EEG-based Imagined Speech Analysis using Functional Connectivity
Meenakshi Bisla1, Radhey Shyam Anand2
1Dr. Meenakshi Bisla, Department of Computer Science, UPES, Dehradun (Uttarakhand), India.
2Dr. Radhey Shyam Anand, Department of Electrical Engineering, Indian Institute of Technology, Roorkee, India.
Manuscript received on 25 February 2026 | First Revised Manuscript received on 02 March 2026 | Second Revised Manuscript received on 10 March 2026 | Manuscript Accepted on 15 March 2026 | Manuscript published on 30 March 2026 | PP: 22-28 | Volume-16 Issue-1, March 2026 | Retrieval Number: 100.1/ijsce.A371116010326 | DOI: 10.35940/ijsce.A3711.16010326
Open Access | Editorial and Publishing Policies | Cite | Zenodo | OJS | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Imagined speech involves a network of brain regions that simulate the experience of speaking or listening, though it lacks the physical movement of the vocal apparatus or external sound. This study analyses the brain’s role in the imagination of words, sentences, and vowels using an amalgamation of innovative methods, including transfer entropy, graph theory, and functional connectivity. Channels that exhibit stronger connectivity than most of the network play a crucial role in information processing or network integration. This helps analyze the most “active” or “influential” channels of imagination of any class. By selecting channels with connectivity values significantly above the average (i.e., those exceeding the mean + 1.5 standard deviations), this method ensures that you focus on the most distinctive and potentially relevant patterns in the data. The analysis is performed on the original EEG dataset acquired. The proposed analysis is also validated using public datasets (Kara one and ASU) to assess the reliability of the methodology. Overall, the findings support the hypothesis that imagined speech engages a distributed but left leaning network of regions, with task-specific patterns modulated by the complexity and phonetic structure of the stimuli.
Keywords: EEG, Cognitive Analysis, Functional Connectivity, Imagined Speech, Graph Theory.
Scope of the Article: Data Analytics