It is proposed that attention to an object requires the simultaneous activity of three brain regions that are interconnected by a triangular circuit. The regions are the cortical site of attentional expression, the thalamic enhancement structure, and the prefrontal area of control. It is also proposed that awareness of an object requires the additional component of attention directed to a representation of the self. The expression of attention to a self-representation may involve activations of cortical sites corresponding to the body (...) landscape and/or verbal-based memories of autobiographical episodes. As in the case of attention to objects, attention to the self is presumed to involve a triangular circuit. The proposed triangular circuit of attention is shown to be generally consistent with current knowledge of brain structures and with data from a broad range of experiments concerned with the functions of neurons in these structures. (shrink)

Human cognition is unique in the way in which it relies on combinatorial (or compositional) structures. Language provides ample evidence for the existence of combinatorial structures, but they can also be found in visual cognition. To understand the neural basis of human cognition, it is therefore essential to understand how combinatorial structures can be instantiated in neural terms. In his recent book on the foundations of language, Jackendoff described four fundamental problems for a neural instantiation of combinatorial structures: the massiveness (...) of the binding problem, the problem of 2, the problem of variables, and the transformation of combinatorial structures from working memory to long-term memory. This paper aims to show that these problems can be solved by means of neural “blackboard” architectures. For this purpose, a neural blackboard architecture for sentence structure is presented. In this architecture, neural structures that encode for words are temporarily bound in a manner that preserves the structure of the sentence. It is shown that the architecture solves the four problems presented by Jackendoff. The ability of the architecture to instantiate sentence structures is illustrated with examples of sentence complexity observed in human language performance. Similarities exist between the architecture for sentence structure and blackboard architectures for combinatorial structures in visual cognition, derived from the structure of the visual cortex. These architectures are briefly discussed, together with an example of a combinatorial structure in which the blackboard architectures for language and vision are combined. In this way, the architecture for language is grounded in perception. Perspectives and potential developments of the architectures are discussed. Key Words: binding; blackboard architectures; combinatorial structure; compositionality; language; dynamic system; neurocognition; sentence complexity; sentence structure; working memory; variables; vision. (shrink)

I argue that dynamicism does not provide a convincing alternative to currently available cognitive theories. First, I show that the attractor dynamics of dynamicist models are inadequate for accounting for high-level cognition. Second, I argue that dynamicist arguments for the rejection of computation and representation are unsound in light of recent empirical findings. This new evidence provides a basis for questioning the importance of continuity to cognitive function, challenging a central commitment of dynamicism. Coupled with a defense of current connectionist (...) theory, these two critiques lead to the conclusion that dynamicists have failed to achieve their goal of providing a new paradigm for understanding cognition. (shrink)

There is some complementarity of models for the origin of the electroencephalogram (EEG) and neural network models for information storage in brainlike systems. From the EEG models of Freeman, of Nunez, and of the authors' group we argue that the wavelike processes revealed in the EEG exhibit linear and near-equilibrium dynamics at macroscopic scale, despite extremely nonlinear – probably chaotic – dynamics at microscopic scale. Simulations of cortical neuronal interactions at global and microscopic scales are then presented. The simulations depend (...) on anatomical and physiological estimates of synaptic densities, coupling symmetries, synaptic gain, dendritic time constants, and axonal delays. It is shown that the frequency content, wave velocities, frequency/wavenumber spectra and response to cortical activation of the electrocorticogram (ECoG) can be reproduced by a “lumped” simulation treating small cortical areas as single-function units. The corresponding cellular neural network simulation has properties that include those of attractor neural networks proposed by Amit and by Parisi. Within the simulations at both scales, sharp transitions occur between low and high cell firing rates. These transitions may form a basis for neural interactions across scale. To maintain overall cortical dynamics in the normal low firing-rate range, interactions between the cortex and the subcortical systems are required to prevent runaway global excitation. Thus, the interaction of cortex and subcortex via corticostriatal and related pathways may partly regulate global dynamics by a principle analogous to adiabatic control of artificial neural networks. (shrink)

The "relational mind" approach to the inner content of consciousness is developed in terms of various control structures and processing strategies and their possible neurobiological identifications in brain sites. This leads naturally to a division of consciousness into a passive and an active part. A global control structure for the "single strand" aspect of consciousness is proposed as the thalamo-nucleus reticularis thalami-cortex coupled system, which is related to experimental data on the electrical stimulation of awareness. Local control, in terms of (...) excitatory transfer from pre-processing sites to posterior working memory regions, is supported by data on subliminal perception timing and disambiguation of poorly defined percepts. The inner content of consciousness is understood as arising from the resulting relational features between inputs and stored pre-processing and episodic memories. Strong analogies are drawn between emergent properties of the model and suggested properties of "raw feels", supporting the thesis that working memories are the initial sites for the emergence of phenomenal awareness, and the frontal lobes for its further adumbration in terms of higher cognitive processing, including the creation of self. (shrink)

This chapter provides an overview of the basic research strategies and analytic techniques deployed in computational cognitive neuroscience. On the one hand, “top-down” strategies are used to infer, from formal characterizations of behavior and cognition, the computational properties of underlying neural mechanisms. On the other hand, “bottom-up” research strategies are used to identify neural mechanisms and to reconstruct their computational capacities. Both of these strategies rely on experimental techniques familiar from other branches of neuroscience, including functional magnetic resonance imaging, single-cell (...) recording, and electroencephalography. What sets computational cognitive neuroscience apart, however, is the explanatory role of analytic techniques from disciplines as varied as computer science, statistics, machine learning, and mathematical physics. These techniques serve to describe neural mechanisms computationally, but also to drive the process of scientific discovery by influencing which kinds of mechanisms are most likely to be identified. For this reason, understanding the nature and unique appeal of computational cognitive neuroscience requires not just an understanding of the basic research strategies that are involved, but also of the formal methods and tools that are being deployed, including those of probability theory, dynamical systems theory, and graph theory. (shrink)

The Gestalt principle of isomorphism reveals the primacy of subjective experience as a valid source of evidence for the information encoded neurophysiologically. This theory invalidates the abstractionist view that the neurophysiological representation can be of lower dimensionality than the percept to which it gives rise.

A standing challenge for the science of mind is to account for the datum that every mind faces in the most immediate – that is, unmediated – fashion: its phenomenal experience. The complementary tasks of explaining what it means for a system to give rise to experience and what constitutes the content of experience (qualia) in computational terms are particularly challenging, given the multiple realizability of computation. In this paper, we identify a set of conditions that a computational theory must (...) satisfy for it to constitute not just a sufficient but a necessary, and therefore naturalistic and intrinsic, explanation of qualia. We show that a common assumption behind many neurocomputational theories of the mind, according to which mind states can be formalized solely in terms of instantaneous vectors of activities of representational units such as neurons, does not meet the requisite conditions, in part because it relies on inactive units to shape presently experienced qualia and implies a homogeneous representation space, which is devoid of intrinsic structure. We then sketch a naturalistic computational theory of qualia, which posits that experience is realized by dynamical activity-space trajectories (rather than points) and that its richness is measured by the representational capacity of the trajectory space in which it unfolds. (shrink)

In this paper I put forward a reconstruction of the evolution of certain explanatory hypotheses on the neural basis of association and learning that are the premises of connectionism in the cybernetic age and of present-day connectionism. The main point of my reconstruction is based on two little-known case studies. The first is the project, published in 1913, of a hydraulic machine through which its author believed it was possible to simulate certain essential elements of the plasticity of nervous connections. (...) The author, S. Bent Russell, was an engineer deeply influenced by the neurological hypotheses on nervous conduction of Herbert Spencer, Max Meyer and Edward L. Thorndike. The second is the project, published in 1929, of an electromechanical machine in which the author, the psychologist J.M. Stephens, believed it was possible to embody Thorndike's law of effect. Thus both Bent Russell and Stephens referred to the principles of learning that Thorndike defined as connectionist . Their attempt was that of simulating by machines at least certain simple aspects of inhibition, association and habit formation that are typical of living organisms. I propose to situate their projects within the frame of thediscovery of a simulative (modelling) methodology which I believe might be considered an important topic of the Culture of the Artificial . Certain more recent steps toward such a methodology made by both connectionism of the 1950s and present-day connectionism are briefly pointed out in the paper. (shrink)

Some concepts relating to learned, structured functioning of local modules in neocortex are clarified in order to ensure that the integration from the small scale to the global attempted by Wright & Liley does not miss the target.

Wright & Liley provide an advance in addressing the interaction of multiple scales of processing in the brain. It should address in more detail the biological evidence that underlies the models it proposes to replace.

Although this is an impressive piece of modeling work, I worry that the two models that Wright & Liley have created do not yet provide us with useful empirical information regarding brain processing.

Does the Wright & Liley model predict: (1) that subdural and hippocampal EEGs coherence tend to rise and fall in parallel for many frequencies, (2) that it is locally high or low within 10mm and falls steeply on average or, (3) that it is in constant flux, mostly rising and falling within 5–15 sec?

Two hypotheses concerning nonlinear elements in complex systems are contrasted: that neurons, intrinsically unstable, are stabilized through embedding in networks and populations; and, conversely, that cortical neurons are intrinsically stable, but are destabilized through embedding in cortical populations and corticostriatal feedback systems. Tests are made by piecewise linearization of nonlinear dynamics at nonequilibriumoperating points, followed by linear stability analysis.

The work in progress reported by Wright & Liley shows great promise, primarily because of their experimental and simulation paradigms. However, their tentative conclusion that macroscopic neocortex may be considered (approximately) a linear near-equilibrium system is premature and does not correspond to tentative conclusions drawn from other studies of neocortex.

Grodzinsky interprets linguistic differences between agrammatic comprehension and production symptoms as supporting the hypothesis that the mechanisms underlying grammatical encoding (sentence formulation) and grammatical decoding (syntactic parsing) are at least partially distinct. This inference is shown to be premature. A range of experimentally established similarities between the encoding and decoding processes is highlighted, testifying to the viability of the hypothesis that receptive and productive syntactic tasks are performed by the same syntactic processor.

To model the organization of levels' of cortical dynamics, at least some general scheme for hierarchy, functional diversity, and proper intrinsic control must be provided. Rhythmic control forces the system to iterate its state by short trajectories, which makes it much more stable and predictable without discarding the desirable ability of chaotic systems to make rapid phase transitions. Rhythmic control provides a fundamentally different systems dynamics, one not provided by models that allow the emergence of continuous trajectories in the systems (...) state space. (shrink)

We focus on one aspect of Wright & Liley's target article: the linearity of the EEG. According to the authors, some nonlinear models of the cortex can be reduced (approximated) to the linear case at the millimetric scale. We argue here that the statement about the linear character of EEG is too strong and that EEG exhibits nonlinear features which cannot be ignored.

We present simulation results of an olfactory cortex model complementing the results presented in Wright & Liley's target article. We show how the cortical dynamics as expressed in EEG can be regulated by neuromodulation and discuss how the system can attain global stability without cortical-subcortical interaction, as presumed necessary by Wright & Liley. Network structure is shown to be crucial.

Simulation of brain dynamics requires the use of accurate empirical data. This commentary points out major errors in some of the empirical data used in Wright & Laley's simulation. The simulation is quantitatively very different from the real cortex, and may also have important qualitative differences.

Event-related potentials (ERPs) – neglected almost entirely by Wright & Liley – allow objective investigation of information processing in the brain. The application of chaos theory to such an analysis broadens this possibility. Through the use of the point correlation dimension (PD2) accurate dimensional analysis of different Event-Related Potential components such as the P3 wave is possible.

Multiscale dynamics, linear approximations, global boundary conditions, experimental verification, and global influences on local cell assemblies are considered in the context of Wright & Liley's work. W&L provide a nice introduction to these issues and a reasonable simulation of intermediate scale dynamics, but the model does not adequately simulate combined local and global processes.

The unification of microscopic and macroscopic models of brain behaviour is of paramount importance and Wright & Liley's target article provides some important groundwork. In this commentary, I propose that a useful approach for the future is to incorporate a developmental perspective into such models. This may be an important constraint, providing a key to understanding the nature of macroscopic measures of brain function such as functional measures like ERP.

For some years there has been a controversy about whether brain state variables such as EEG or neuronal spike trains exhibit chaotic behaviour. Wright & Liley claim that the local dynamics measured by spike trains or local field potentials exhibit chaotic behaviour, but global measures like EEG should be governed by linear dynamics. We propose a different scheme. Based on simulation studies and various experiments, we suggest that the pointwise dimension of EEG time series may provide some valuable information about (...) underlying neuronal generators. (shrink)

Wright & Liley contrast their theory that the global dynamics of the EEG are linear with that of Freeman, who hypothesizes an EEG governed by (nonlinear) deterministic-chaotic dynamics. A “call for further discussion” on the part of the authors is made as to how either theory fits with experimental findings indicating that EEG dynamics are non-linear but stochastic.

Some dichotomies related to modeling electrocortical activities are analyzed. Attractor neural networks versus biologically motivated models, near-equilibrium versus nonequilibrium processes, linear and nonlinear dynamics, stochastic and chaotic patterns, local and global scale simulation of cortical activities are discussed.

I would like to emphasize the significance of chaotic dynamics at both local and macroscopic levels in the cortex. The basic notions dealt with in this commentary will be noise-induced order, chaotic “itinerancy” and dissipative structure. Wright & Laley's theory would be partially misleading, since emergent nonlinearity rather than the linearity at even a macroscopic level can actually subserve cortical functions.

We outline fresh findings that show that our macroscopic electrocorticographic (ECoG) simulations can account for synchronous multiunit pulse oscillations at separate, simultaneously activated cortical sites and the associated gamma-band ECoG activity. We clarify our views on the approximations of dynamic class applicable to neural events at macroscopic and microscopic scales, and the analogies drawn to classes of ANN behaviour. We accept the need to introduce memory processes and detailed anatomical and physiological information into any future developments of our simulations. On (...) the issue of intrinsic cortical stability and the role of extrinsic fibre systems in maintaining stability, we argue that this position is not in extreme contradiction to those of our commentators, and that the mechanisms implicit in our simulations' properties imply rich computational possibilities. We discuss some of the reasons for and against the existence of significant global resonances in the brain and explain why such behaviour appears absent in our simulations. Last, we discuss other phenomena, such as rhythmic driving of the cortex, which have not yet been introduced into our models, and indicate lines for future development of the simulations. (shrink)

The network model of EEG formation has revealed a unified mechanism for disparate EEG phenomena: for various reactions as well as for ontogenetic and phylogenetic differences. EEG rhythmicity was shown to be an external manifestation of the functioning of the intracortical stabilizing system which provides normal informational operations in the cerebral cortex.