Development of a Novel Non-Invasive Blood Pressure Measurement Technique Using Near-Infrared Spectroscopy and Deep Neural Networks

Non-invasive monitoring of blood pressure is essential for health management in daily life, particularly for elderly individuals and patients with autonomic dysfunction. However, conventional cuff-based sphygmomanometers are uncomfortable and provide discontinuous monitoring. In this study, we propose a novel non-invasive technique for blood pressure change assessment based on prefrontal near-infrared spectroscopy (NIRS) and deep neural networks. Unlike existing approaches that rely primarily on electrocardiography or photoplethysmography signals, the proposed method independently utilizes prefrontal NIRS signals to capture optical responses associated with postural changes. This enables early prediction of blood pressure changes following standing. NIRS is a non-invasive, real-time, portable, high temporal resolution, and cost-effective in comparison with other brain monitoring modalities. NIRS signals were collected from 40 healthy adults (aged 19-39 years) as test and training groups, 40 entirely new volunteers as an independent evaluation group, and during sit-to-stand transitions using a two-channel system, while reference blood pressure measurements were obtained using an automatic sphygmomanometer. An adaptive filtering framework was applied to suppress physiological noise, and multiple deep learning architectures were evaluated for regression analysis. A hybrid convolutional neural network-gated recurrent unit (CNN-GRU) model achieved the best performance, yielding root mean square errors of 1.95 mmHg and 1.75 mmHg for systolic and diastolic blood pressure, respectively. The results demonstrate that prefrontal NIRS combined with deep learning provides a reliable and accurate framework for non-invasive blood pressure change monitoring. This technique shows strong potential for wearable applications in elderly populations and individuals at risk of orthostatic hypotension, and it establishes a foundation for future clinical deployment.