The field of micro-expression recognition and analysis is moving towards more robust and accurate frameworks, with a focus on addressing key challenges such as key-frame index errors and information redundancy. Recent developments have introduced novel architectures and techniques, including temporal state transition mechanisms and motion prompt tuning, to improve the characterization of micro-expressions and their recognition. Additionally, there is a growing interest in multimodal approaches, including the development of custom attention networks for recognizing subjective self-disclosure in human-robot interactions. Noteworthy papers include: MPT, which introduces a pioneering method for subtle motion prompt tuning, and ME-TST+, which proposes a synergy strategy for spotting and recognition at both the feature and result levels. ME-TST+ achieves state-of-the-art performance, and MPT consistently surpasses state-of-the-art approaches on widely used MER datasets.