Language and semantic processing rely on complex neural networks that remain poorly understood at the single-neuron level. By leveraging recordings from patients with pharmacoresistant epilepsy who were implanted with depth electrodes, this research team explore single-neuron activity within the human amygdala and hippocampus.

Across several language tasks, the study provides insight into how individual neurons respond to linguistic input. Notably, during the natural story task, which most closely resembles real-life communication, more than half of the recorded neurons showed selective responses, highlighting a strong involvement of these structures in semantic comprehension.

Because SEEG offers millisecond temporal resolution and single-unit recordings capture activity at the cellular level, this dataset provides detailed characterization of single-neuron activity underlying language and semantic processing in the human medial temporal lobe. These recordings are made possible using micro-macro electrodes, which enable precise targeting and high-resolution data collection.

Importantly, the researchers have made all data and tasks openly available in the MATLAB format, enabling the scientific community to further investigate the cellular basis of human language.

link to the publication : https://www.nature.com/articles/s41597-025-05839-3

link to the dataset : https://osf.io/26sgq/

‍