Title: AI, Innovation and Global Transformation: Interdisciplinary Perspectives on Technology, Business and Society
Editors: Dr. J. Preetha, and Dr. Siddhartha Mehrotra
ISBN: 978-81-69857-64-2
Chapter: 12
DOI: https://doi.org/10.59646/809/12
Author: Dr. R. Vijayalakshmi
Abstract
Brain–Computer Interfaces (BCIs) and neural communication frameworks provide a direct communication pathway between the human nervous system and external computational devices, bypassing conventional neuromuscular output channels. While invasive electrocorticography and non-invasive electroencephalography (EEG) have enabled motor prosthetics and assistive communication, real-time deployment is hindered by low signal-to-noise ratios, volume conduction, and non-stationary neural dynamics across sessions. This chapter introduces a Spatio-Temporal Riemannian Manifold Alignment and Adaptive Neural Decoding (STRM-AND) framework engineered for high-throughput, non-invasive neural communication. By coupling covariance-based Riemannian geometric projections with dynamic domain adaptation and attention-gated temporal convolutions, the proposed architecture aligns inter-session latent neural spaces and decodes high-dimensional motor imagery and steady-state visual evoked potentials (SSVEP). Benchmarked across multi-subject datasets, STRM-AND achieves an average information transfer rate of 148.5 bits/min and a decoding accuracy of 89.6%, reducing cross-session calibration overhead by 68.4% and advancing scalable, adaptive neurotechnologies.
Keywords: Brain–Computer Interfaces, Neural Decoding, Riemannian Geometry, Electroencephalography, Domain Adaptation, Information Transfer Rate