The name of the nose: Automatic fusiform face area activity predicts stimulus type before conscious categorization of face-related words

HESSE, EUGENIA; GARCÍA, ADOLFO M.; IBÁÑEZ, AGUSTÍN; SEDEÑO, LUCAS; ADOLFI, FEDERICO; MIKULAN, EZEQUIEL

Buenos Aires

Encuentro; 20th Annual Meeting of the Association for the Scientific Study of Consciousness; 2016

Association for the Scientific Study of Consciousness

Upon reading a word, we are able to make fast, explicit judgments on its meaning. However, this skill could be driven by fine-grained automatic processes operating significantly before we gain conscious access to relevant semantic information. Here, profiting from the excellent spatio-temporal resolution of intracranial recordings, we assessed whether face-related words yield ultra-rapid activity boosts in the highly specialized fusiform face area (FFA), and whether such patterns predict stimulus content prior to conscious categorization. Methods: Participants performed a semantic decision task on words denoting facial body parts, non-facial body parts, and familiar objects. Direct cortical recordings were obtained from implanted semi-rigid, multi-lead electrodes. We defined two target regions of interest (ROIs), one comprising key FFA hubs (right fusiform, ventral/rostral lingual, and calcarine gyri) and another one including temporal areas implicated in lexico-semantic processing (posterior middle temporal, rostral angular, and supramarginal gyri). A parietal and an occipital ROI were also included for control purposes. Time-frequency charts for each word type were obtained by analyzing the digitized signals via a windowed Fourier transform. Significant power increases and decreases across time against baseline values were analyzed with Monte Carlo permutation tests combined with bootstrapping. Also, logistic regression analyses were performed to evaluate whether power activity across each trial in very early windows predicted conscious stimulus categorization. Results: Relative to names of non-facial body parts and objects, words denoting facial body parts yielded significant activation increases in both target ROIs at low frequencies (0-20 Hz). This effect was stronger and earlier in the FFA (100-200 ms) than in the temporal (200-350 ms) ROI. In addition, logistic regressions based on such patterns (particularly, early FFA activity) predicted each word type with more than 70% accuracy. Discussion: Our findings unambiguously show that face-related words trigger preconscious embodiment effects which then lead to further operations in more general (arguably amodal) semantic networks. Crucially, ultra-rapid activity in the FFA consistently predicted word type prior to the subjects? conscious categorization. These results offer new insights into the interplay between automatic and conscious mechanisms during language processing, including robust empirical constraints for neurofunctional models of language. [Work partially supported by CONICET, CONICYT/FONDECYT Regular (1130920), COLCIENCIAS (1115-545-31374 and 1115-569-33858), FONCyT-PICT 2012-0412, FONCyT-PICT 2012-1309, FONDAP 15150012, and INECO Foundation.]