The exponential growth of multi-scale biomedical and behavioral data introduces both challenges and opportunities for Image 1 -driven analytics. Effectively managing the complexity and variability of these data sources requires advanced computational techniques for accurate interpretation and robust decision-making. Integrating Image 2 with colorimetric biosensing and multimodal data fusion offers scalable solutions that can improve diagnostic accuracy, enable early disease detection, and support personalized medicine. This work explores mobile-based colorimetry, an Image 3 -driven approach that uses image processing and Image 4 to detect colorimetric changes in chemical and biological solutions. We propose a modular conceptual framework that integrates mobile-based colorimetry with multimodal biomedical data, such as clinical, imaging, and environmental datasets, to develop scalable, low-cost tools for predictive modeling, real-time health monitoring, and personalized diagnostics. We review recent advancements in Image 5 -enabled colorimetric analysis and multimodal data fusion for healthcare applications, emphasizing innovations in Image 6 -assisted biosensors, Image 7 -driven biomedical imaging, and multimodal fusion techniques. In addition, we highlight the need for robust data management systems and interpretable AI/ML models to ensure security, privacy, and reliability in biomedical and behavioral research. This work also highlights practical directions for improving diagnostic accuracy and accessibility, particularly in resource-limited settings.

Phosphoenolpyruvate-carbohydrate phosphotransferase system (PTS) is a highly conserved, multistep chemical process which uses phosphate transfer to regulate the intake and use of sugars and other carbohydrates by bacteria. In addition to controlling sugar uptake, the PTS regulates several bacterial cellular functions such as chemotaxis, glycogen metabolism, catabolite repression and biofilm formation. Previous studies have shown that the phosphoenolpyruvate (PEP) to pyruvate ratio is a critical determinant of PTS functions. This study shows that 2-oxo-4-phenyl-2,5-dihydro-3-furancarbonitrile (MW01), a compound with structural similarity to known natural products, induces Vibrio cholerae to grow preferentially in the biofilm mode in a mechanism that involves interaction with pyruvate. Spectrophotometric assays were used to monitor bacterial growth kinetics in microtiter plates and quantitatively evaluate biofilm formation in borosilicate glass tubes. Evidence of MW01 and pyruvate interactions was determined by nuclear magnetic resonance spectroscopy. Given the established connection between PTS activity and biofilm formation, this study also highlights the potential impact that small-molecule modulators of the PTS may have in the development of innovative approaches to manage desired and undesired microbial cultures in clinical, industrial and environmental settings.

Bacterial adhesion to biotic and abiotic surfaces under fluid shear stress plays a major role in the pathogenesis of infections linked to medical implants and tissues. This study employed an automated BioFlux 200 microfluidic system and video microscopy to conduct real-time adhesion assays, examining the influence of shear stress on adhesion kinetics and spatial distribution of Staphylococcus aureus on glass surfaces. The adhesion rate exhibited a non-linear relationship with shear stress, with notable variations at intermediate levels. Empirical adhesion events were simulated with COMSOL Multiphysics^® and Python. Overall, COMSOL accurately predicted the experimental trend of higher rates of bacterial adhesion with decreasing shear stress but poorly characterized the plateauing phenomena observed over time. Python provided a robust mathematical representation of the non-linear relationship between cell concentration, shear stress, and time but its polynomial regression approach was not grounded on theoretical physical concepts. These insights, combined with advancements in AI and machine learning, underscore the potential for synergistic computational techniques to enhance our understanding of bacterial adhesion to surfaces, offering a promising avenue for developing novel therapeutic strategies.

Biofilms formed by Staphylococcus aureus on medical devices and tissue surfaces are a major contributor to persistent infections due to their resistance to antibiotics. Hydrodynamic forces in physiological and device-associated environments significantly influence biofilm development, yet the dynamics of detachment and regrowth under flow remain poorly quantified. In this study, biofilm surface coverage was measured in microfluidic flow assays across combinations of shear rates and nutrient concentrations. A computational workflow was used to segment biofilm trajectories into three kinetic phases—growth, exodus, and regrowth—based on surface coverage dynamics. Each phase was modeled using parametric functions, and fitted parameters were interpolated across experimental conditions to reconstruct biofilm lifecycles throughout the flow–nutrient conditions. The analysis revealed that intermediate shear rates triggered early detachment events while suppressing subsequent regrowth, whereas lower and higher shear regimes favored biofilm persistence. The resulting model enables quantitative comparison of condition-specific biofilm behaviors and identifies key thresholds in mechanical and nutritional inputs that modulate biofilm stability. These findings establish a phase-resolved framework for studying S. aureus biofilms under hydrodynamic stress and support future development of targeted strategies to control biofilm progression in clinical and engineered systems.

Physiological stress and emotion recognition are important for health monitoring and affective computing. In this work, we present a comprehensive evaluation of deep learning models such as Long Short-Term Memory (LSTM), Temporal Convolutional Networks (TCN), and Transformer on the WESAD dataset for multimodal affect recognition using wrist and chest sensor signals. We perform ablation studies to assess the individual contributions of each modality by training models on wrist-only and chest-only inputs. In addition, we implement a late-fusion ensemble strategy that combines predictions from all three architectures trained on multimodal input. We also employ early fusion at the sensor level by concatenating wrist and chest signals before feeding them into each model. Our results show that Transformer models consistently achieve the highest accuracy in multimodal settings, while TCN models perform best in the wrist-only configuration. The ensemble method yields the highest overall accuracy (98.91 ± 0.13%) and macro-F1 score (98.56 ± 0.17%). These findings demonstrate the effectiveness of sensor fusion and ensemble-based fusion in developing robust systems for physiological emotion recognition.