Deep learning improves classification of brain conscious states
Deep Learning Methods in Neuroscience: From Modeling Molecular Mechanisms to Classifying States of Consciousness
Computation and LanguageNeural and Evolutionary Computing
Summary
Knowing how conscious someone is can be hard to measure directly. This paper looks at how deep learning, a type of AI, helps classify brain states using data like EEG and brain scans. The authors found these AI methods work well for recognizing different states, especially when under anesthesia. However, they also note challenges like understanding how the AI makes decisions and finding markers that clearly indicate consciousness. Their review suggests combining AI with brain biology knowledge may help make better tools for brain monitoring in medicine.
What this means in practice
- •For clinical neuroscientists: Improve real-time monitoring of patients' consciousness levels during anesthesia using neural network analysis of EEG and LFP data.
- •For medical device developers: Develop physiologically informed hybrid AI models that classify brain states to enhance anesthesia safety and diagnosis accuracy.
A survey. It maps existing work.
Authors
Elena Benderskaya, Anastasiia Alifanova, Svetlana Batalova, Vasilisa Zhuk, Anna Kovalenko
Abstract
A critical analysis of contemporary approaches to the study of conscious states. The review focuses on methods of classification, clustering, modeling of brain states under anesthesia and identification of measurable neurobiological characteristics of brain function. A comparative analysis was conducted in the following three major areas: automatic detection of states of consciousness using neural networks based on EEG and fMRI data; modeling of the structural-functional dynamics of the brain under the effects of anesthetics; and detection of neurophysiological indicators which correlate with the level of consciousness. The obtained conclusions demonstrate the growing effectiveness of deep neural models in the classification and prediction of brain states and the analysis of dynamic structural-functional connectivity. Nonetheless, significant limitations were also identified, including the limited interpretability of the models, the lack of standardized metrics, and the problem of the specificity of consciousness markers. Our findings support the need for developing hybrid, generalizible, physiologically grounded architectures. Furthermore, such approaches may improve the translational potential of computational models in clinical neuroscience. Diverse methods of machine and computational modeling have demonstrated their effectiveness in tasks of automatic clustering and classification of brain states, the development of multilevel models and the identification of connectivity patterns correlated with levels of consciousness. A larger-scale analysis and a larger dataset, as well as the implementation of model interpretability approaches are required for the practical application of the analyzed models. The models based on EEG and LFP are the most promising for clinical application due to their availability and the possibility of real-time monitoring.