KANWISHER, N. G. | bioRxiv Science

bioRxiv2025

Visual experience contributes to separation of face and language responses in the ventral stream

Human ventral occipitotemporal cortex (vOTC) contains specialized regions that support visual recognition of behaviorally-relevant categories, including faces, written language and places (e.g., 1-4). An open question is how experience interacts with innate constraints to enable functional specialization. We investigate this question by comparing vOTC function across sighted and congenitally blind adults. In sighted adults, a region in lateral vOTC called the fusiform face area (FFA) responds preferentially to faces, whereas distinct left-lateralized portions of vOTC respond to written language 1,2,5-12. In blind people, lateral vOTC responds to face touching, braille and speech, but their functional co-localization has not been tested 13-16. The same group of congenitally blind adults (n=20) touched faces and spatial layouts (Experiment 1) and performed a reading (braille) and spoken language task (Experiment 2). Sighted adults performed analogous tasks in the visual modality (n=28). Using within subject analyses, we replicate the separation of faces and written language in sighted adults: written language responses are found only in left vOTC and within that hemisphere they are separate from faces. By contrast, left and right vOTC responds to language in people born blind and in the left hemisphere face and language responses overlap. These findings suggest that visual experience contributes to segregating responses to face and language in vOTC. Co-localization of face and language responses suggests an innate predisposition for communication-relevant processing in lateral vOTC.

neuroscience↗

bioRxiv2020

Artificial Neural Networks Accurately Predict Language Processing in the Brain

The neuroscience of perception has recently been revolutionized with an integrative modeling approach in which computation, brain function, and behavior are linked across many datasets and many computational models. By revealing trends across models, this approach yields novel insights into cognitive and neural mechanisms in the target domain. We here present a first systematic study taking this approach to higher-level cognition: human language processing, our species signature cognitive skill. We find that the most powerful transformer models predict nearly 100% of explainable variance in neural responses to sentences and generalize across different datasets and imaging modalities (fMRI, ECoG). Models neural fits ( brain score) and fits to behavioral responses are both strongly correlated with model accuracy on the next-word prediction task (but not other language tasks). Model architecture appears to substantially contribute to neural fit. These results provide computationally explicit evidence that predictive processing fundamentally shapes the language comprehension mechanisms in the human brain. SignificanceLanguage is a quintessentially human ability. Research has long probed the functional architecture of language processing in the mind and brain using diverse brain imaging, behavioral, and computational modeling approaches. However, adequate neurally mechanistic accounts of how meaning might be extracted from language are sorely lacking. Here, we report an important first step toward addressing this gap by connecting recent artificial neural networks from machine learning to human recordings during language processing. We find that the most powerful models predict neural and behavioral responses across different datasets up to noise levels. Models that perform better at predicting the next word in a sequence also better predict brain measurements - providing computationally explicit evidence that predictive processing fundamentally shapes the language comprehension mechanisms in the human brain.

neuroscience↗