Innovative signal processing algorithms: Machine learning, deep learning, AI

The human brain generates an extraordinary volume of complex, high-dimensional signals, from electroencephalographic (EEG) and magnetoencephalographic (MEG) recordings to functional MRI time series and intracranial local field potentials. Decoding these signals with sufficient precision to advance our understanding of neural function and to translate that understanding into clinical tools, represents one of the defining challenges of modern neuroscience. Conventional signal processing approaches, while foundational, are increasingly insufficient in the face of data complexity, inter-subject variability, and the subtlety of neural dynamics that underlie cognition, perception, and disease.The rapid maturation of machine learning (ML), deep learning (DL), and artificial intelligence (AI) now offers transformative opportunities to overcome these limitations, enabling feature extraction, classification, and predictive modeling at a scale and resolution previously unattainable.This Research Topic invites original contributions that advance or critically evaluate the application of ML, DL, and AI-driven signal processing algorithms to human neuroscience data. We welcome work spanning methodological development, benchmarking, and applied studies, provided that a clear connection to neural signals or neuroscientific questions is maintained. Both invasive and non-invasive recording modalities are in scope, as are studies drawing on clinical or healthy populations.Contributions may address, but are not limited to, the following themes:Novel neural decoding architectures: Convolutional, recurrent, transformer-based, and hybrid deep learning models applied to EEG, MEG, fMRI, fNIRS, or multiunit recordings for decoding cognitive states, motor intentions, or affective responsesBrain-computer interfaces (BCIs): Real-time signal processing pipelines integrating adaptive ML algorithms Artifact removal and signal enhancement: Data-driven methods for separating neural signals from physiological and environmental noise, including independent component analysis extensions, autoencoders, and generative adversarial networksClinical translation: Applications in neurological and psychiatric disorders, including epilepsy, Parkinson’s disease, stroke, depression, and disorders of consciousness Topic editors: Baidaa Al-Bander, Fatma Gargouri Source: https://www.frontiersin.org/research-topics/82296/innovative-signal-processing-algorithms-machine-learning-deep-learning-ai