and panels (c), (d), (g), (h), (k), (l), (o), and (p) to recurrence plots (RP).

FIGURE 7. Mutual Information (MI) as a function of time delay (τ) ranging from 1 to 30, computed between interpolated R-to-R peak interval (RRI) and breath-to-breath (BB) signals in control (blue) and type 2 diabetes mellitus (T2DM, red) groups. (a) MI during the Paced respiration (R) stage, where the control group exhibits consistently higher MI values along delays compared to the T2DM group. (b) MI during the Valsalva maneuver (V) stage, showing an opposite trend, where the T2DM group presents higher MI values at larger time delays. Data are expressed as mean ± SEM.