A neuroanatomically grounded Hebbian-learning model of attention-language interactions in the human brain

Distributed representations of prediction error signals across the cortical hierarchy are synergistic

A relevant question concerning inter-areal communication in the cortex is whether these interactions are synergistic. Synergy refers to the complementary effect of multiple brain signals conveying more information than the sum of each isolated signal. Redundancy, on the other hand, refers to the common information shared between brain signals. Here, we dissociated cortical interactions encoding complementary information (synergy) from those sharing common information (redundancy) during prediction error (PE) processing. We analyzed auditory and frontal electrocorticography (ECoG) signals in five common awake marmosets performing two distinct auditory oddball tasks and investigated to what extent event-related potentials (ERP) and broadband (BB) dynamics encoded synergistic and redundant information about PE processing. The information conveyed by ERPs and BB signals was synergistic even at lower stages of the hierarchy in the auditory cortex and between auditory and frontal regions. Using a brain-constrained neural network, we simulated the synergy and redundancy observed in the experimental results and demonstrated that the emergence of synergy between auditory and frontal regions requires the presence of strong, long-distance, feedback, and feedforward connections. These results indicate that distributed representations of PE signals across the cortical hierarchy can be highly synergistic.

The impact of early and late blindness on language and verbal working memory: A brain-constrained neural model

Neural circuits related to language exhibit a remarkable ability to reorganize and adapt in response to visual deprivation. Particularly, early and late blindness induce distinct neuroplastic changes in the visual cortex, repurposing it for language and semantic processing. Interestingly, these functional changes provoke a unique cognitive advantage - enhanced verbal working memory, particularly in early blindness. Yet, the underlying neuromechanisms and the impact on language and memory-related circuits remain not fully understood. Here, we applied a brain-constrained neural network mimicking the structural and functional features of the frontotemporal-occipital cortices, to model conceptual acquisition in early and late blindness. The results revealed differential expansion of conceptual-related neural circuits into deprived visual areas depending on the timing of visual loss, which is most prominent in early blindness. This neural recruitment is fundamentally governed by the biological principles of neural circuit expansion and the absence of uncorrelated sensory input. Critically, the degree of these changes is constrained by the availability of neural matter previously allocated to visual experiences, as in the case of late blindness. Moreover, we shed light on the implication of visual deprivation on the neural underpinnings of verbal working memory, revealing longer reverberatory neural activity in 'blind models' as compared to the sighted ones. These findings provide a better understanding of the interplay between visual deprivations, neuroplasticity, language processing and verbal working memory.

Breakdown of category-specific word representations in a brain-constrained neurocomputational model of semantic dementia

The neurobiological nature of semantic knowledge, i.e., the encoding and storage of conceptual information in the human brain, remains a poorly understood and hotly debated subject. Clinical data on semantic deficits and neuroimaging evidence from healthy individuals have suggested multiple cortical regions to be involved in the processing of meaning. These include semantic hubs (most notably, anterior temporal lobe, ATL) that take part in semantic processing in general as well as sensorimotor areas that process specific aspects/categories according to their modality. Biologically inspired neurocomputational models can help elucidate the exact roles of these regions in the functioning of the semantic system and, importantly, in its breakdown in neurological deficits. We used a neuroanatomically constrained computational model of frontotemporal cortices implicated in word acquisition and processing, and adapted it to simulate and explain the effects of semantic dementia (SD) on word processing abilities. SD is a devastating, yet insufficiently understood progressive neurodegenerative disease, characterised by semantic knowledge deterioration that is hypothesised to be specifically related to neural damage in the ATL. The behaviour of our brain-based model is in full accordance with clinical data-namely, word comprehension performance decreases as SD lesions in ATL progress, whereas word repetition abilities remain less affected. Furthermore, our model makes predictions about lesion- and category-specific effects of SD: our simulation results indicate that word processing should be more impaired for object- than for action-related words, and that degradation of white matter should produce more severe consequences than the same proportion of grey matter decay. In sum, the present results provide a neuromechanistic explanatory account of cortical-level language impairments observed during the onset and progress of semantic dementia.

Neurobiological mechanisms for language, symbols and concepts: Clues from brain-constrained deep neural networks

Neural networks are successfully used to imitate and model cognitive processes. However, to provide clues about the neurobiological mechanisms enabling human cognition, these models need to mimic the structure and function of real brains. Brain-constrained networks differ from classic neural networks by implementing brain similarities at different scales, ranging from the micro- and mesoscopic levels of neuronal function, local neuronal links and circuit interaction to large-scale anatomical structure and between-area connectivity. This review shows how brain-constrained neural networks can be applied to study in silico the formation of mechanisms for symbol and concept processing and to work towards neurobiological explanations of specifically human cognitive abilities. These include verbal working memory and learning of large vocabularies of symbols, semantic binding carried by specific areas of cortex, attention focusing and modulation driven by symbol type, and the acquisition of concrete and abstract concepts partly influenced by symbols. Neuronal assembly activity in the networks is analyzed to deliver putative mechanistic correlates of higher cognitive processes and to develop candidate explanations founded in established neurobiological principles.

Influence of language on perception and concept formation in a brain-constrained deep neural network model

A neurobiologically constrained model of semantic learning in the human brain was used to simulate the acquisition of concrete and abstract concepts, either with or without verbal labels. Concept acquisition and semantic learning were simulated using Hebbian learning mechanisms. We measured the network's category learning performance, defined as the extent to which it successfully (i) grouped partly overlapping perceptual instances into a single (abstract or concrete) conceptual representation, while (ii) still distinguishing representations for distinct concepts. Co-presence of linguistic labels with perceptual instances of a given concept generally improved the network's learning of categories, with a significantly larger beneficial effect for abstract than concrete concepts. These results offer a neurobiological explanation for causal effects of language structure on concept formation and on perceptuo-motor processing of instances of these concepts: supplying a verbal label during concept acquisition improves the cortical mechanisms by which experiences with objects and actions along with the learning of words lead to the formation of neuronal ensembles for specific concepts and meanings. Furthermore, the present results make a novel prediction, namely, that such 'Whorfian' effects should be modulated by the concreteness/abstractness of the semantic categories being acquired, with language labels supporting the learning of abstract concepts more than that of concrete ones. This article is part of the theme issue 'Concepts in interaction: social engagement and inner experiences'.

When neuro-robots go wrong: A review

Neuro-robots are a class of autonomous machines that, in their architecture, mimic aspects of the human brain and cognition. As such, they represent unique artifacts created by humans based on human understanding of healthy human brains. European Union's Convention on Roboethics 2025 states that the design of all robots (including neuro-robots) must include provisions for the complete traceability of the robots' actions, analogous to an aircraft's flight data recorder. At the same time, one can anticipate rising instances of neuro-robotic failure, as they operate on imperfect data in real environments, and the underlying AI behind such neuro-robots has yet to achieve explainability. This paper reviews the trajectory of the technology used in neuro-robots and accompanying failures. The failures demand an explanation. While drawing on existing explainable AI research, we argue explainability in AI limits the same in neuro-robots. In order to make robots more explainable, we suggest potential pathways for future research.

Embodied time: Effect of reading expertise on the spatial representation of past and future

How do people grasp the abstract concept of time? It has been argued that abstract concepts, such as future and past, are grounded in sensorimotor experience. When responses to words that refer to the past or the future are either spatially compatible or incompatible with a left-to-right timeline, a space-time congruency effect is observed. In the present study, we investigated whether reading expertise determines the strength of the space-time congruency effect, which would suggest that learning to read and write drives the effect. Using a temporal categorization task, we compared two types of space-time congruency effects, one where spatial incongruency was generated by the location of the stimuli on the screen and one where it was generated by the location of the responses on the keyboard. While the first type of incongruency was visuo-spatial only, the second involved the motor system. Results showed stronger space-time congruency effects for the second type of incongruency (i.e., when the motor system was involved) than for the first type (visuo-spatial). Crucially, reading expertise, as measured by a standardized reading test, predicted the size of the space-time congruency effects. Altogether, these results reinforce the claim that the spatial representation of time is partially mediated by the motor system and partially grounded in spatially-directed movement, such as reading or writing.

Brain correlates of action word memory revealed by fMRI

Understanding language semantically related to actions activates the motor cortex. This activation is sensitive to semantic information such as the body part used to perform the action (e.g. arm-/leg-related action words). Additionally, motor movements of the hands/feet can have a causal effect on memory maintenance of action words, suggesting that the involvement of motor systems extends to working memory. This study examined brain correlates of verbal memory load for action-related words using event-related fMRI. Seventeen participants saw either four identical or four different words from the same category (arm-/leg-related action words) then performed a nonmatching-to-sample task. Results show that verbal memory maintenance in the high-load condition produced greater activation in left premotor and supplementary motor cortex, along with posterior-parietal areas, indicating that verbal memory circuits for action-related words include the cortical action system. Somatotopic memory load effects of arm- and leg-related words were observed, but only at more anterior cortical regions than was found in earlier studies employing passive reading tasks. These findings support a neurocomputational model of distributed action-perception circuits (APCs), according to which language understanding is manifest as full ignition of APCs, whereas working memory is realized as reverberant activity receding to multimodal prefrontal and lateral temporal areas.

Modelling concrete and abstract concepts using brain-constrained deep neural networks

A neurobiologically constrained deep neural network mimicking cortical areas relevant for sensorimotor, linguistic and conceptual processing was used to investigate the putative biological mechanisms underlying conceptual category formation and semantic feature extraction. Networks were trained to learn neural patterns representing specific objects and actions relevant to semantically 'ground' concrete and abstract concepts. Grounding sets consisted of three grounding patterns with neurons representing specific perceptual or action-related features; neurons were either unique to one pattern or shared between patterns of the same set. Concrete categories were modelled as pattern triplets overlapping in their 'shared neurons', thus implementing semantic feature sharing of all instances of a category. In contrast, abstract concepts had partially shared feature neurons common to only pairs of category instances, thus, exhibiting family resemblance, but lacking full feature overlap. Stimulation with concrete and abstract conceptual patterns and biologically realistic unsupervised learning caused formation of strongly connected cell assemblies (CAs) specific to individual grounding patterns, whose neurons were spread out across all areas of the deep network. After learning, the shared neurons of the instances of concrete concepts were more prominent in central areas when compared with peripheral sensorimotor ones, whereas for abstract concepts the converse pattern of results was observed, with central areas exhibiting relatively fewer neurons shared between pairs of category members. We interpret these results in light of the current knowledge about the relative difficulty children show when learning abstract words. Implications for future neurocomputational modelling experiments as well as neurobiological theories of semantic representation are discussed.

Biological constraints on neural network models of cognitive function

Neural network models are potential tools for improving our understanding of complex brain functions. To address this goal, these models need to be neurobiologically realistic. However, although neural networks have advanced dramatically in recent years and even achieve human-like performance on complex perceptual and cognitive tasks, their similarity to aspects of brain anatomy and physiology is imperfect. Here, we discuss different types of neural models, including localist, auto-associative, hetero-associative, deep and whole-brain networks, and identify aspects under which their biological plausibility can be improved. These aspects range from the choice of model neurons and of mechanisms of synaptic plasticity and learning to implementation of inhibition and control, along with neuroanatomical properties including areal structure and local and long-range connectivity. We highlight recent advances in developing biologically grounded cognitive theories and in mechanistically explaining, on the basis of these brain-constrained neural models, hitherto unaddressed issues regarding the nature, localization and ontogenetic and phylogenetic development of higher brain functions. In closing, we point to possible future clinical applications of brain-constrained modelling.

Deep Gated Hebbian Predictive Coding Accounts for Emergence of Complex Neural Response Properties Along the Visual Cortical Hierarchy

Predictive coding provides a computational paradigm for modeling perceptual processing as the construction of representations accounting for causes of sensory inputs. Here, we developed a scalable, deep network architecture for predictive coding that is trained using a gated Hebbian learning rule and mimics the feedforward and feedback connectivity of the cortex. After training on image datasets, the models formed latent representations in higher areas that allowed reconstruction of the original images. We analyzed low- and high-level properties such as orientation selectivity, object selectivity and sparseness of neuronal populations in the model. As reported experimentally, image selectivity increased systematically across ascending areas in the model hierarchy. Depending on the strength of regularization factors, sparseness also increased from lower to higher areas. The results suggest a rationale as to why experimental results on sparseness across the cortical hierarchy have been inconsistent. Finally, representations for different object classes became more distinguishable from lower to higher areas. Thus, deep neural networks trained using a gated Hebbian formulation of predictive coding can reproduce several properties associated with neuronal responses along the visual cortical hierarchy.

The Same Ultra-Rapid Parallel Brain Dynamics Underpin the Production and Perception of Speech

The temporal dynamics by which linguistic information becomes available is one of the key properties to understand how language is organized in the brain. An unresolved debate between different brain language models is whether words, the building blocks of language, are activated in a sequential or parallel manner. In this study, we approached this issue from a novel perspective by directly comparing the time course of word component activation in speech production versus perception. In an overt object naming task and a passive listening task, we analyzed with mixed linear models at the single-trial level the event-related brain potentials elicited by the same lexico-semantic and phonological word knowledge in the two language modalities. Results revealed that both word components manifested simultaneously as early as 75 ms after stimulus onset in production and perception; differences between the language modalities only became apparent after 300 ms of processing. The data provide evidence for ultra-rapid parallel dynamics of language processing and are interpreted within a neural assembly framework where words recruit the same integrated cell assemblies across production and perception. These word assemblies ignite early on in parallel and only later on reverberate in a behavior-specific manner.

A computational model of language functions in flexible goal-directed behaviour

The function of language in high-order goal-directed human cognition is an important topic at the centre of current debates. Experimental evidence shows that inner speech, representing a self-directed form of language, empowers cognitive processes such as working memory, perception, categorization, and executive functions. Here we study the relations between inner speech and processes like feedback processing and cognitive flexibility. To this aim we propose a computational model that controls an artificial agent who uses inner speech to internally manipulate its representations. The agent is able to reproduce human behavioural data collected during the solution of the Wisconsin Card Sorting test, a neuropsychological test measuring cognitive flexibility, both in the basic condition and when a verbal shadowing protocol is used. The components of the model were systematically lesioned to clarify the specific impact of inner speech on the agent’s behaviour. The results indicate that inner speech improves the efficiency of internal representation manipulation. Specifically, it makes the representations linked to specific visual features more disentangled, thus improving the agent’s capacity to engage/disengage attention on stimulus features after positive/negative action outcomes. Overall, the model shows how inner speech could improve goal-directed internal manipulation of representations and enhance behavioural flexibility.

Action sound-shape congruencies explain sound symbolism

Sound symbolism, the surprising semantic relationship between meaningless pseudowords (e.g., 'maluma', 'takete') and abstract (round vs. sharp) shapes, is a hitherto unexplained human-specific knowledge domain. Here we explore whether abstract sound symbolic links can be explained by those between the sounds and shapes of bodily actions. To this end, we asked human subjects to match pseudowords with abstract shapes and, in a different experimental block, the sounds of actions with the shapes of the trajectories of the actions causing these same sounds. Crucially, both conditions were also crossed. Our findings reveal concordant matching in the sound symbolic and action domains, and, importantly, significant correlations between them. We conclude that the sound symbolic knowledge interlinking speech sounds and abstract shapes is explained by audiovisual information immanent to action experience along with acoustic similarities between speech and action sounds. These results demonstrate a fundamental role of action knowledge for abstract sound symbolism, which may have been key to human symbol-manipulation ability.

Tracking the Effects of Top-Down Attention on Word Discrimination Using Frequency-tagged Neuromagnetic Responses

Discrimination of words from nonspeech sounds is essential in communication. Still, how selective attention can influence this early step of speech processing remains elusive. To answer that question, brain activity was recorded with magnetoencephalography in 12 healthy adults while they listened to two sequences of auditory stimuli presented at 2.17 Hz, consisting of successions of one randomized word (tagging frequency = 0.54 Hz) and three acoustically matched nonverbal stimuli. Participants were instructed to focus their attention on the occurrence of a predefined word in the verbal attention condition and on a nonverbal stimulus in the nonverbal attention condition. Steady-state neuromagnetic responses were identified with spectral analysis at sensor and source levels. Significant sensor responses peaked at 0.54 and 2.17 Hz in both conditions. Sources at 0.54 Hz were reconstructed in supratemporal auditory cortex, left superior temporal gyrus (STG), left middle temporal gyrus, and left inferior frontal gyrus. Sources at 2.17 Hz were reconstructed in supratemporal auditory cortex and STG. Crucially, source strength in the left STG at 0.54 Hz was significantly higher in verbal attention than in nonverbal attention condition. This study demonstrates speech-sensitive responses at primary auditory and speech-related neocortical areas. Critically, it highlights that, during word discrimination, top-down attention modulates activity within the left STG. This area therefore appears to play a crucial role in selective verbal attentional processes for this early step of speech processing.

Visual cortex recruitment during language processing in blind individuals is explained by Hebbian learning

In blind people, the visual cortex takes on higher cognitive functions, including language. Why this functional reorganisation mechanistically emerges at the neuronal circuit level is still unclear. Here, we use a biologically constrained network model implementing features of anatomical structure, neurophysiological function and connectivity of fronto-temporal-occipital areas to simulate word-meaning acquisition in visually deprived and undeprived brains. We observed that, only under visual deprivation, distributed word-related neural circuits 'grew into' the deprived visual areas, which therefore adopted a linguistic-semantic role. Three factors are crucial for explaining this deprivation-related growth: changes in the network's activity balance brought about by the absence of uncorrelated sensory input, the connectivity structure of the network, and Hebbian correlation learning. In addition, the blind model revealed long-lasting spiking neural activity compared to the sighted model during word recognition, which is a neural correlate of enhanced verbal working memory. The present neurocomputational model offers a neurobiological account for neural changes following sensory deprivation, thus closing the gap between cellular-level mechanisms, system-level linguistic and semantic function.

A Neurobiologically Constrained Cortex Model of Semantic Grounding With Spiking Neurons and Brain-Like Connectivity

One of the most controversial debates in cognitive neuroscience concerns the cortical locus of semantic knowledge and processing in the human brain. Experimental data revealed the existence of various cortical regions relevant for meaning processing, ranging from semantic hubs generally involved in semantic processing to modality-preferential sensorimotor areas involved in the processing of specific conceptual categories. Why and how the brain uses such complex organization for conceptualization can be investigated using biologically constrained neurocomputational models. Here, we improve pre-existing neurocomputational models of semantics by incorporating spiking neurons and a rich connectivity structure between the model ‘areas’ to mimic important features of the underlying neural substrate. Semantic learning and symbol grounding in action and perception were simulated by associative learning between co-activated neuron populations in frontal, temporal and occipital areas. As a result of Hebbian learning of the correlation structure of symbol, perception and action information, distributed cell assembly circuits emerged across various cortices of the network. These semantic circuits showed category-specific topographical distributions, reaching into motor and visual areas for action- and visually-related words, respectively. All types of semantic circuits included large numbers of neurons in multimodal connector hub areas, which is explained by cortical connectivity structure and the resultant convergence of phonological and semantic information on these zones. Importantly, these semantic hub areas exhibited some category-specificity, which was less pronounced than that observed in primary and secondary modality-preferential cortices. The present neurocomputational model integrates seemingly divergent experimental results about conceptualization and explains both semantic hubs and category-specific areas as an emergent process causally determined by two major factors: neuroanatomical connectivity structure and correlated neuronal activation during language learning.

Task-free auditory EEG paradigm for probing multiple levels of speech processing in the brain

While previous studies on language processing highlighted several ERP components in relation to specific stages of sound and speech processing, no study has yet combined them to obtain a comprehensive picture of language abilities in a single session. Here, we propose a novel task-free paradigm aimed at assessing multiple levels of speech processing by combining various speech and nonspeech sounds in an adaptation of a multifeature passive oddball design. We recorded EEG in healthy adult participants, who were presented with these sounds in the absence of sound-directed attention while being engaged in a primary visual task. This produced a range of responses indexing various levels of sound processing and language comprehension: (a) P1-N1 complex, indexing obligatory auditory processing; (b) P3-like dynamics associated with involuntary attention allocation for unusual sounds; (c) enhanced responses for native speech (as opposed to nonnative phonemes) from ∼50 ms from phoneme onset, indicating phonological processing; (d) amplitude advantage for familiar real words as opposed to meaningless pseudowords, indexing automatic lexical access; (e) topographic distribution differences in the cortical activation of action verbs versus concrete nouns, likely linked with the processing of lexical semantics. These multiple indices of speech-sound processing were acquired in a single attention-free setup that does not require any task or subject cooperation; subject to future research, the present protocol may potentially be developed into a useful tool for assessing the status of auditory and linguistic functions in uncooperative or unresponsive participants, including a range of clinical or developmental populations.

Flexibility in Language Action Interaction: The Influence of Movement Type

Recent neuropsychological studies in neurological patients and healthy subjects suggest a close functional relationship between the brain systems for language and action. Facilitation and inhibition effects of motor system activity on language processing have been demonstrated as well as causal effects in the reverse direction, from language processes on motor excitability or performance. However, as the documented effects between motor and language systems were sometimes facilitatory and sometimes inhibitory, the “sign” of these effects still remains to be explained. In a previous study, we reported a word-category-specific differential impairment of verbal working memory for concordant arm- and leg-related action words brought about by complex sequential movements of the hands and feet. In this article, we seek to determine whether the sign of the functional interaction between language and action systems of the human brain can be changed in a predictable manner by changing movement type. We here report that the sign of the effect of motor movement on action word memory can be reversed from interference to facilitation if, instead of complex movement sequences, simple repetitive movements are performed. Specifically, when engaged in finger tapping, subjects were able to remember relatively more arm-related action words (as compared to control conditions), thus documenting an enhancement of working memory brought about by simple hand movements. In contrast, when performing complex sequences of finger movements, an effector-specific degradation of action word memory was found. By manipulating the sign of the effect in accord with theory-driven predictions, these findings provide support for shared neural bases for motor movement and verbal working memory for action-related words and strengthen the argument that motor systems play a causal and functionally relevant role in language processing semantically related to action.

Motor (but not auditory) attention affects syntactic choice

Understanding the determinants of syntactic choice in sentence production is a salient topic in psycholinguistics. Existing evidence suggests that syntactic choice results from an interplay between linguistic and non-linguistic factors, and a speaker’s attention to the elements of a described event represents one such factor. Whereas multimodal accounts of attention suggest a role for different modalities in this process, existing studies examining attention effects in syntactic choice are primarily based on visual cueing paradigms. Hence, it remains unclear whether attentional effects on syntactic choice are limited to the visual modality or are indeed more general. This issue is addressed by the current study. Native English participants viewed and described line drawings of simple transitive events while their attention was directed to the location of the agent or the patient of the depicted event by means of either an auditory (monaural beep) or a motor (unilateral key press) lateral cue. Our results show an effect of cue location, with participants producing more passive-voice descriptions in the patient-cued conditions. Crucially, this cue location effect emerged in the motor-cue but not (or substantially less so) in the auditory-cue condition, as confirmed by a reliable interaction between cue location (agent vs. patient) and cue type (auditory vs. motor). Our data suggest that attentional effects on the speaker’s syntactic choices are modality-specific and limited to the visual and motor, but not the auditory, domain.

Following instructions in a dual-task paradigm: Evidence for a temporary motor store in working memory

Evidence from dual-task studies suggests that working memory supports the retention and implementation of verbal instructions. One key finding that is not readily accommodated by existing models of working memory is that participants are consistently more accurate at physically performing rather than verbally repeating a sequence of commands. This action advantage has no obvious source within the multi-component model of working memory and has been proposed to be driven by an as yet undetected limited-capacity store dedicated to the temporary maintenance of spatial, motoric, and temporal features of intended movements. To test this hypothesis, we sought to selectively disrupt the action advantage with concurrent motor suppression. In three dual-task experiments, young adults' immediate memory for sequences of spoken instructions was assessed by both action-based and spoken recall. In addition to classic interference tasks known to tax the phonological loop and central executive, motor suppression tasks designed to impair the encoding and retention of motoric representations were included. These required participants to produce repetitive sequences of either fine motor gestures (Experiment 1, N = 16) or more basic ones (Experiments 2, N = 16, and 3, N = 16). The benefit of action-based recall was reduced following the production of basic gestures but remained intact under all other interference conditions. These results suggest that the mnemonic advantage of enacted recall depends on a cognitive system dedicated to the temporary maintenance of motoric representations of planned action sequences.

Spread the word: MMN brain response reveals whole-form access of discontinuous particle verbs

The status of particle verbs such as rise (…) up as either lexically stored or combinatorially assembled is an issue which so far has not been settled decisively. In this study, we use the mismatch negativity (MMN) brain response to observe neurophysiological responses to discontinuous particle verbs. The MMN can be used to distinguish between whole-form storage and combinatorial processes, as it is enhanced to stored words compared to unknown pseudowords, whereas combinatorially legal strings elicit a reduced MMN relative to ungrammatical ones. Earlier work had found larger MMNs to congruent than to incongruent verb-particle combinations when particle and verb appeared as adjacent elements, thus suggesting whole-form storage at least in this case. However, it is still possible that particle verbs discontinuously spread out across a sentence would elicit the combinatorial, grammar-violation response pattern instead. Here, we tested the brain signatures of discontinuous verb-particle combinations, orthogonally varying congruence and semantic transparency. The results show for the first time brain indices of whole-form storage for discontinuous constituents, thus arguing in favour of access to whole-form-stored lexical elements in the processing of particle verbs, irrespective of their semantic opacity. Results are discussed in the context of linguistic debates about the status of particle verbs as words, lexical elements or syntactically generated combinations. The explanation of the pattern of results within a neurobiological language model is highlighted.

Using a Large-scale Neural Model of Cortical Object Processing to Investigate the Neural Substrate for Managing Multiple Items in Short-term Memory

Many cognitive and computational models have been proposed to help understand working memory. In this article, we present a simulation study of cortical processing of visual objects during several working memory tasks using an extended version of a previously constructed large-scale neural model [Tagamets, M. A., & Horwitz, B. Integrating electrophysiological and anatomical experimental data to create a large-scale model that simulates a delayed match-to-sample human brain imaging study. Cerebral Cortex, 8, 310-320, 1998]. The original model consisted of arrays of Wilson-Cowan type of neuronal populations representing primary and secondary visual cortices, inferotemporal (IT) cortex, and pFC. We added a module representing entorhinal cortex, which functions as a gating module. We successfully implemented multiple working memory tasks using the same model and produced neuronal patterns in visual cortex, IT cortex, and pFC that match experimental findings. These working memory tasks can include distractor stimuli or can require that multiple items be retained in mind during a delay period (Sternberg's task). Besides electrophysiology data and behavioral data, we also generated fMRI BOLD time series from our simulation. Our results support the involvement of IT cortex in working memory maintenance and suggest the cortical architecture underlying the neural mechanisms mediating particular working memory tasks. Furthermore, we noticed that, during simulations of memorizing a list of objects, the first and last items in the sequence were recalled best, which may implicate the neural mechanism behind this important psychological effect (i.e., the primacy and recency effect).

Preventing Neurodegenerative Memory Loss in Hopfield Neuronal Networks Using Cerebral Organoids or External Microelectronics

Developing technologies have made significant progress towards linking the brain with brain-machine interfaces (BMIs) which have the potential to aid damaged brains to perform their original motor and cognitive functions. We consider the viability of such devices for mitigating the deleterious effects of memory loss that is induced by neurodegenerative diseases and/or traumatic brain injury (TBI). Our computational study considers the widely used Hopfield network, an autoassociative memory model in which neurons converge to a stable state pattern after receiving an input resembling the given memory. In this study, we connect an auxiliary network of neurons, which models the BMI device, to the original Hopfield network and train it to converge to its own auxiliary memory patterns. Injuries to the original Hopfield memory network, induced through neurodegeneration, for instance, can then be analyzed with the goal of evaluating the ability of the BMI to aid in memory retrieval tasks. Dense connectivity between the auxiliary and Hopfield networks is shown to promote robustness of memory retrieval tasks for both optimal and nonoptimal memory sets. Our computations estimate damage levels and parameter ranges for which full or partial memory recovery is achievable, providing a starting point for novel therapeutic strategies.

Neural reuse of action perception circuits for language, concepts and communication

Neurocognitive and neurolinguistics theories make explicit statements relating specialized cognitive and linguistic processes to specific brain loci. These linking hypotheses are in need of neurobiological justification and explanation. Recent mathematical models of human language mechanisms constrained by fundamental neuroscience principles and established knowledge about comparative neuroanatomy offer explanations for where, when and how language is processed in the human brain. In these models, network structure and connectivity along with action- and perception-induced correlation of neuronal activity co-determine neurocognitive mechanisms. Language learning leads to the formation of action perception circuits (APCs) with specific distributions across cortical areas. Cognitive and linguistic processes such as speech production, comprehension, verbal working memory and prediction are modelled by activity dynamics in these APCs, and combinatorial and communicative-interactive knowledge is organized in the dynamics within, and connections between APCs. The network models and, in particular, the concept of distributionally-specific circuits, can account for some previously not well understood facts about the cortical 'hubs' for semantic processing and the motor system's role in language understanding and speech sound recognition. A review of experimental data evaluates predictions of the APC model and alternative theories, also providing detailed discussion of some seemingly contradictory findings. Throughout, recent disputes about the role of mirror neurons and grounded cognition in language and communication are assessed critically.

Neurocomputational Consequences of Evolutionary Connectivity Changes in Perisylvian Language Cortex

The human brain sets itself apart from that of its primate relatives by specific neuroanatomical features, especially the strong linkage of left perisylvian language areas (frontal and temporal cortex) by way of the arcuate fasciculus (AF). AF connectivity has been shown to correlate with verbal working memory-a specifically human trait providing the foundation for language abilities-but a mechanistic explanation of any related causal link between anatomical structure and cognitive function is still missing. Here, we provide a possible explanation and link, by using neurocomputational simulations in neuroanatomically structured models of the perisylvian language cortex. We compare networks mimicking key features of cortical connectivity in monkeys and humans, specifically the presence of relatively stronger higher-order "jumping links" between nonadjacent perisylvian cortical areas in the latter, and demonstrate that the emergence of working memory for syllables and word forms is a functional consequence of this structural evolutionary change. We also show that a mere increase of learning time is not sufficient, but that this specific structural feature, which entails higher connectivity degree of relevant areas and shorter sensorimotor path length, is crucial. These results offer a better understanding of specifically human anatomical features underlying the language faculty and their evolutionary selection advantage.SIGNIFICANCE STATEMENT Why do humans have superior language abilities compared to primates? Recently, a uniquely human neuroanatomical feature has been demonstrated in the strength of the arcuate fasciculus (AF), a fiber pathway interlinking the left-hemispheric language areas. Although AF anatomy has been related to linguistic skills, an explanation of how this fiber bundle may support language abilities is still missing. We use neuroanatomically structured computational models to investigate the consequences of evolutionary changes in language area connectivity and demonstrate that the human-specific higher connectivity degree and comparatively shorter sensorimotor path length implicated by the AF entail emergence of verbal working memory, a prerequisite for language learning. These results offer a better understanding of specifically human anatomical features for language and their evolutionary selection advantage.

Model-based functional neuroimaging using dynamic neural fields: An integrative cognitive neuroscience approach

A fundamental challenge in cognitive neuroscience is to develop theoretical frameworks that effectively span the gap between brain and behavior, between neuroscience and psychology. Here, we attempt to bridge this divide by formalizing an integrative cognitive neuroscience approach using dynamic field theory (DFT). We begin by providing an overview of how DFT seeks to understand the neural population dynamics that underlie cognitive processes through previous applications and comparisons to other modeling approaches. We then use previously published behavioral and neural data from a response selection Go/Nogo task as a case study for model simulations. Results from this study served as the 'standard' for comparisons with a model-based fMRI approach using dynamic neural fields (DNF). The tutorial explains the rationale and hypotheses involved in the process of creating the DNF architecture and fitting model parameters. Two DNF models, with similar structure and parameter sets, are then compared. Both models effectively simulated reaction times from the task as we varied the number of stimulus-response mappings and the proportion of Go trials. Next, we directly simulated hemodynamic predictions from the neural activation patterns from each model. These predictions were tested using general linear models (GLMs). Results showed that the DNF model that was created by tuning parameters to capture simultaneously trends in neural activation and behavioral data quantitatively outperformed a Standard GLM analysis of the same dataset. Further, by using the GLM results to assign functional roles to particular clusters in the brain, we illustrate how DNF models shed new light on the neural populations' dynamics within particular brain regions. Thus, the present study illustrates how an interactive cognitive neuroscience model can be used in practice to bridge the gap between brain and behavior.

A Spiking Neurocomputational Model of High-Frequency Oscillatory Brain Responses to Words and Pseudowords

Experimental evidence indicates that neurophysiological responses to well-known meaningful sensory items and symbols (such as familiar objects, faces, or words) differ from those to matched but novel and senseless materials (unknown objects, scrambled faces, and pseudowords). Spectral responses in the high beta- and gamma-band have been observed to be generally stronger to familiar stimuli than to unfamiliar ones. These differences have been hypothesized to be caused by the activation of distributed neuronal circuits or cell assemblies, which act as long-term memory traces for learned familiar items only. Here, we simulated word learning using a biologically constrained neurocomputational model of the left-hemispheric cortical areas known to be relevant for language and conceptual processing. The 12-area spiking neural-network architecture implemented replicates physiological and connectivity features of primary, secondary, and higher-association cortices in the frontal, temporal, and occipital lobes of the human brain. We simulated elementary aspects of word learning in it, focussing specifically on semantic grounding in action and perception. As a result of spike-driven Hebbian synaptic plasticity mechanisms, distributed, stimulus-specific cell-assembly (CA) circuits spontaneously emerged in the network. After training, presentation of one of the learned "word" forms to the model correlate of primary auditory cortex induced periodic bursts of activity within the corresponding CA, leading to oscillatory phenomena in the entire network and spontaneous across-area neural synchronization. Crucially, Morlet wavelet analysis of the network's responses recorded during presentation of learned meaningful "word" and novel, senseless "pseudoword" patterns revealed stronger induced spectral power in the gamma-band for the former than the latter, closely mirroring differences found in neurophysiological data. Furthermore, coherence analysis of the simulated responses uncovered dissociated category specific patterns of synchronous oscillations in distant cortical areas, including indirectly connected primary sensorimotor areas. Bridging the gap between cellular-level mechanisms, neuronal-population behavior, and cognitive function, the present model constitutes the first spiking, neurobiologically, and anatomically realistic model able to explain high-frequency oscillatory phenomena indexing language processing on the basis of dynamics and competitive interactions of distributed cell-assembly circuits which emerge in the brain as a result of Hebbian learning and sensorimotor experience.

Is the Sensorimotor Cortex Relevant for Speech Perception and Understanding? An Integrative Review

In the neuroscience of language, phonemes are frequently described as multimodal units whose neuronal representations are distributed across perisylvian cortical regions, including auditory and sensorimotor areas. A different position views phonemes primarily as acoustic entities with posterior temporal localization, which are functionally independent from frontoparietal articulatory programs. To address this current controversy, we here discuss experimental results from functional magnetic resonance imaging (fMRI) as well as transcranial magnetic stimulation (TMS) studies. On first glance, a mixed picture emerges, with earlier research documenting neurofunctional distinctions between phonemes in both temporal and frontoparietal sensorimotor systems, but some recent work seemingly failing to replicate the latter. Detailed analysis of methodological differences between studies reveals that the way experiments are set up explains whether sensorimotor cortex maps phonological information during speech perception or not. In particular, acoustic noise during the experiment and ‘motor noise’ caused by button press tasks work against the frontoparietal manifestation of phonemes. We highlight recent studies using sparse imaging and passive speech perception tasks along with multivariate pattern analysis (MVPA) and especially representational similarity analysis (RSA), which succeeded in separating acoustic-phonological from general-acoustic processes and in mapping specific phonological information on temporal and frontoparietal regions. The question about a causal role of sensorimotor cortex on speech perception and understanding is addressed by reviewing recent TMS studies. We conclude that frontoparietal cortices, including ventral motor and somatosensory areas, reflect phonological information during speech perception and exert a causal influence on language understanding.

Embedding Task-Based Neural Models into a Connectome-Based Model of the Cerebral Cortex

A number of recent efforts have used large-scale, biologically realistic, neural models to help understand the neural basis for the patterns of activity observed in both resting state and task-related functional neural imaging data. An example of the former is The Virtual Brain (TVB) software platform, which allows one to apply large-scale neural modeling in a whole brain framework. TVB provides a set of structural connectomes of the human cerebral cortex, a collection of neural processing units for each connectome node, and various forward models that can convert simulated neural activity into a variety of functional brain imaging signals. In this paper, we demonstrate how to embed a previously or newly constructed task-based large-scale neural model into the TVB platform. We tested our method on a previously constructed large-scale neural model (LSNM) of visual object processing that consisted of interconnected neural populations that represent, primary and secondary visual, inferotemporal, and prefrontal cortex. Some neural elements in the original model were “non-task-specific” (NS) neurons that served as noise generators to “task-specific” neurons that processed shapes during a delayed match-to-sample (DMS) task. We replaced the NS neurons with an anatomical TVB connectome model of the cerebral cortex comprising 998 regions of interest interconnected by white matter fiber tract weights. We embedded our LSNM of visual object processing into corresponding nodes within the TVB connectome. Reciprocal connections between TVB nodes and our task-based modules were included in this framework. We ran visual object processing simulations and showed that the TVB simulator successfully replaced the noise generation originally provided by NS neurons; i.e., the DMS tasks performed with the hybrid LSNM/TVB simulator generated equivalent neural and fMRI activity to that of the original task-based models. Additionally, we found partial agreement between the functional connectivities using the hybrid LSNM/TVB model and the original LSNM. Our framework thus presents a way to embed task-based neural models into the TVB platform, enabling a better comparison between empirical and computational data, which in turn can lead to a better understanding of how interacting neural populations give rise to human cognitive behaviors.

The proactive bilingual brain: Using interlocutor identity to generate predictions for language processing

The present study investigated the proactive nature of the human brain in language perception. Specifically, we examined whether early proficient bilinguals can use interlocutor identity as a cue for language prediction, using an event-related potentials (ERP) paradigm. Participants were first familiarized, through video segments, with six novel interlocutors who were either monolingual or bilingual. Then, the participants completed an audio-visual lexical decision task in which all the interlocutors uttered words and pseudo-words. Critically, the speech onset started about 350 ms after the beginning of the video. ERP waves between the onset of the visual presentation of the interlocutors and the onset of their speech significantly differed for trials where the language was not predictable (bilingual interlocutors) and trials where the language was predictable (monolingual interlocutors), revealing that visual interlocutor identity can in fact function as a cue for language prediction, even before the onset of the auditory-linguistic signal.

Conceptual grounding of language in action and perception: a neurocomputational model of the emergence of category specificity and semantic hubs

Current neurobiological accounts of language and cognition offer diverging views on the questions of 'where' and 'how' semantic information is stored and processed in the human brain. Neuroimaging data showing consistent activation of different multi-modal areas during word and sentence comprehension suggest that all meanings are processed indistinctively, by a set of general semantic centres or 'hubs'. However, words belonging to specific semantic categories selectively activate modality-preferential areas; for example, action-related words spark activity in dorsal motor cortex, whereas object-related ones activate ventral visual areas. The evidence for category-specific and category-general semantic areas begs for a unifying explanation, able to integrate the emergence of both. Here, a neurobiological model offering such an explanation is described. Using a neural architecture replicating anatomical and neurophysiological features of frontal, occipital and temporal cortices, basic aspects of word learning and semantic grounding in action and perception were simulated. As the network underwent training, distributed lexico-semantic circuits spontaneously emerged. These circuits exhibited different cortical distributions that reached into dorsal-motor or ventral-visual areas, reflecting the correlated category-specific sensorimotor patterns that co-occurred during action- or object-related semantic grounding, respectively. Crucially, substantial numbers of neurons of both types of distributed circuits emerged in areas interfacing between modality-preferential regions, i.e. in multimodal connection hubs, which therefore became loci of general semantic binding. By relating neuroanatomical structure and cellular-level learning mechanisms with system-level cognitive function, this model offers a neurobiological account of category-general and category-specific semantic areas based on the different cortical distributions of the underlying semantic circuits.

Sex Differences in Neurophysiological Activation Patterns During Phonological Input Processing: An Influencing Factor for Normative Data

In the context of neurophysiological normative data, it has been established that aging has a significant impact on neurophysiological correlates of auditory phonological input processes, such as phoneme discrimination (PD) and word recognition (WR). Besides age, sex is another demographic factor that influences several language processes. We aimed to disentangle whether sex has a similar effect on PD and WR. Event-related potentials (ERPs) were recorded in 20 men and 24 women. During PD, three phonemic contrasts (place and manner of articulation and voicing) were compared using the attentive P300 and pre-attentive Mismatch Negativity. To investigate WR, real words were contrasted with pseudowords in a pre-attentive oddball task. Women demonstrated a larger sensitivity to spectrotemporal differences, as evidenced by larger P300 responses to the place of articulation (PoA) contrast and larger P300 and MMN responses than men in PoA-based PD. Men did not display such sensitivity. Attention played an important role, considering that women needed more attentional resources to differentiate between PoA and the other phonemic contrasts. During WR, pseudowords evoked larger amplitudes already 100 ms post-stimulus independent of sex. However, women had decreased P200 latencies, but longer N400 latencies in response to pseudowords, whereas men showed increased N400 latencies compared to women in response to real words. The current results demonstrate significant sex-related influences on phonological input processes. Therefore, existing neurophysiological normative data for age should be complemented for the factor sex.

Manual praxis in stone tool manufacture: implications for language evolution

Alternative functions of the left-hemisphere dominant Broca's region have induced hypotheses regarding the evolutionary parallels between manual praxis and language in humans. Many recent studies on Broca's area reveal several assumptions about the cognitive mechanisms that underlie both functions, including: (1) an accurate, finely controlled body schema, (2) increasing syntactical abilities, particularly for goal-oriented actions, and (3) bilaterality and fronto-parietal connectivity. Although these characteristics are supported by experimental paradigms, many researchers have failed to acknowledge a major line of evidence for the evolutionary development of these traits: stone tools. The neuroscience of stone tool manufacture is a viable proxy for understanding evolutionary aspects of manual praxis and language, and may provide key information for evaluating competing hypotheses on the co-evolution of these cognitive domains in our species.

Ultra-rapid access to words in chronic aphasia: the effects of intensive language action therapy (ILAT)

Effects of intensive language action therapy (ILAT) on automatic language processing were assessed using Magnetoencephalography (MEG). Auditory magnetic mismatch negativity (MMNm) responses to words and pseudowords were recorded in twelve patients with chronic aphasia before and immediately after two weeks of ILAT. Following therapy, Patients showed significant clinical improvements of auditory comprehension as measured by the Token Test and in word retrieval and naming as measured by the Boston Naming Test. Neuromagnetic responses dissociated between meaningful words and meaningless word-like stimuli ultra-rapidly, approximately 50 ms after acoustic information first allowed for stimulus identification. Over treatment, there was a significant increase in the left-lateralisation of this early word-elicited activation, observed in perilesional fronto-temporal regions. No comparable change was seen for pseudowords. The results may reflect successful, therapy-induced, language restitution in the left hemisphere.

Neurophysiological evidence for whole form retrieval of complex derived words: a mismatch negativity study

Complex words can be seen as combinations of elementary units, decomposable into stems and affixes according to morphological rules. Alternatively, complex forms may be stored as single lexical entries and accessed as whole forms. This study uses an event-related potential brain response capable of indexing both whole-form retrieval and combinatorial processing, the Mismatch Negativity (MMN), to investigate early brain activity elicited by morphologically complex derived words in German. We presented complex words consisting of stems “sicher” (secure), or “sauber” (clean) combined with abstract nominalizing derivational affixes -heit or -keit, to form either congruent derived words: “Sicherheit” (security) and “Sauberkeit” (cleanliness), or incongruent derived pseudowords: *“Sicherkeit”, and *“Sauberheit”. Using this orthogonal design, it was possible to record brain responses for -heit and -keit in both congruent and incongruent contexts, therefore balancing acoustic variance. Previous research has shown that incongruent combinations of symbols elicit a stronger MMN than congruent combinations, but that single words or constructions stored as whole forms elicit a stronger MMN than pseudowords or non-existent constructions. We found that congruent derived words elicited a stronger MMN than incongruent derived words, beginning about 150 ms after perception of the critical morpheme. This pattern of results is consistent with whole-form storage of morphologically complex derived words as lexical units, or mini-constructions. Using distributed source localization methods, the MMN enhancement for well-formed derivationally complex words appeared to be most prominent in the left inferior anterior-temporal, bilateral superior parietal and bilateral post-central, supra-marginal areas. In addition, neurophysiological results reflected the frequency of derived forms, thus providing further converging evidence for whole form storage and against a combinatorial mechanism.

Thinking in circuits: toward neurobiological explanation in cognitive neuroscience

Cognitive theory has decomposed human mental abilities into cognitive (sub) systems, and cognitive neuroscience succeeded in disclosing a host of relationships between cognitive systems and specific structures of the human brain. However, an explanation of why specific functions are located in specific brain loci had still been missing, along with a neurobiological model that makes concrete the neuronal circuits that carry thoughts and meaning. Brain theory, in particular the Hebb-inspired neurocybernetic proposals by Braitenberg, now offers an avenue toward explaining brain-mind relationships and to spell out cognition in terms of neuron circuits in a neuromechanistic sense. Central to this endeavor is the theoretical construct of an elementary functional neuronal unit above the level of individual neurons and below that of whole brain areas and systems: the distributed neuronal assembly (DNA) or thought circuit (TC). It is shown that DNA/TC theory of cognition offers an integrated explanatory perspective on brain mechanisms of perception, action, language, attention, memory, decision and conceptual thought. We argue that DNAs carry all of these functions and that their inner structure (e.g., core and halo subcomponents), and their functional activation dynamics (e.g., ignition and reverberation processes) answer crucial localist questions, such as why memory and decisions draw on prefrontal areas although memory formation is normally driven by information in the senses and in the motor system. We suggest that the ability of building DNAs/TCs spread out over different cortical areas is the key mechanism for a range of specifically human sensorimotor, linguistic and conceptual capacities and that the cell assembly mechanism of overlap reduction is crucial for differentiating a vocabulary of actions, symbols and concepts.

The bridge of iconicity: from a world of experience to the experience of language

Iconicity, a resemblance between properties of linguistic form (both in spoken and signed languages) and meaning, has traditionally been considered to be a marginal, irrelevant phenomenon for our understanding of language processing, development and evolution. Rather, the arbitrary and symbolic nature of language has long been taken as a design feature of the human linguistic system. In this paper, we propose an alternative framework in which iconicity in face-to-face communication (spoken and signed) is a powerful vehicle for bridging between language and human sensori-motor experience, and, as such, iconicity provides a key to understanding language evolution, development and processing. In language evolution, iconicity might have played a key role in establishing displacement (the ability of language to refer beyond what is immediately present), which is core to what language does; in ontogenesis, iconicity might play a critical role in supporting referentiality (learning to map linguistic labels to objects, events, etc., in the world), which is core to vocabulary development. Finally, in language processing, iconicity could provide a mechanism to account for how language comes to be embodied (grounded in our sensory and motor systems), which is core to meaningful communication.

Brain and behavioral correlates of action semantic deficits in autism

Action-perception circuits containing neurons in the motor system have been proposed as the building blocks of higher cognition; accordingly, motor dysfunction should entail cognitive deficits. Autism spectrum conditions (ASC) are marked by motor impairments but the implications of such motor dysfunction for higher cognition remain unclear. We here used word reading and semantic judgment tasks to investigate action-related motor cognition and its corresponding fMRI brain activation in high-functioning adults with ASC. These participants exhibited hypoactivity of motor cortex in language processing relative to typically developing controls. Crucially, we also found a deficit in semantic processing of action-related words, which, intriguingly, significantly correlated with this underactivation of motor cortex to these items. Furthermore, the word-induced hypoactivity in the motor system also predicted the severity of ASC as expressed by the number of autistic symptoms measured by the Autism-Spectrum Quotient (Baron-Cohen etal., 2001). These significant correlations between word-induced activation of the motor system and a newly discovered semantic deficit in a condition known to be characterized by motor impairments, along with the correlation of such activation with general autistic traits, confirm critical predictions of causal theories linking cognitive and semantic deficits in ASC, in part, to dysfunctional action-perception circuits and resultant reduction of motor system activation.

Neuronal correlates of decisions to speak and act: Spontaneous emergence and dynamic topographies in a computational model of frontal and temporal areas

The neural mechanisms underlying the spontaneous, stimulus-independent emergence of intentions and decisions to act are poorly understood. Using a neurobiologically realistic model of frontal and temporal areas of the brain, we simulated the learning of perception-action circuits for speech and hand-related actions and subsequently observed their spontaneous behaviour. Noise-driven accumulation of reverberant activity in these circuits leads to their spontaneous ignition and partial-to-full activation, which we interpret, respectively, as model correlates of action intention emergence and action decision-and-execution. Importantly, activity emerged first in higher-association prefrontal and temporal cortices, subsequently spreading to secondary and finally primary sensorimotor model-areas, hence reproducing the dynamics of cortical correlates of voluntary action revealed by readiness-potential and verb-generation experiments. This model for the first time explains the cortical origins and topography of endogenous action decisions, and the natural emergence of functional specialisation in the cortex, as mechanistic consequences of neurobiological principles, anatomical structure and sensorimotor experience.