Role of AI and AI-Derived Techniques in Brain and Behavior Computing

Human brains are full of neurons and nerve cells connected to each other through dendrites and axons. These neurons and nerve cells are at work for every action and help brain in learning. Brain integrates sensory information and directs motor responses through the limbic system and neocortex. The limbic brain is responsible for primal and survival urges whereas neocortex deals with logical functions. The brain computing is conceptual and computational information processing within the human brain and brain’s architecture. Brain–Computer Interfaces (BCIs) acquire brain signals, analyze them, and translate them into instructions to be performed by machines. But the most important challenge for the BCI is that everyone has different brain. Machine learning enables machines to detect complex patterns, automatically generate hypotheses, and create interpretable models revealing biological mechanisms. Using Machine Learning (ML), for every new session, the BCI must learn from the user’s brain by adapting so as to properly classify its thoughts. In this chapter, machines learning basics and existing models for classification, regression, and clustering methods for BCI are explored. Available deep learning models for brain computing are also investigated. Deep learning models like restricted Boltzmann machine (RBM), recurrent neural network (RNN), and long short-term memory (LSTM) with a high number of parameters are found to perform better for brain signals. Results obtained for Convolution Neural Network (CNN) and combination of CNN with modified RNN are comparable with existing BCI models. Performance parameters and databases play very important role for the validation of results. Standard performance parameters and available databases and their impact on results are also analyzed.