Research

Research topics

• Detecting atherosclerosis early and easily

  Early diagnosis support system for atherosclerosis

  A diagonal crease on the earlobe is known to be associated with atherosclerosis. We are developing deep-learning systems that estimate vascular health from images of the ear, and methods that assess endothelial function with the same oscillometric technique used in home blood pressure monitors, so that a vascular check-up no longer needs specialist equipment.

  Key papersAm J Cardiol 2017Sci Rep 2018

• Reading pain and stress from the stiffness of blood vessels

  Quantitative assessment of sensory perception based on vascular mechanical properties

  When we feel pain or stress, the sympathetic nervous system stiffens peripheral blood vessels. Using models that estimate arterial viscoelasticity from pulse waves, we turn sensations that are hard to put into words into objective measurements, and combine them with brain imaging using fMRI.

  Key papersSci Rep 2018Sci Rep 2021

• Measuring blood pressure and heart rate just by wearing or sitting

  Non-invasive continuous blood pressure monitoring system

  We develop technologies that measure blood pressure and heart rate continuously from pulse waves without squeezing the arm with a cuff, including a palpable pulse-wave sensor based on electromagnetic induction, unconstrained monitoring from the backrest of a chair, and an earphone-type device with biodegradable piezoelectric sensors.

  Key papersSci Rep 2019SII 2026

• Making cardiopulmonary bypass safer

  Prediction of haemoglobin concentration during cardiopulmonary bypass

  When cardiopulmonary bypass begins during heart surgery, the blood is diluted and haemoglobin concentration falls. We predict this value from patient data with machine learning, to support clinical engineers in preparing and managing the heart-lung machine.

  Key papersJ Extra Corpor Technol 2026

• Predicting recovery and complications after stroke

  Post-stroke delirium prediction system

  Using machine learning on clinical assessments and brain images of stroke patients, we have clarified how lesions and functions relate to mood disorders, driving ability, and swallowing and cognitive function. Building on these findings, we are developing a system that predicts post-stroke delirium at an early stage.

  Key papersSci Rep 2020J NeuroEng Rehabil 2023Sci Rep 2026

• Spotting changes in health early

  Predicting declines in health from biosignals

  Small changes can appear in biosignals such as heart rate and pulse waves before a person's health deteriorates. We aim to detect these changes with signal processing and machine learning and give early warning of illness.

  Key papersSci Rep 2020

• Supporting dialysis at home

  Vascular access monitoring system for home haemodialysis

  For patients on home haemodialysis, keeping the vascular access (an arteriovenous fistula created for dialysis) in good condition is essential. Working with the Blood Purification Centre at Fujita Health University Hospital, we aim to develop a system that lets patients check their access easily at home.