Max Garagnani

Goldsmiths, University of London · Department of Computing

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...

The ability to coactivate (or “superpose”) multiple conceptual representations is a fundamental function that we constantly rely upon; this is crucial in complex cognitive tasks requiring multi-item working memory, such as mental arithmetic, abstract reasoning, and language comprehension. As such, an artificial system aspiring to implement any of t...

We are pleased to announce that the presentations and posters of the Annual Computational Neuroscience Meeting (CNS*2023) have become available. Discover the detailed program on the official website https://cns2023.sched.com ... Join us at Annual Computational Neuroscience Meeting. See also https://link.springer.com/article/10.1007/s10827-024-008...

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 meani...

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 th...

The nature of semantic knowledge – conceptual information stored in the brain – is highly debated in the field of cognitive science. Experimental and clinical data specify various cortical regions involved in the processing of meaning. Those include semantic hubs that take part in semantic processing in general as well as sensorimotor areas that pr...

Foreword from the editors. We hosted four keynote speakers: Wolf Singer, Bill Bialek, Danielle Bassett, and Sonja Gruen. They enlightened us about computations in the cerebral cortex, the reduction of high-dimensional data, the emerging field of computational psychiatry, and the significance of spike patterns in motor cortex. From the submissions,...

Embodied theories of grounded semantics postulate that, when word meaning is first acquired, a link is established between symbol (word form) and corresponding semantic information present in modality-specific—including primary—sensorimotor cortices of the brain. Direct experimental evidence documenting the emergence of such a link (i.e., showing t...

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 fr...

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 senso...

Neuroimaging and patient studies show that different areas of cortex respectively specialize for general and selective, or category-specific, semantic processing. Why are there both semantic hubs and category-specificity, and how come that they emerge in different cortical regions? Can the activation time-course of these areas be predicted and expl...

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 t...

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...

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 multi-modal areas during word and sentence comprehension suggest that all meanings are processed indistinctively, b...

In this work we describe how an existing neural model for learning Cell Assemblies (CAs) across multiple neuroanatomical brain areas has been integrated with a humanoid robot simulation to explore the learning of associations of visual and motor modalities. The results show that robust CAs are learned to enable pattern completion to select a correc...

Memory cells, the ultimate neurobiological substrates of working memory, remain active for several seconds and are most commonly found in prefrontal cortex and higher multisensory areas. However, if correlated activity in “embodied” sensorimotor systems underlies the formation of memory traces, why should memory cells emerge in areas distant from t...

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...

Humans are unique in developing large lexicons; to achieve this, they are able to learn new words rapidly. However, the neural bases of this rapid learning, which may be an expression of a more general mechanism rooted in plasticity at cellular and synaptic levels, are not yet understood. Here, we highlight a selection of recent EEG and fMRI studie...

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 s...

Lateral inhibition and neuronal adaptation, the main candidate mechanisms previously proposed to underlie the generation of the mismatch negativity (MMN) response as elicited in passive oddball experiments, fail to explain why the MMN is larger for familiar stimuli (e.g., words) than for unfamiliar ones (e.g., pseudowords). We aimed to identify a s...

Most animals detect sudden changes in trains of repeated stimuli but only some can learn a wide range of sensory patterns and recognise them later, a skill crucial for the evolutionary success of higher mammals. Here we use a neural model mimicking the cortical anatomy of sensory and motor areas and their connections to explain brain activity index...

Background: Sensory dysfunctions are frequently reported in children with autism spectrum disorders (ASD). It has been suggested these symptoms are associated with their difficulties processing language as well as with social isolation observed in people with autism. It is speculated that impaired feature extraction in early sensory processing coul...

This work concerns the investigation of the neuronal mechanisms at the basis of language acquisition and processing, and the complex interactions of language and attention processes in the human brain. In particular, this research was motivated by two sets of existing neurophysiological data which cannot be reconciled on the basis of current psycho...

Current cognitive theories postulate either localist representations of knowledge or fully overlapping, distributed ones. We use a connectionist model that closely replicates known anatomical properties of the cerebral cortex and neurophysiological principles to show that Hebbian learning in a multi-layer neural network leads to memory traces (cell...

Recent simulations obtained with a neurobiologically realistic neural network model of the left perisylvian language cortex suggest that words are represented in the human brain as strongly connected circuits that are both distributed and functionally discrete (Garagnani et al., 2008). Such a model replicates, explains and reconciles existing diver...

Recent EEG and MEG studies have revealed that brain responses to the same speech sounds differ if the stimuli are presented in different task contexts: when subjects are not paying attention to the auditory input, their mismatch negativity (MMN) response is greater for words than for matched meaningless pseudowords. However, greater late N400 respo...

Recent results obtained with a neural-network model of the language cortex suggest that the memory circuits developing for words are both distributed and functionally discrete. This model makes testable predictions about brain responses to words and pseudowords under variable availability of attentional resources. In particular, due to their strong...

Meaningful familiar stimuli and senseless unknown materials lead to different patterns of brain activation. A late major neurophysiological response indexing 'sense' is the negative component of event-related potential peaking at around 400 ms (N400), an event-related potential that emerges in attention-demanding tasks and is larger for senseless m...

We modelled language-learning processes in a brain-inspired model of the language cortex. The network consisted of neuron-like elements (graded-response units) and mimicked the neuroanatomical areas in the perisylvian language cortex and the intrinsic and mutual connections within and between them. Speaking words creates correlated activity in moto...

This paper demonstrates how associative neural networks as standard models for Hebbian cell assemblies can be extended to implement language processes in large-scale brain simulations. To this end the classical auto- and hetero-associative paradigms of attractor nets and synfire chains (SFCs) are combined and complemented by conditioned association...

Sentential and diagrammatic representations are two different formalisms for describing domains and problems. Sentential descriptions are usually more expres- sive than diagrammatic ones, but tend to present a more complex and less intuitive notation. All modern planning domain description languages are sentential. The com- plexity of sentential fo...

This chapter describes a model and an underlying theoretical framework for hybrid planning. Modern planning domain description languages are based on sentential representations. Sentential formalisms produce problem encodings that often lead the system to carry out large amounts of superfluous operations, causing a loss in performance. This chapter...

Sentential and analogical representations constitute two complementary formalisms for describing problems and domains. Experimental evidence indicates that different domain types can have their most efficient encoding in different representations. While real-world problems typically involve a combination of different types of domains, all modern pl...

This chapter describes a model and an underlying theoretical framework for hybrid planning. Modern planning domain-description formalisms are based on purely sentential languages. Sentential representations produce problem encodings that often require the system to carry out an unnecessary amount of trivial deductions, preventing it from concentrat...

This paper proposes a non-propositional representation framework for planning in physical domains. Physical planning problems involve identi-fying a correct sequence (plan) of object manipulations, transformations and spatial rearrangements to achieve an assigned goal. The problem of the ramification of action effects causes most current (propositi...

We describe a model-based planning representation, aimed at capturing more efficiently the basic topological and structural properties of a domain. We specify the syntax of a domain-modelling language based on the proposed representation. We report the experimental results obtained with a prototype system (called PMP, Pattern-Matching Planner) able...

This report illustrates how new methods and techniques from the area of knowledge representation and reasoning can be adopted and exploited in planning to produce new, more efficient domain-description languages. Planning domain description formalisms should be expressive and customisable, and yet be able to produce domain encodings that allow the...

We present an automatic procedure for pre-processing planning problems containing language axioms, a specific type of domain axioms. The axioms considered are assumed to be in the form p ^ P2 ^ ...^ P, --> c. The pre-processing approach described consists of encoding the language axioms directly inside the given operators, contrary to other (not al...

We present an automatic procedure for pre-processing planning problems containing language axioms, a specific type of domain axioms. The axioms considered are assumed to be in the form p1 ∧ p2 ∧ …∧ pn → c. The pre-processing approach described consists of encoding the language axioms directly inside the given operators, contrary to other (not alway...

A connectionist model for emergent planning behavior is proposed. The model demonstrates that a simple planning schema, acting in concert with two general purpose cognitive functionalities, namely, episodic memory and perception, can solve a restricted class of planning problems by backchaining from the goal to the current state. In spite of its si...

This paper describes a polynomial algorithm for preprocessing planning problems which contain domain axioms (DAs) in the form p1 ∧ p2 ∧. . . ∧ pn → c. The algorithm presented is an improved version of the (incorrect) transformation for DAs described by Gazen and Knoblock in [6]. The first result presented consists of a counter-example showing that...

This short paper proposes an extension of the Graphplan planning system. The extension allows Graphplan to solve planning problems containing Domain Axioms (DAs) in the form p1 /\ p2 /\…/\ pn --> c.

Using sophisticated and expressive belief languages for generating complex discourse plans requires a clear distinction between the belief system and the planning machinery which makes use of it. The main results presented in this paper consist of: 1) the formalisation of an expressive 'speaker-hearer' belief system which allows the representation...

The presence of domain axioms (DAs) in a planning problem can be easily originated by the adoption of expressive domain definition languages. The paper proposes an improved version of the pre-processing algorithm for the solution of planning problems with DAs described by Gazen and Knoblock in [3]. The new method avoids the inefficiencies of the pr...

Given a planning domain, the choice of the language for the state definition, goals and operator specification is, often, not simple: if a language is sophisticated enough to allow a accurate ad natural description of these elements, it is likely to contain expressions which are related by logical formulae, determined by the specific characteristic...

Given a planning domain, if the language adopted for the state representation is sophisticated enough to allow an accurate, simple and natural description of a problem, it is likely to contain `redundancies', i.e. expressions whose validity (or truth) can be deduced from that of other expressions currently present in the state. The relations exp...

Communication between rational agents can be seen as an exchange of internal concepts expressed using a shared language. Most of the communication acts between humans are performed through Natural Language (NL), and the meaning of the expressions of the language adopted is assumed to be known and properly understood by all of the participants. This...

In a multiagent system, conflicts may arise because of two basic reasons: 1) different agents have contrasting goals; 2) different agents have inconsistent knowledge. This situation can originate when agents are autonomous and strongly motivated by their own interests, and when heterogeneous agents, with different skills, 'histories' and beliefs, c...

Agents in a multi-agent environment must often cooperate to achieve their objectives. In this paper an agent, B, cooperates with another agent, A, if B adopts a goal that furthers A's objectives in the environment. If agents are independent and motivated by their own interests, cooperation cannot be relied upon and it may be necessary for A to pers...

Agents in a multi-agent environment must often cooperate to achieve their objectives. In this paper an agent, B, cooperates with another agent, A, if B adopts a goal that furthers A's objectives in the environment. If agents are independent and motivated by their own interests, cooperation cannot be relied upon and it may be necessary for A to pers...