Recent advances in artificial intelligence (AI) and machine learning (ML) have transformed our ability to decode complex neural signals, opening new frontiers in understanding brain function and enabling practical applications such as brain-computer interfaces, neuroprosthetics, cognitive monitoring, and data security. Sophisticated deep learning architectures, generative models, and multimodal data fusion now allow the extraction of speech, imagery, emotional states, and intent directly from neural activity with unprecedented precision. Despite these breakthroughs, significant challenges remain in data acquisition, algorithmic transparency, and real-time deployment. Integrating heterogeneous data sources, such as fMRI, EEG, MEG, and intracranial recordings, while ensuring robust performance across individuals and tasks requires innovative methods that bridge neuroscience, computational modelling, and clinical translation.

This research topic aims to bring together cutting-edge research from neuroscience, computer science, and biomedical engineering to accelerate the development of AI-driven decoding methods. By adopting interdisciplinary collaboration, it will provide a platform to explore novel techniques, share benchmark datasets, and identify ethical frameworks for responsible use of brain-decoding technologies. We seek contributions that (i) enhance signal quality and lab-to-clinic transfer through domain-adaptive and causal representations; (ii) elevate fidelity in reconstructing language, vision, and fine-grained cognitive/affective states; (iii) achieve low-latency, closed-loop control for BCIs and neuroprosthetics via adaptive and resource-efficient inference; and (iv) establish end-to-end security, privacy, and governance for neurodata and networked BCI systems.

We invite submissions including Original Research, Methods, Brief Research Reports, Systematic Reviews, Mini Reviews, Technology & Code papers, Data Reports, Perspectives, and Opinions, provided they meet the journal’s standards for ethics, reproducibility, and transparent reporting (e.g., data/code availability, evaluation details, and limitations).

Suggested themes include, but are not limited to:

- AL/ML for Speech and Language Decoding: End-to-end AI pipelines for mapping neural signals to natural language and conversational speech.

- Visual and Perceptual Reconstruction: Generative and diffusion models for recreating visual experiences or imagined scenes from brain activity.

- Emotion, Cognition, and Mental State Modelling: AI methods to infer affective states, decision processes, memory, and attention from multimodal recordings.

- Multimodal Neural Data Fusion: Integration of fMRI, EEG, MEG, ECoG, and behavioural data for high-fidelity decoding and cross-subject generalisation.

- Real-Time Brain–Computer Interfaces and Neuroprosthetics: Low-latency algorithms and adaptive learning for closed-loop neural control.

- Neuromorphic and Edge AI Implementations: Energy-efficient architectures and on-device inference for scalable brain decoding.

- Explainability, Robustness, and Causality: Interpretable AI frameworks and causal inference for trustworthy neural decoding.

- Clinical and Translational Applications: AI-driven diagnostics and therapeutic systems for neurological and psychiatric disorders.

Are you interested in contributing a manuscript to this Research Topic?

AI for Decoding the Brain: Advances in Neural Signal Interpretation and Cognitive State Modelling