Constructing Mamdani-Intuitionistic Fuzzy Rules Set to Detect the Relaxed State by Transforming Spatio-Temporal EEG Data

The state of relaxation is an emotional condition that may be detected through various physical indicators in humans, such as sweat gland intensity, heart rate, breathing patterns, muscle tension and brain activity. Observation using physical patterns is challenging because each human has their own distinct pattern. In order to address this problem, neurologists are employing electroencephalography (EEG) to monitor the brain activity of patients. During this observation, neurologists depend on their own intuition to determine the relaxed state. However, the traditional observation method has drawbacks because each neurologist has their own subjective interpretation, which might lead to ambiguities. Thus, neurologists necessitate an automated recognition system capable of suggesting states of relaxation. To better assess the relaxed state, we divided the alpha band into two parts alpha sub-band: high and low alpha band. In order to obtain Spatio-Temporal features, both signals are transformed by wavelet. The ReliefF is used to select the features toward obtain optimal features. The maximum amplitude and standard deviation are two optimal features utilised as input to Mamdani-Intuitionistic Fuzzy Rules Set. The proposed approach is developed by integrating the fuzzy rules concept of Mamdani and Intuitionistic. In order to validate our model, we are collaborating with three neurologist experts and utilising majority decision to provide label annotation. According to this annotation, our model is performing well with an accuracy score of 92.45%. This investigation employs the DEAP public dataset. The level of accuracy seen in all examined subjects remained consistently high.