Source Reconstruction of Brain Potentials Using Bayesian Model Averaging to Analyze Face Intra-Domain vs. Face-Occupation Cross-Domain Processing

Abstract

We investigated the neural correlates of the access to and retrieval of face structure information in contrast to those concerning the access to and retrieval of person-related verbal information, triggered by faces. We experimentally induced stimulus familiarity via a systematic learning procedure including faces with and without associated verbal information. Then, we recorded event-related potentials (ERPs) in both intra-domain (face-feature) and cross-domain (face-occupation) matching tasks while N400-like responses were elicited by incorrect eyes-eyebrows completions and occupations, respectively. A novel Bayesian source reconstruction approach plus conjunction analysis of group effects revealed that in both cases the generated N170s were of similar amplitude but had different neural origin. Thus, whereas the N170 of faces was associated predominantly to right fusiform and occipital regions (the so-called “Fusiform Face Area”, “FFA” and “Occipital Face Area”, “OFA”, respectively), the N170 of occupations was associated to a bilateral very posterior activity, suggestive of basic perceptual processes. Importantly, the right-sided perceptual P200 and the face-related N250 were evoked exclusively in the intra-domain task, with sources in OFA and extensively in the fusiform region, respectively. Regarding later latencies, the intra-domain N400 seemed to be generated in right posterior brain regions encompassing mainly OFA and, to some extent, the FFA, likely reflecting neural operations triggered by structural incongruities. In turn, the cross-domain N400 was related to more anterior left-sided fusiform and temporal inferior sources, paralleling those described previously for the classic verbal N400. These results support the existence of differentiated neural streams for face structure and person-related verbal processing triggered by faces, which can be activated differentially according to specific task demands.

External (Versus Internal) Facial Features Contribute Most to Repetition Priming in Facial Recognition: ERP Evidence

“…Consistent with our expectation and previous research (Alzueta et al., 2019; Feuerriegel et al., 2019; Hashemi et al., 2019; Herzmann, 2016; Huang et al., 2017; Mohr et al., 2018; Olivares et al., 2018; Özkaya & Yıldırım, 2018; Schweinberger & Neumann, 2016; Wiese et al., 2013), we observed that N250r and N400 activations seemed to represent, respectively, the processes of accessing a stored structural representation of a face and retrieving knowledge about a person from long-term memory. We observed that both familiar faces and familiar external feature faces were associated with reliable N250r activity for the positive-going sub-component of N250r over prefrontal and frontal regions, indicating that these two types of target faces could prompt the brain to match the input with a known face’s representation, as activated by the FRU, corresponding to similar findings of previous investigators who used famous/familiar faces as stimuli (Engst et al., 2006; Mohr et al., 2018; Nie, Li, et al., 2016; Nie, Ye, et al., 2016).…”

“…On the other hand, all of our three types of targets (familiar faces, familiar internal feature faces, and familiar external feature faces) were associated with reliable N400 activity, consistent with findings from previous investigations with famous/familiar faces (Huang et al., 2017; Olivares et al., 2018; Özkaya & Yıldırım, 2018; Schweinberger & Neumann, 2016). As mentioned in our Introduction, the function of N400 activity is to trigger the PIN to access the semantic information about the person being identified by the FRU or to indicate the retrieval or activation of knowledge relevant to the faces from long-term memory during repetition priming (Huang et al., 2017; Olivares et al., 2018; Özkaya & Yıldırım, 2018; Schweinberger & Neumann, 2016). Thus, within the time course of such priming, the waveform differences between the presentations of primes and targets are all termed as N400, regardless of whether the amplitudes of the waveforms of the target stimuli are more positive- or negative-going than those of the primes (Schweinberger & Neumann, 2016).…”

“…Extensive studies have shown that the N400 ERP component typically peaks at around 400 ms post-stimulus onset, predominantly over the centro-parietal region (Kutas & Federmeier, 2011; Olivares et al., 2018; Özkaya & Yıldırım, 2018; Schweinberger & Neumann, 2016; Wakui et al., 2016). Many other studies have found, however, that this component could be simultaneously triggered over the prefrontal and/or frontal regions (Nie, Li, et al., 2016; Nie, Ye, et al., 2016).…”

“…Many other studies have found, however, that this component could be simultaneously triggered over the prefrontal and/or frontal regions (Nie, Li, et al., 2016; Nie, Ye, et al., 2016). Various studies have assumed, in the case of repetition priming of faces, that N400 activation reflects the processing of semantic knowledge about a person whose face is being retrieved from long-term memory (Huang et al., 2017; Olivares et al., 2018; Özkaya & Yıldırım, 2018; Schweinberger & Neumann, 2016). Similar to N250r, the amplitudes of N400 are much larger for famous/familiar faces vs unknown faces, strongly suggesting that the pre-existing (or experimentally triggered) memory representation serves as a mediator for N400 activation (Schweinberger & Neumann, 2016).…”

Role of line junctions in expert object recognition: The case of musical notation

“…ERP studies found that the N170 component supports expert object categorization (Bentin et al, 1996; Maurer et al, 2008; Scott et al, 2006) while the N250 component supports expert object individuation (Gentile & Jansma, 2012; Rossion & Jacques, 2012; Scott et al, 2006). Because both ERP components are potentially generated at the fusiform gyrus (Bobes et al, 2019; Kaufmann et al, 2009; Olivares et al, 2018; Schweinberger et al, 2002), these results show that expert object recognition is supported by the higher visual cortex, in particular, the fusiform gyrus and the inferotemporal cortex.…”