The Primacy of Experience in Language Processing: Semantic Priming Is Driven Primarily by Experiential Similarity.

The organization of semantic memory, including memory for word meanings, has long been a central question in cognitive science. Although there is general agreement that word meaning representations must make contact with sensory-motor and affective experiences in a non-arbitrary fashion, the nature of this relationship remains controversial. One prominent view proposes that word meanings are represented directly in terms of their experiential content (i.e., sensory-motor and affective representations). Opponents of this view argue that the representation of word meanings reflects primarily taxonomic structure, that is, their relationships to natural categories. In addition, the recent success of language models based on word co-occurrence (i.e., distributional) information in emulating human linguistic behavior has led to proposals that this kind of information may play an important role in the representation of lexical concepts. We used a semantic priming paradigm designed for representational similarity analysis (RSA) to quantitatively assess how well each of these theories explains the representational similarity pattern for a large set of words. Crucially, we used partial correlation RSA to account for intercorrelations between model predictions, which allowed us to assess, for the first time, the unique effect of each model. Semantic priming was driven primarily by experiential similarity between prime and target, with no evidence of an independent effect of distributional or taxonomic similarity. Furthermore, only the experiential models accounted for unique variance in priming after partialling out explicit similarity ratings. These results support experiential accounts of semantic representation and indicate that, despite their good performance at some linguistic tasks, the distributional models evaluated here do not encode the same kind of information used by the human semantic system.

The neural representation of body part concepts.

Neuropsychological and neuroimaging studies provide evidence for a degree of category-related organization of conceptual knowledge in the brain. Some of this evidence indicates that body part concepts are distinctly represented from other categories; yet, the neural correlates and mechanisms underlying these dissociations are unclear. We expand on the limited prior data by measuring functional magnetic resonance imaging responses induced by body part words and performing a series of analyses investigating the cortical representation of this semantic category. Across voxel-level contrasts, pattern classification, representational similarity analysis, and vertex-wise encoding analyses, we find converging evidence that the posterior middle temporal gyrus, the supramarginal gyrus, and the ventral premotor cortex in the left hemisphere play important roles in the preferential representation of this category compared to other concrete objects.

How does the "default mode" network contribute to semantic cognition?

This review examines whether and how the "default mode" network (DMN) contributes to semantic processing. We review evidence implicating the DMN in the processing of individual word meanings and in sentence- and discourse-level semantics. Next, we argue that the areas comprising the DMN contribute to semantic processing by coordinating and integrating the simultaneous activity of local neuronal ensembles across multiple unimodal and multimodal cortical regions, creating a transient, global neuronal ensemble. The resulting ensemble implements an integrated simulation of phenomenological experience - that is, an embodied situation model - constructed from various modalities of experiential memory traces. These situation models, we argue, are necessary not only for semantic processing but also for aspects of cognition that are not traditionally considered semantic. Although many aspects of this proposal remain provisional, we believe it provides new insights into the relationships between semantic and non-semantic cognition and into the functions of the DMN.

The Primacy of Experience in Language Processing: Semantic Priming Is Driven Primarily by Experiential Similarity.

The organization of semantic memory, including memory for word meanings, has long been a central question in cognitive science. Although there is general agreement that word meaning representations must make contact with sensory-motor and affective experiences in a non-arbitrary fashion, the nature of this relationship remains controversial. One prominent view proposes that word meanings are represented directly in terms of their experiential content (i.e., sensory-motor and affective representations). Opponents of this view argue that the representation of word meanings reflects primarily taxonomic structure, that is, their relationships to natural categories. In addition, the recent success of language models based on word co-occurrence (i.e., distributional) information in emulating human linguistic behavior has led to proposals that this kind of information may play an important role in the representation of lexical concepts. We used a semantic priming paradigm designed for representational similarity analysis (RSA) to quantitatively assess how well each of these theories explains the representational similarity pattern for a large set of words. Crucially, we used partial correlation RSA to account for intercorrelations between model predictions, which allowed us to assess, for the first time, the unique effect of each model. Semantic priming was driven primarily by experiential similarity between prime and target, with no evidence of an independent effect of distributional or taxonomic similarity. Furthermore, only the experiential models accounted for unique variance in priming after partialling out explicit similarity ratings. These results support experiential accounts of semantic representation and indicate that, despite their good performance at some linguistic tasks, the distributional models evaluated here do not encode the same kind of information used by the human semantic system.

A Distributed Network for Multimodal Experiential Representation of Concepts.

Neuroimaging, neuropsychological, and psychophysical evidence indicate that concept retrieval selectively engages specific sensory and motor brain systems involved in the acquisition of the retrieved concept. However, it remains unclear which supramodal cortical regions contribute to this process and what kind of information they represent. Here, we used representational similarity analysis of two large fMRI datasets with a searchlight approach to generate a detailed map of human brain regions where the semantic similarity structure across individual lexical concepts can be reliably detected. We hypothesized that heteromodal cortical areas typically associated with the default mode network encode multimodal experiential information about concepts, consistent with their proposed role as cortical integration hubs. In two studies involving different sets of concepts and different participants (both sexes), we found a distributed, bihemispheric network engaged in concept representation, composed of high-level association areas in the anterior, lateral, and ventral temporal lobe; inferior parietal lobule; posterior cingulate gyrus and precuneus; and medial, dorsal, ventrolateral, and orbital prefrontal cortex. In both studies, a multimodal model combining sensory, motor, affective, and other types of experiential information explained significant variance in the neural similarity structure observed in these regions that was not explained by unimodal experiential models or by distributional semantics (i.e., word2vec similarity). These results indicate that during concept retrieval, lexical concepts are represented across a vast expanse of high-level cortical regions, especially in the areas that make up the default mode network, and that these regions encode multimodal experiential information.SIGNIFICANCE STATEMENT Conceptual knowledge includes information acquired through various modalities of experience, such as visual, auditory, tactile, and emotional information. We investigated which brain regions encode mental representations that combine information from multiple modalities when participants think about the meaning of a word. We found that such representations are encoded across a widely distributed network of cortical areas in both hemispheres, including temporal, parietal, limbic, and prefrontal association areas. Several areas not traditionally associated with semantic cognition were also implicated. Our results indicate that the retrieval of conceptual knowledge during word comprehension relies on a much larger portion of the cerebral cortex than previously thought and that multimodal experiential information is represented throughout the entire network.

Decoding the information structure underlying the neural representation of concepts.

The nature of the representational code underlying conceptual knowledge remains a major unsolved problem in cognitive neuroscience. We assessed the extent to which different representational systems contribute to the instantiation of lexical concepts in high-level, heteromodal cortical areas previously associated with semantic cognition. We found that lexical semantic information can be reliably decoded from a wide range of heteromodal cortical areas in the frontal, parietal, and temporal cortex. In most of these areas, we found a striking advantage for experience-based representational structures (i.e., encoding information about sensory-motor, affective, and other features of phenomenal experience), with little evidence for independent taxonomic or distributional organization. These results were found independently for object and event concepts. Our findings indicate that concept representations in the heteromodal cortex are based, at least in part, on experiential information. They also reveal that, in most heteromodal areas, event concepts have more heterogeneous representations (i.e., they are more easily decodable) than object concepts and that other areas beyond the traditional "semantic hubs" contribute to semantic cognition, particularly the posterior cingulate gyrus and the precuneus.

Deep Artificial Neural Networks Reveal a Distributed Cortical Network Encoding Propositional Sentence-Level Meaning.

Understanding how and where in the brain sentence-level meaning is constructed from words presents a major scientific challenge. Recent advances have begun to explain brain activation elicited by sentences using vector models of word meaning derived from patterns of word co-occurrence in text corpora. These studies have helped map out semantic representation across a distributed brain network spanning temporal, parietal, and frontal cortex. However, it remains unclear whether activation patterns within regions reflect unified representations of sentence-level meaning, as opposed to superpositions of context-independent component words. This is because models have typically represented sentences as "bags-of-words" that neglect sentence-level structure. To address this issue, we interrogated fMRI activation elicited as 240 sentences were read by 14 participants (9 female, 5 male), using sentences encoded by a recurrent deep artificial neural-network trained on a sentence inference task (InferSent). Recurrent connections and nonlinear filters enable InferSent to transform sequences of word vectors into unified "propositional" sentence representations suitable for evaluating intersentence entailment relations. Using voxelwise encoding modeling, we demonstrate that InferSent predicts elements of fMRI activation that cannot be predicted by bag-of-words models and sentence models using grammatical rules to assemble word vectors. This effect occurs throughout a distributed network, which suggests that propositional sentence-level meaning is represented within and across multiple cortical regions rather than at any single site. In follow-up analyses, we place results in the context of other deep network approaches (ELMo and BERT) and estimate the degree of unpredicted neural signal using an "experiential" semantic model and cross-participant encoding.SIGNIFICANCE STATEMENT A modern-day scientific challenge is to understand how the human brain transforms word sequences into representations of sentence meaning. A recent approach, emerging from advances in functional neuroimaging, big data, and machine learning, is to computationally model meaning, and use models to predict brain activity. Such models have helped map a cortical semantic information-processing network. However, how unified sentence-level information, as opposed to word-level units, is represented throughout this network remains unclear. This is because models have typically represented sentences as unordered "bags-of-words." Using a deep artificial neural network that recurrently and nonlinearly combines word representations into unified propositional sentence representations, we provide evidence that sentence-level information is encoded throughout a cortical network, rather than in a single region.

An Integrated Neural Decoder of Linguistic and Experiential Meaning.

The brain is thought to combine linguistic knowledge of words and nonlinguistic knowledge of their referents to encode sentence meaning. However, functional neuroimaging studies aiming at decoding language meaning from neural activity have mostly relied on distributional models of word semantics, which are based on patterns of word co-occurrence in text corpora. Here, we present initial evidence that modeling nonlinguistic "experiential" knowledge contributes to decoding neural representations of sentence meaning. We model attributes of peoples' sensory, motor, social, emotional, and cognitive experiences with words using behavioral ratings. We demonstrate that fMRI activation elicited in sentence reading is more accurately decoded when this experiential attribute model is integrated with a text-based model than when either model is applied in isolation (participants were 5 males and 9 females). Our decoding approach exploits a representation-similarity-based framework, which benefits from being parameter free, while performing at accuracy levels comparable with those from parameter fitting approaches, such as ridge regression. We find that the text-based model contributes particularly to the decoding of sentences containing linguistically oriented "abstract" words and reveal tentative evidence that the experiential model improves decoding of more concrete sentences. Finally, we introduce a cross-participant decoding method to estimate an upper bound on model-based decoding accuracy. We demonstrate that a substantial fraction of neural signal remains unexplained, and leverage this gap to pinpoint characteristics of weakly decoded sentences and hence identify model weaknesses to guide future model development.SIGNIFICANCE STATEMENT Language gives humans the unique ability to communicate about historical events, theoretical concepts, and fiction. Although words are learned through language and defined by their relations to other words in dictionaries, our understanding of word meaning presumably draws heavily on our nonlinguistic sensory, motor, interoceptive, and emotional experiences with words and their referents. Behavioral experiments lend support to the intuition that word meaning integrates aspects of linguistic and nonlinguistic "experiential" knowledge. However, behavioral measures do not provide a window on how meaning is represented in the brain and tend to necessitate artificial experimental paradigms. We present a model-based approach that reveals early evidence that experiential and linguistically acquired knowledge can be detected in brain activity elicited in reading natural sentences.

Multiple Regions of a Cortical Network Commonly Encode the Meaning of Words in Multiple Grammatical Positions of Read Sentences.

Deciphering how sentence meaning is represented in the brain remains a major challenge to science. Semantically related neural activity has recently been shown to arise concurrently in distributed brain regions as successive words in a sentence are read. However, what semantic content is represented by different regions, what is common across them, and how this relates to words in different grammatical positions of sentences is weakly understood. To address these questions, we apply a semantic model of word meaning to interpret brain activation patterns elicited in sentence reading. The model is based on human ratings of 65 sensory/motor/emotional and cognitive features of experience with words (and their referents). Through a process of mapping functional Magnetic Resonance Imaging activation back into model space we test: which brain regions semantically encode content words in different grammatical positions (e.g., subject/verb/object); and what semantic features are encoded by different regions. In left temporal, inferior parietal, and inferior/superior frontal regions we detect the semantic encoding of words in all grammatical positions tested and reveal multiple common components of semantic representation. This suggests that sentence comprehension involves a common core representation of multiple words' meaning being encoded in a network of regions distributed across the brain.

Predicting Neural Activity Patterns Associated with Sentences Using a Neurobiologically Motivated Model of Semantic Representation.

We introduce an approach that predicts neural representations of word meanings contained in sentences then superposes these to predict neural representations of new sentences. A neurobiological semantic model based on sensory, motor, social, emotional, and cognitive attributes was used as a foundation to define semantic content. Previous studies have predominantly predicted neural patterns for isolated words, using models that lack neurobiological interpretation. Fourteen participants read 240 sentences describing everyday situations while undergoing fMRI. To connect sentence-level fMRI activation patterns to the word-level semantic model, we devised methods to decompose the fMRI data into individual words. Activation patterns associated with each attribute in the model were then estimated using multiple-regression. This enabled synthesis of activation patterns for trained and new words, which were subsequently averaged to predict new sentences. Region-of-interest analyses revealed that prediction accuracy was highest using voxels in the left temporal and inferior parietal cortex, although a broad range of regions returned statistically significant results, showing that semantic information is widely distributed across the brain. The results show how a neurobiologically motivated semantic model can decompose sentence-level fMRI data into activation features for component words, which can be recombined to predict activation patterns for new sentences.

Heteromodal Cortical Areas Encode Sensory-Motor Features of Word Meaning.

UNLABELLED: The capacity to process information in conceptual form is a fundamental aspect of human cognition, yet little is known about how this type of information is encoded in the brain. Although the role of sensory and motor cortical areas has been a focus of recent debate, neuroimaging studies of concept representation consistently implicate a network of heteromodal areas that seem to support concept retrieval in general rather than knowledge related to any particular sensory-motor content. We used predictive machine learning on fMRI data to investigate the hypothesis that cortical areas in this "general semantic network" (GSN) encode multimodal information derived from basic sensory-motor processes, possibly functioning as convergence-divergence zones for distributed concept representation. An encoding model based on five conceptual attributes directly related to sensory-motor experience (sound, color, shape, manipulability, and visual motion) was used to predict brain activation patterns associated with individual lexical concepts in a semantic decision task. When the analysis was restricted to voxels in the GSN, the model was able to identify the activation patterns corresponding to individual concrete concepts significantly above chance. In contrast, a model based on five perceptual attributes of the word form performed at chance level. This pattern was reversed when the analysis was restricted to areas involved in the perceptual analysis of written word forms. These results indicate that heteromodal areas involved in semantic processing encode information about the relative importance of different sensory-motor attributes of concepts, possibly by storing particular combinations of sensory and motor features. SIGNIFICANCE STATEMENT: The present study used a predictive encoding model of word semantics to decode conceptual information from neural activity in heteromodal cortical areas. The model is based on five sensory-motor attributes of word meaning (color, shape, sound, visual motion, and manipulability) and encodes the relative importance of each attribute to the meaning of a word. This is the first demonstration that heteromodal areas involved in semantic processing can discriminate between different concepts based on sensory-motor information alone. This finding indicates that the brain represents concepts as multimodal combinations of sensory and motor representations.

Toward a brain-based componential semantic representation.

Componential theories of lexical semantics assume that concepts can be represented by sets of features or attributes that are in some sense primitive or basic components of meaning. The binary features used in classical category and prototype theories are problematic in that these features are themselves complex concepts, leaving open the question of what constitutes a primitive feature. The present availability of brain imaging tools has enhanced interest in how concepts are represented in brains, and accumulating evidence supports the claim that these representations are at least partly "embodied" in the perception, action, and other modal neural systems through which concepts are experienced. In this study we explore the possibility of devising a componential model of semantic representation based entirely on such functional divisions in the human brain. We propose a basic set of approximately 65 experiential attributes based on neurobiological considerations, comprising sensory, motor, spatial, temporal, affective, social, and cognitive experiences. We provide normative data on the salience of each attribute for a large set of English nouns, verbs, and adjectives, and show how these attribute vectors distinguish a priori conceptual categories and capture semantic similarity. Robust quantitative differences between concrete object categories were observed across a large number of attribute dimensions. A within- versus between-category similarity metric showed much greater separation between categories than representations derived from distributional (latent semantic) analysis of text. Cluster analyses were used to explore the similarity structure in the data independent of a priori labels, revealing several novel category distinctions. We discuss how such a representation might deal with various longstanding problems in semantic theory, such as feature selection and weighting, representation of abstract concepts, effects of context on semantic retrieval, and conceptual combination. In contrast to componential models based on verbal features, the proposed representation systematically relates semantic content to large-scale brain networks and biologically plausible accounts of concept acquisition.

Concept Representation Reflects Multimodal Abstraction: A Framework for Embodied Semantics.

Recent research indicates that sensory and motor cortical areas play a significant role in the neural representation of concepts. However, little is known about the overall architecture of this representational system, including the role played by higher level areas that integrate different types of sensory and motor information. The present study addressed this issue by investigating the simultaneous contributions of multiple sensory-motor modalities to semantic word processing. With a multivariate fMRI design, we examined activation associated with 5 sensory-motor attributes--color, shape, visual motion, sound, and manipulation--for 900 words. Regions responsive to each attribute were identified using independent ratings of the attributes' relevance to the meaning of each word. The results indicate that these aspects of conceptual knowledge are encoded in multimodal and higher level unimodal areas involved in processing the corresponding types of information during perception and action, in agreement with embodied theories of semantics. They also reveal a hierarchical system of abstracted sensory-motor representations incorporating a major division between object interaction and object perception processes.

Predicting brain activation patterns associated with individual lexical concepts based on five sensory-motor attributes.

While major advances have been made in uncovering the neural processes underlying perceptual representations, our grasp of how the brain gives rise to conceptual knowledge remains relatively poor. Recent work has provided strong evidence that concepts rely, at least in part, on the same sensory and motor neural systems through which they were acquired, but it is still unclear whether the neural code for concept representation uses information about sensory-motor features to discriminate between concepts. In the present study, we investigate this question by asking whether an encoding model based on five semantic attributes directly related to sensory-motor experience - sound, color, visual motion, shape, and manipulation - can successfully predict patterns of brain activation elicited by individual lexical concepts. We collected ratings on the relevance of these five attributes to the meaning of 820 words, and used these ratings as predictors in a multiple regression model of the fMRI signal associated with the words in a separate group of participants. The five resulting activation maps were then combined by linear summation to predict the distributed activation pattern elicited by a novel set of 80 test words. The encoding model predicted the activation patterns elicited by the test words significantly better than chance. As expected, prediction was successful for concrete but not for abstract concepts. Comparisons between encoding models based on different combinations of attributes indicate that all five attributes contribute to the representation of concrete concepts. Consistent with embodied theories of semantics, these results show, for the first time, that the distributed activation pattern associated with a concept combines information about different sensory-motor attributes according to their respective relevance. Future research should investigate how additional features of phenomenal experience contribute to the neural representation of conceptual knowledge.

Where is the action? Action sentence processing in Parkinson's disease.

According to an influential view of conceptual representation, action concepts are understood through motoric simulations, involving motor networks of the brain. A stronger version of this embodied account suggests that even figurative uses of action words (e.g., grasping the concept) are understood through motoric simulations. We investigated these claims by assessing whether Parkinson's disease (PD), a disorder affecting the motor system, is associated with selective deficits in comprehending action-related sentences. Twenty PD patients and 21 age-matched controls performed a sentence comprehension task, where sentences belonged to one of four conditions: literal action, non-idiomatic metaphoric action, idiomatic action, and abstract. The same verbs (referring to hand/arm actions) were used in the three action-related conditions. Patients, but not controls, were slower to respond to literal and idiomatic action than to abstract sentences. These results indicate that sensory-motor systems play a functional role in semantic processing, including processing of figurative action language.

Parkinson's disease disrupts both automatic and controlled processing of action verbs.

The problem of how word meaning is processed in the brain has been a topic of intense investigation in cognitive neuroscience. While considerable correlational evidence exists for the involvement of sensory-motor systems in conceptual processing, it is still unclear whether they play a causal role. We investigated this issue by comparing the performance of patients with Parkinson's disease (PD) with that of age-matched controls when processing action and abstract verbs. To examine the effects of task demands, we used tasks in which semantic demands were either implicit (lexical decision and priming) or explicit (semantic similarity judgment). In both tasks, PD patients' performance was selectively impaired for action verbs (relative to controls), indicating that the motor system plays a more central role in the processing of action verbs than in the processing of abstract verbs. These results argue for a causal role of sensory-motor systems in semantic processing.

Common and dissociable prefrontal loci associated with component mechanisms of analogical reasoning.

The ability to draw analogies requires 2 key cognitive processes, relational integration and resolution of interference. The present study aimed to identify the neural correlates of both component processes of analogical reasoning within a single, nonverbal analogy task using event-related functional magnetic resonance imaging. Participants verified whether a visual analogy was true by considering either 1 or 3 relational dimensions. On half of the trials, there was an additional need to resolve interference in order to make a correct judgment. Increase in the number of dimensions to integrate was associated with increased activation in the lateral prefrontal cortex as well as lateral frontal pole in both hemispheres. When there was a need to resolve interference during reasoning, activation increased in the lateral prefrontal cortex but not in the frontal pole. We identified regions in the middle and inferior frontal gyri which were exclusively sensitive to demands on each component process, in addition to a partial overlap between these neural correlates of each component process. These results indicate that analogical reasoning is mediated by the coordination of multiple regions of the prefrontal cortex, of which some are sensitive to demands on only one of these 2 component processes, whereas others are sensitive to both.

Are cortical motor maps based on body parts or coordinated actions? Implications for embodied semantics.

The embodied cognition approach to the study of the mind proposes that higher order mental processes such as concept formation and language are essentially based on perceptual and motor processes. Contrary to the classical approach in cognitive science, in which concepts are viewed as amodal, arbitrary symbols, embodied semantics argues that concepts must be "grounded" in sensorimotor experiences in order to have meaning. In line with this view, neuroimaging studies have shown a roughly somatotopic pattern of activation along cortical motor areas (broadly construed) for the observation of actions involving different body parts, as well as for action-related language comprehension. These findings have been interpreted in terms of a mirror-neuron system, which automatically matches observed and executed actions. However, the somatotopic pattern of activation found in these studies is very coarse, with significant overlap between body parts, and sometimes with multiple representations for the same body part. Furthermore, the localization of the respective activations varies considerably across studies. Based on recent work on the motor cortex in monkeys, we suggest that these discrepancies result from the organization of the primate motor cortex (again, broadly construed), which probably includes maps of the coordinated actions making up the individual's motor repertoire, rather than a single, continuous map of the body. We review neurophysiological and neuroimaging data supporting this hypothesis and discuss ways in which this framework can be used to further test the links between neural mirroring and linguistic processing.

The effects of bilateral presentations on lateralized lexical decision.

We investigated how lateralized lexical decision is affected by the presence of distractors in the visual hemifield contralateral to the target. The study had three goals: first, to determine how the presence of a distractor (either a word or a pseudoword) affects visual field differences in the processing of the target; second, to identify the stage of the process in which the distractor is affecting the decision about the target; and third, to determine whether the interaction between the lexicality of the target and the lexicality of the distractor ("lexical redundancy effect") is due to facilitation or inhibition of lexical processing. Unilateral and bilateral trials were presented in separate blocks. Target stimuli were always underlined. Regarding our first goal, we found that bilateral presentations (a) increased the effect of visual hemifield of presentation (right visual field advantage) for words by slowing down the processing of word targets presented to the left visual field, and (b) produced an interaction between visual hemifield of presentation (VF) and target lexicality (TLex), which implies the use of different strategies by the two hemispheres in lexical processing. For our second goal of determining the processing stage that is affected by the distractor, we introduced a third condition in which targets were always accompanied by "perceptual" distractors consisting of sequences of the letter "x" (e.g., xxxx). Performance on these trials indicated that most of the interaction occurs during lexical access (after basic perceptual analysis but before response programming). Finally, a comparison between performance patterns on the trials containing perceptual and lexical distractors indicated that the lexical redundancy effect is mainly due to inhibition of word processing by pseudoword distractors.