Rapid VR adoption across a variety of industries has improved user immersion while also posing problems like cybersickness, which has a substantial negative impact on user satisfaction and retention. Although the Simulator Sickness Questionnaire (SSQ) and other traditional measures have been used for a long time to assess symptoms, new technologies are opening up new assessment options. This study compares state-of-the-art models, such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Autoencoders, to examine recent developments in deep learning methods for detecting cybersickness. The paper also highlights gaps in present research methodologies and charts the evolution of algorithmic development.