The field of human motion analysis and event-based vision is rapidly evolving, with a focus on developing innovative methods for estimating hand load, recognizing human activities, and predicting motion. Researchers are exploring the use of deep learning techniques, such as latent variable models and spiking neural networks, to improve the accuracy and efficiency of these systems. Additionally, there is a growing interest in leveraging auxiliary information, such as thermal sensing and baseline gait patterns, to enhance the performance of these models. The development of adaptive vision sampling methods and event autoencoders is also enabling more practical and power-efficient solutions for real-time motion analysis. Noteworthy papers in this area include Gait-Based Hand Load Estimation via Deep Latent Variable Models, which proposes a novel load estimation framework that incorporates auxiliary information to improve accuracy. Another notable paper is THOR: Thermal-guided Hand-Object Reasoning, which introduces a real-time adaptive spatio-temporal RGB frame sampling method that leverages thermal sensing to capture hand-object patches and classify them in real-time.