ObjectiveEx vivo normothermic machine perfusion (NMP) supports kidney preservation and assessment, but conventional measurements mainly describe operating-point or endpoint conditions. This proof-of-concept study examined whether controlled flow perturbation and online system identification can characterize renal pressure–flow dynamics during rabbit kidney NMP.MethodsThree isolated rabbit kidneys (n=3) were perfused on a custom NMP platform. Pseudo-random binary sequence perturbations were applied to nominal pump-commanded perfusion flow, with renal arterial pressure as the output. Second-, third-, and fourth-order transfer-function models were identified using recursive least squares with adaptive multiple forgetting factors (RLS-MFF). Chronological online tracking was assessed using FIT and RMSE; fixed training-end models were evaluated using AIC/BIC, residual autocorrelation, and Ljung–Box testing. Parameter trajectories and frequency responses were also examined.ResultsOnline adaptive tracking accuracy increased with model order. For the selected third-order structure, online-evaluation FIT ranged from 84.81% to 86.00%, with a mean RMSE of 1.48 mmHg. Fourth-order models achieved the highest FIT and lowest RMSE but required two additional coefficients. AIC/BIC favored the third-order model in two records, while both third- and fourth-order structures satisfied the primary 10-lag Ljung–Box criterion in two of three records.ConclusionControlled hydraulic perturbation and adaptive online identification provided kidney-specific pressure–flow dynamic descriptors during rabbit kidney NMP. The third-order model offered a parsimonious common structure across records. With only three kidneys and no independent kidney-quality reference outcome, the findings demonstrate methodological feasibility rather than validated organ assessment.
Stroke is a leading cause of long-term disability, and severe upper-limb paresis often persists despite conventional rehabilitation. Brain–computer interfaces (BCIs) have been proposed to support motor recovery by pairing motor imagery with contingent sensory feedback, although clinical evidence remains heterogeneous, particularly for robot-assisted systems. This pilot study evaluates the feasibility of a noninvasive motor-imagery (MI)-based BCI delivering simultaneous multimodal feedback (embodied visual avatar and robot-assisted hand movement), and characterizes motor, cognitive, and functional outcomes in stroke patients receiving it alongside conventional therapy. Sixteen individuals with first-ever stroke were allocated to a control group (CG, n=6) receiving conventional rehabilitation or an experimental group (EG, n=10) receiving the same therapy plus 23 one-hour BCI sessions. Motor, cognitive, and functional outcomes were assessed pre- and post-intervention and reported descriptively, consistent with guidance for pilot and feasibility trials. The EG showed pre-post point estimates whose 95% confidence intervals excluded zero across several motor (Fugl–Meyer Assessment total score: 25.0 to 33.6), cognitive (episodic memory, visuoconstructive performance), and functional (general health, functional independence) measures. The CG showed a similar pattern for some measures, including functional independence and physical and social health perceptions, indicating these gains were not exclusive to the EG; between-group effect sizes favored the EG over the CG for episodic memory and semantic verbal fluency. Usability was positively rated (mean 76.3/100; 95% CI: 64.8–87.7). These preliminary findings support the feasibility of the proposed therapy and motivate larger, adequately powered, ideally sham-controlled trials to confirm added clinical benefit over conventional rehabilitation.
Background: Metabolic simulators help accelerate the development of diabetes technologies, but existing mechanistic models for type 2 diabetes (T2D) are difficult to individualize and provide limited explanation of the complex interplay between physiology and available treatments. Methods: First, we trained NeuralOGTT, a neural network state-space model that replicates glucose-insulin dynamics during the oral glucose tolerance test (OGTT) in T2D. Next, we determined the minimal model architecture using a variance-based sensitivity analysis. Finally, we created digital twins (DT-NeuralOGTT) by adding auxiliary sub-networks to the NeuralOGTT trained using individuals data from a free-living observational study (N=38, age: 50 ± 10 years; 27 females; 1 high risk of diabetes, 9 pre-T2D, and 28 T2D). We simulated study scenarios and assessed the digital twins ability to replicate real-world glucose outcomes: time in range (TIR,70−180 mg/dL), time in tight range (TITR,70−140 mg/dL), time below range (TBR, < 70 mg/dL), and time above range (TAR, > 180 mg/dL). Results: NeuralOGTT replicated glucose dynamics during OGTTs. Personalized DT-NeuralOGTT enabled generalization under free-living conditions across 12-h simulations, achieving simulated glucose outcomes that are statistically identical to observed outcomes: TIR: 91.8 (95% CI:82.2–96.0)% vs. 92.1 (95% CI:83.5–95.8)%, TITR: 56.0 (95% CI:36.6–77.4)% vs. 59.3 (95% CI:48.7–76.4)%, TAR: 6.3 (95% CI:1.6–15.8)% vs. 7.4 (95% CI:3.8–15.7)%, and TBR: 0.5 (95% CI:0.0–1.0)% vs. 0.1 (95% CI:0.1–0.2)%. Conclusion: NeuralOGTT models glucose–insulin dynamics using a structure that preserves explainability while requiring small amounts of training data. DT-NeuralOGTT incorporates individual variability to better capture dynamics in pre-T2D and T2D. Code available: https://github.com/mosqueralopez/T2DSim_AI
The assessment of cardiovascular dynamics from heart period (RR) and systolic arterial pressure (SAP) time series requires analytical tools capable of capturing the complexity and nonlinearity of the underlying regulatory mechanisms and the directionality of their interactions. This work presented and tested a methodological framework combining Mutual Information Rate (MIR) with multiple surrogate data analysis to quantify and statistically validate complexity, causality, and nonlinearity in cardiovascular dynamics. Specifically, MIR was estimated through a model-free nearest-neighbor approach and decomposed into measures of complexity (i.e., entropy rate) and causality (i.e., transfer entropy). The capability of the method to characterize cardiovascular regulation was evaluated across young normotensive healthy subjects, older healthy individuals, and post-acute myocardial infarction (AMI) patients, undergoing an orthostatic stress test. In healthy young subjects, orthostatic stress induced the expected reduction in cardiovascular entropy rate and nonlinear coupling, reflecting a regulated baroreflex response. In contrast, older adults and post-AMI patients exhibited altered dynamics already at rest and lacked the adaptive modulation observed in the young group. These results demonstrated the potential of MIR decomposition, supported by surrogate analysis, as a robust framework to characterize the modulation of cardiovascular control mechanisms in response to orthostatic stress and their alterations across pathophysiological conditions.
