A Tri-network Model of Human Semantic Processing

Xu, Yangwen; He, Yong; Bi, Yanchao

doi:10.3389/fpsyg.2017.01538

Cited by 70 publications

(59 citation statements)

References 129 publications

(212 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These dimensions also tend to be associated with different large-scale brain networks, as shown here and previously (Jackson et al, 2016). The clustering of these different semantic dimensions may make the ATL one of the ideal regions to represent a multidimensional semantic space in a population-coding fashion by pooling multiple semantic and linguistic dimensions (Xu et al, 2017). Future studies should try to understand the relationships between the neuronal structures of these distinct ATL subregions and their corresponding semantic dimensions, as well as the regions outside the ATL for a given semantic dimension.…”

Section: Discussionsupporting

confidence: 57%

“…In addition to the three dimensions shown here, previous studies also reported the involvement of the ATL in artifacts (Bi et al, 2011), unique entities (Ross & Olson, 2012;Wang et al, 2016), concrete words in general Striem-Amit et al, 2018), and object perceptibility (Striem-Amit et al, 2018; for a review, see Lambon Ralph et al, 2017). The clustering of these different semantic dimensions may make the ATL one of the ideal regions to represent a multidimensional semantic space in a population-coding fashion by pooling multiple semantic and linguistic dimensions (Xu et al, 2017). The clustering of these different semantic dimensions may make the ATL one of the ideal regions to represent a multidimensional semantic space in a population-coding fashion by pooling multiple semantic and linguistic dimensions (Xu et al, 2017).…”

Section: Discussionsupporting

confidence: 51%

See 1 more Smart Citation

Close yet independent: Dissociation of social from valence and abstract semantic dimensions in the left anterior temporal lobe

Wang

Asan

2019

Human Brain Mapping

Self Cite

View full text Add to dashboard Cite

The anterior temporal lobe (ATL) is engaged in various types of semantic dimensions. One consistently reported dimension is social information, with abstract words describing social behaviors inducing stronger activations in the ATL than nonsocial words. One potential factor that has been systematically confounded in this finding is emotional valence, given that abstract social words tend to be associated with emotional feelings. We investigated which factors drove the ATL sensitivity using a 2 (social/nonsocial) × 2 (valenced/neutral) factorial design in an fMRI study with relatively high spatial resolutions. We found that sociality and valence were processed in different ATL regions without significant interactions: The social effect was found in the left anterior superior temporal sulcus (aSTS), whereas the valence effect activated small clusters in the bilateral temporal poles (TP). In the left ATL, the social‐ and valence‐related clusters were distinct from another superior ATL area that exhibited a general “abstractness” effect with little modulation of sociality or valence. These subregions exhibited distinct whole‐brain functional connectivity patterns during the resting state, with the social cluster functionally connected to the default mode network, the valence cluster connected to the adjacent temporal regions and amygdala, and the abstractness cluster connected to a distributed network including a set of language‐related regions. These results of activation profiles and connectivity patterns together indicate that the way in which the left ATL supports semantic processing is highly fine‐grained, with the neural substrate for social semantic effects dissociated from those for emotional valence and abstractness.

show abstract

Section: Discussionsupporting

confidence: 57%

Section: Discussionsupporting

confidence: 51%

Close yet independent: Dissociation of social from valence and abstract semantic dimensions in the left anterior temporal lobe

Wang

Asan

2019

Human Brain Mapping

Self Cite

View full text Add to dashboard Cite

show abstract

“…Another semantic model considered three functional networks as the basis of semantic processing, comprising (1) a perisylvian "language-supported system" (partially overlapping with the core language network as defined here), (2) a "multimodal experiential system" also addressed as the "default mode network" integrating experience-based knowledge across multiple modalities (see below), and (3) a left-dominant frontoparietal network as a semantic control system (Xu et al, 2017). These three networks are linked together in hub regions, comprising the anterior temporal lobe, posterior middle temporal gyrus, posterior intraparietal sulcus, angular gyrus, and parts of superior and middle frontal gyrus.…”

Section: The Semantic Systemmentioning

confidence: 99%

The Margins of the Language Network in the Brain

2020

View full text Add to dashboard Cite

This review paper summarizes the various brain modules that are involved in speech and language communication in addition to a left-dominant "core" language network that, for the present purpose, has been restricted to elementary formal-linguistic and more or less disembodied functions such as abstract phonology, syntax, and very basic lexical functions. This left-dominant perisylvian language network comprises parts of inferior frontal gyrus, premotor cortex, and upper temporal lobe, and a temporoparietal interface. After introducing this network, first, the various roles of neighboring and functionally connected brain regions are discussed. As a second approach, entire additional networks were considered rather than single regions, mainly motivated by resting-state studies indicating more or less stable connectivity patterns within these networks. Thirdly, some examples are provided for language tasks with functional demands exceeding the operating domain of the core language network. The rationale behind this approach is to present some outline of how the brain produces and perceives language, accounting, first, for a bulk of clinical studies showing typical forms of aphasia in case of left-hemispheric lesions in the core language network and second, for widespread activation patterns beyond this network in various experimental studies with language tasks. Roughly, the brain resources that complement the core language system in a task-specific way can be described as a number of brain structures and networks that are related to (1) motor representations, (2) sensory-related representations, (3) non-verbal memory structures, (4) affective/emotional processing, (5) social cognition and theory of mind, (6) meaning in context, and (7) cognitive control. After taking into account all these aspects, first, it seems clear that natural language communication cannot really work without additional systems. Second, it also becomes evident that during language acquisition the core language network has to be built up from outside, that is, from various neuronal activations that are related to sensory input, motor imitation, nursing, pre-linguistic sound communication, and pre-linguistic pragmatics. Furthermore, it might be worth considering that also in cases of aphasia the language network might be restored by being trained from outside.

show abstract

“…Although object knowledge can be acquired both through sensory experience (seeing red roses) and through language description (e.g., being told by others that roses are “red”), one prominent theory assumes that such knowledge is stored in sensory association cortices, derived from sensory experiences (i.e., seeing the colors of roses) (Barsalou et al, 2003; Martin, 2016; Simmons et al, 2007). The alternative possibility is that even sensory-derived knowledge is also represented at an abstract conceptual level distinct from sensory representations (Leshinskaya and Caramazza, 2016; Mahon and Caramazza, 2008; Shallice, 1988, 1987; Wang et al, 2018; Xu et al, 2017).…”

Section: Introductionmentioning

confidence: 99%

Two Forms of Knowledge Representations in the Human Brain

Wang

Men

Gao

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

AbstractSensory experience shapes what and how knowledge is stored in the brain -- our knowledge about the color of roses depends in part on the activity of color-responsive neurons based on experiences of seeing roses. We study the brain basis of color knowledge in congenitally blind individuals whose color knowledge can only be obtained through language descriptions. We found that some regions support color knowledge only in the sighted. More importantly, a region in the left dorsal anterior temporal lobe supports object color knowledge in both the blind and sighted groups, indicating the existence of a sensory-independent knowledge coding system in both groups. Thus, there are (at least) two forms of object knowledge representations in the human brain: sensory-derived and cognitively-derived knowledge, supported by different brain systems.

show abstract

A Tri-network Model of Human Semantic Processing

Cited by 70 publications

References 129 publications

Close yet independent: Dissociation of social from valence and abstract semantic dimensions in the left anterior temporal lobe

Close yet independent: Dissociation of social from valence and abstract semantic dimensions in the left anterior temporal lobe

The Margins of the Language Network in the Brain

Two Forms of Knowledge Representations in the Human Brain

Contact Info

Product

Resources

About