Deep learning reveals cross-modal MEG representations of auditory and visual mental imagery

• A Journal of Neurophysiology Ahead-of-Print paper applies deep learning to MEG recordings of auditory and visual mental imagery.¹
• The study reports cross-modal neural representations shared across imagined auditory and visual content.¹
• Deep-learning models decode these imagery-related patterns from noninvasive MEG signals.¹
• Decoded signals reflect internal mental states rather than external sensory input alone.¹
• The authors frame the approach as relevant to imagery-based brain–computer interfaces.¹
• The work is primarily a methods paper focused on decoding performance rather than a clinical device path.^{1 1}

Footnotes

1. https://journals.physiology.org/doi/abs/10.1152/jn.00563.2025?af=R ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷