Rehabilitation following total knee replacement (TKR) principally aims to reduce pain and restore knee muscle strength and function, which necessitates objective monitoring. Given the growing interest in wearable technologies for rehabilitation, this paper investigates the feasibility of using electrical impedance tomography (EIT) to monitor and estimate muscle activity, specifically focusing on resultant knee angle (KA) and muscle strength (MS). A data collection system was designed to simultaneously capture EIT data, kinematic and physiological reference data, namely KA and MS, followed by two experiments. Data processing methods, such as region of interest (ROI) selection and modelling of the correlation between EIT and reference data, were used to assess feasibility. The results indicate that the long short-term memory (LSTM) model achieved an R2 value of 0.97, suggesting a strong correlation between the EIT data and KA. While EIT demonstrated potential in detecting KA, it did not show effectiveness in capturing changes in muscle activity and MS during isometric contractions.

This work presents a novel interaction method using bioimpedance (Bio-Z) to enhance virtual reality (VR) experience. The platform consists of two electrode bands (with 4 and 5 electrodes), a high-performance bioimpedance measurement system, a recurrent neural network deployed on a laptop and a VR headset. The system captures the Bio-Z features from muscle contraction and bone movement in the wrist and upper arm. After training, the regression model can predict hand grasp angle and forearm motion in both horizontal and vertical directions. This system can be used as a VR human-computer interaction (HCI) peripheral. Visitors will experience VR games using this demonstration platform.

This paper introduces a linear weighted integrate-and-fire (I&F) neuron architecture for Ion-Sensitive Field-Effect Transistors (ISFETs). It contributes to the next generation of neuromorphic lab-on-chip (LoC) platforms with the aim to integrate electrochemical sensors with neural networks on-silicon to compensate for sensor non-idealities. The neuron consists of a readout circuit, a linear voltage-controlled weighting circuit, and a robust I&F circuit with low power consumption. Notably, the readout in the neuron circuit achieves linear conversion of input voltage to output current at low power. This work also presents a cluster architecture for spatial correlation of trapped charge effects and process variations. The calibration system integrated in each cluster is realized using a current bit cell chain inspired from the current mode algorithmic ADC. The compensation range for each pixel ranges from −157.7 mV to 128.1 mV. The system is implemented as a $25\times 20$ cluster array, and the sensitivity of each cluster is 10.92 kHz/pH.