Search for: All records

Abstract Brain dynamics are highly complex and yet hold the key to understanding brain function and dysfunction. The dynamics captured by resting-state functional magnetic resonance imaging data are noisy, high-dimensional, and not readily interpretable. The typical approach of reducing this data to low-dimensional features and focusing on the most predictive features comes with strong assumptions and can miss essential aspects of the underlying dynamics. In contrast, introspection of discriminatively trained deep learning models may uncover disorder-relevant elements of the signal at the level of individual time points and spatial locations. Yet, the difficulty of reliable training on high-dimensional low sample size datasets and the unclear relevance of the resulting predictive markers prevent the widespread use of deep learning in functional neuroimaging. In this work, we introduce a deep learning framework to learn from high-dimensional dynamical data while maintaining stable, ecologically valid interpretations. Results successfully demonstrate that the proposed framework enables learning the dynamics of resting-state fMRI directly from small data and capturing compact, stable interpretations of features predictive of function and dysfunction. 

In this work, we demonstrate an adjustable microfluidic tactile sensor for measurement of post-exercise response of local arterial parameters. The sensor entailed a polydimethylsiloxane (PDMS) microstructure embedded with a 5×1 resistive transducer array. The pulse signal in an artery deflected the microstructure and registered as a resistance change by the transducer aligned at the artery. PDMS layers of different thicknesses were added to adjust the microstructure thickness for achieving good sensor-artery conformity at the radial artery (RA) and the carotid artery (CA). Pulse signals of nine (n=9) young healthy male subjects were measured at-rest and at different times post-exercise, and a medical instrument was used to simultaneously measure their blood pressure and heart rate. Vibration-model-based analysis was conducted on a measured pulse signal to estimate local arterial parameters: elasticity, viscosity, and radius. The arterial elasticity and viscosity increased, and the arterial radius decreased at the two arteries 1min post-exercise, relative to at-rest. The changes in pulse pressure (PP) and mean blood pressure (MAP) between at-rest and 1min post-exercise were not correlated with that of heart rate and arterial parameters. After the large 1min post-exercise response, the arterial parameters and PP all went back to their at-rest values over time post-exercise.Clinical Relevance— The study results show the potential application of an affordable, user-friendly device for a more comprehensive arterial health assessment.