Cogni-Sonance: Navigating the Neurological Orchestra of Imagined Speech

The silent expression of language in the mind, or imagined speech, is an intriguing yet mysterious feature of human thought. This paper offers a thorough analysis of the illusory speech using Electroencephalography signals. We examine the complex interactions between brain oscillations and functional connectivity patterns related to imagined speech tasks, building on recent developments in EEG methodology and processing tools. In order to clarify the cortical networks and activation patterns involved in the production and processing of inner speech, we investigate the spatiotemporal dynamics of EEG signals. The dataset contains 3 different paradigms of imagination i.e. imagined digits, imagined alphabets and imagined images. The bagged tree model has given the best classification accuracy. For three class the classification accuracy is 56.7% and for binary class the maximum obtained accuracy is 70.1%. We have also emphasized on having lighter models which have less training time. EEG research provides us with a window into the inner workings of the human mind and may open up new avenues for helping those who struggle with language comprehension or speech. Through the use of EEG to decode the neurological foundations of imagined speech, we have found the source localization for all the different modalities of imagination. we hope to shed light on human inner monologue and open new avenues for creative applications in neurorehabilitation, brain-computer interfaces, and mental health therapies.