Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Cellular‑Scale Language Processing Across Cortical Depth
Poster Session C, Thursday, October 1, 10:45 am - 12:45 pm, Wangari Maathai
Mohsen Jamali1, Benjamin Mesh, Jing Cai, Analise Bottinger, Richard Hardstone, William Muñoz, Brian Hsueh, Angelique Paulk, Sydney Cash, Ziv Williams; 1MGH, Harvard Medical School
The human capacity for communication depends on the brain’s ability to rapidly transform acoustic signals into meaningful linguistic representations and to generate appropriate spoken responses. The ability to perceive and interpret speech therefore relies on complex neural computations distributed across the human brain. Although large‑scale language networks have been extensively characterized, the fine‑scale cortical architecture and rapidly evolving neural dynamics that support speech processing at the level of individual cell populations remain poorly understood. Addressing this gap requires direct measurement of neural activity with both high spatial and temporal precision during naturalistic communication. In this study, we used ultra‑high‑resolution Neuropixels recordings in human participants as they engaged with a diverse set of speech conditions, including naturalistic listening, systematically manipulated linguistic stimuli (e.g., jabberwocky), and overt spoken responses. This approach enabled us to characterize columnar‑scale organization of neuronal activity across communicative contexts, capturing both receptive and expressive modes of language use with unprecedented resolution. By analyzing spiking activity and local field potentials across cortical layers, we identified neural populations in the human frontotemporal cortices that selectively tracked speech features and communicative demands. These populations exhibited distinct response profiles that differentiated among linguistic structures, levels of engagement, and modes of participation. Neural activity patterns varied systematically across superficial, middle, and deep cortical layers, revealing depth‑dependent computations that contributed uniquely to speech perception and production. Layer‑specific dynamics further highlighted how cortical columns integrate sensory input, linguistic structure, and communicative intent, pointing to functional specialization within the microcircuitry of human language‑related cortex. Together, these findings provide a high‑resolution view of the functional architecture underlying human speech processing. They demonstrate that the cortical column contains specialized, depth‑organized neural populations whose activity reflects both the structure of incoming speech and the behavioral context in which communication occurs. This work advances our understanding of the cellular‑level mechanisms that support language comprehension and production, offering new insight into how the human brain implements the computations necessary for flexible, interactive communication.
Topic Areas: Meaning: Lexical Semantics, Speech Perception