• 026990

    Single-cellular representations of semantic meaning during natural language perception

     

    Mohsen Jamali, Evelina Fedorenko, Jing Cai, Dan Lee, Benjamin Grannan, Ziv Williams

     

    Introduction: Humans extract meaning from language over fine temporal scales and myriad contexts.  Functional imaging studies suggest that language processing recruits distributed frontotemporal and subcortical networks.  The role of single neuron physiology in semantic processing, however, remains unknown.

    Objective: We test the hypothesis that dominant prefrontal neurons are involved in the rapid computation of semantic-level content during language comprehension.

    Methods: We performed acute recordings of left prefrontal neurons in 10 right-handed patients undergoing DBS surgery while participants listened to pre-recorded sentences and non-ordered word lists.  Study words were clustered into 9 semantic domains using a word-embedding and dimensionality reduction approach.  Single-neuron responses that were selective for certain semantic domains were detected by comparing firing rates in response to words from one domain with responses to all other words.  A categorical logistic regression model was used to test the semantic decoding capacity of the neuronal population.

    Results: We recorded from 220 prefrontal neurons, of which 47 (21.4%) were selective for at least one semantic domain (p<0.025, false-discovery correction).  Most neurons (N=31, 66.0%) were selective for one domain; no neurons were selective for more than 3 domains.  Population activity correctly predicted the instantaneous semantic domain 28% of the time (chance prediction=11.1%).  This pattern was specific to contextual language comprehension (i.e. sentences).  During word-list processing, only 5.5% of neurons demonstrated selective activity and population-level prediction performance was comparable with chance.

    Conclusions: During language comprehension, dominant prefrontal neurons demonstrate activity that is selective for contextual semantic content.  Such neurons likely contribute to the construction of meaning from language.

We use cookies to improve the performance of our site, to analyze the traffic to our site, and to personalize your experience of the site. You can control cookies through your browser settings. Please find more information on the cookies used on our site here. Privacy Policy