Machine learning enhanced noninvasive transcranial stroke detection using a portable eddy current damping sensor.

Stroke remains a leading cause of morbidity and mortality globally, with timely diagnosis critical for effective treatment. Current imaging modalities, such as computed tomography (CT) and magnetic resonance imaging (MRI), often face limitations in availability and timeliness, leading to diagnostic delays that worsen patient outcomes. This study presents the development and validation of a portable, noninvasive eddy current damping (ECD) sensor for rapid stroke detection. Building on previous research and optimized through benchtop and phantom studies, this device incorporates advanced coil designs and algorithms to distinguish between hemorrhagic stroke patients and healthy individuals by detecting electrical conductivity variations between normal brain tissue and accumulated blood. Results demonstrate high degrees of accuracy and specificity, promising to enable real-time bedside differentiation of stroke patients. This ECD sensor may expand access to timely stroke care across prehospital, clinical, and remote settings and improve patient outcomes by expediting targeted treatment and rapid triage.