Natural Language Processing Decodes Memory's Neural Patterns

A new study bridges neuroscience and artificial intelligence by using natural language processing to decode memory contents from neural patterns. Led by Kim, Koh, and Ranganath and published in Communications Psychology in 2026, the work combines functional MRI with advanced NLP to map the encoding and retrieval of autobiographical memories. The researchers exposed participants to complex stimuli, then asked them to recall details while brain activity was recorded, and used NLP to extract semantic features from the verbal recalls. By aligning these linguistic features with neural data, the team identified shared neural patterns linking memory encoding and retrieval, signaling a shift toward multimodal decoding of memory content with broad implications for mental health and AI.