Scientific Understanding of Consciousness
Consciousness as an Emergent Property of Thalamocortical Activity

Expert’s Theories Sometimes Impose Incongruous Interfaces

Interfaces between experts’ ideas do not always blend smoothly.

Very frequently, similar ideas are expressed in different terms.

Amalgamation and reconciliation of ideas of experts.

Important to be able to combine ideas. People who are especially inventive have a gift for connecting elements that at first glance may seem to have nothing in common. (Kraft; Unleashing Creativity, 17)

Creativity depends primarily on the ability to integrate pieces of disparate data in novel ways. (Kraft; Unleashing Creativity, 18)

Further research work would be very productive and provide much insight.

Hierarchy

Associative Networks

Memory and its retrieval are based on association. (Fuster; Memory in Cerebral Cortex, 213)

In an associative network in higher cortex, an entire class in memory, rapidly can establish new connections with incoming stimuli. (Fuster; Memory in Cerebral Cortex, 95)

The actual autoassociation could be implemented by autoassociatively modifiable synapses between connected pyramidal cells within an area, or by the forward and backward connections between adjacent cortical areas in a hierarchy. (Rolls; Memory, Attention, and Decision-Making, 375)

Each region of the brain contributes to the recognition of a 'chair,' which explains why recognition can be triggered by a number of different sensory elements. (Ratey; User's Guide to Brain, 143)

Convergence zones enable us to automatically conceive of objects, ideas, or interactions as a whole, if the pieces have been put together enough times. (Ratey; User's Guide to Brain, 187)

Associative properties arise materially from the fact that each different member of the degenerate set of circuits used at different times has different alternative network connections. (Edelman; Universe of Consciousness, 98)

Research with brain surgery patients has shown that stimulation to certain brain areas results in complicated feelings such as that experience when having made a social faux pas at a cocktail party. (Ratey; User's Guide to Brain, 226)

Conscious processing requires a particular dynamical state of cortical networks that is characterized by a brief episode of very precise phase locking of high frequency oscillatory activity. (Singer; Neuronal Synchronization, 49)

The autoassociative network could provide the very precise phase locking. High frequency oscillatory activity is a product of the autoassociative networks.

Autoassociation network for recognition memory. (diagram) (Squire; Fundamental Neuroscience, 1295)

When associations flying through our brains self-organize to form a new idea, the result is creativity. If the associations either fail to self-organize, or if they self-organize to create an erroneous idea, the result is psychosis. (Andreasen; Creating Brain, 102)

Response Time of Neuron Processing

Neurons in each cortical stage respond for 20-30 ms when an object can just be seen. (Rolls; Memory, Attention, and Decision-Making, 31)

In the 30 ms necessary to apprehend the gist of a scene, top-down attention cannot play much of a role (because gist is a property associated with the entire image). (Tsuchiya & Koch; Consciousness and Attention, 68)

Experimental observations have concluded that sensory information is processed in discrete time segments as low as 12 ms. (Llinás & Paré; Brain Modulated by Senses, 12)

At more local levels, for instance, in posterior temporal cortices, the scale would be smaller, in the order of tens of milliseconds. The return projections necessary for the reconstruction are aimed toward layers I and V of the cortex. (Domasio; Convergence Zone, 71)

Two thirds of neurons often fire in bursts of 2-4 spikes within 2-6 msec and show a small peak in the 25-50 Hz band of the power spectrum that is related to the propensity of spikes to fire in bursts. The statistical properties can be fitted by Poisson-distributed bursts with a burst-dependent refractory period. (Koch and Crick; Neuronal Basis, 104)

The remaining third of their cells have an autocorrelation function and an interspike interval distribution compatible with the notion that spikes are Poisson distributed with a refractory period. (Koch and Crick; Neuronal Basis, 104)

The variability in interspike intervals observed for many neurons in various locations in the nervous system is consistent with the presence of a random Poisson process. (Motter; Neurophysiology, 52)

Neurons are much slower than a computer. In a half second, information entering the brain can traverse a chain of approximately 100 neurons. (Gazzaniga; Human, 367)

Response Time for Conscious Cognition

In the large scale reconstruction from higher-order cortices such as those in anterior temporal lobe, the time scale of the synchronization would be in the order of several hundred msec, and even beyond 1000 msec, the scale required for meaningful, conscious cognition. (Domasio; Convergence Zone, 71)

Prefrontal cortex-dependent top-down enhancement of visual association cortex activity occurring in the first few hundred milliseconds of the visual processing. (Gazzaley; Top-down WM, 205)

Convergence

In the hierarchy of cortical stages, convergence and competition are key aspects of the processing. (Rolls; Memory, Attention, and Decision-Making, 31)

Convergence divergence zones (CDZs) record the coincidence of activity in neurons hailing from different brain sites, neurons that had been made active by, for example, the mapping of a certain object. (Damasio; Self Comes to Mind, 141)

There are in the order of thousands of convergence zones, which are all microscopic neuron ensembles, located within the macroscopic convergence regions that have been cytoarchitectonically defined and that number about one hundred. (Domasio; Convergence Zone, 71)

Fuster's perception-action cycle

Perception-action cycle is the circular processing of information between posterior and frontal cortices in the integration of sensory -- motor behavior, as well is in higher cognitive activities such as language. (Fuster; Cortex and Mind, 218)

Brain uses a carrot-and-stick system to ensure that we pursue and achieve the things we need to survive. A stimulus from the outside or from the body is registered by the limbic system, which creates an urge, which registers consciously as desire. The cortex then instructs the body to act in whatever way is necessary to achieve its desire. The activity sends messages back to the limbic system, which release opioid-like neurotransmitters, which raise circulating dopamine levels and create a feeling of satisfaction. (Carter; Mapping the Mind, 63)

There is a ceaseless production of new activity states, in early sensory cortices and in motor cortices, across time. It is these successive neural states, one after the other, that can be said to constitute "regresses" for the previous state. Neural assemblies are activated (refreshed) in a pulse-like manner by reentrant signals circulating in the network hierarchy. It is the perpetually recursive property of corticocortical systems that permits this special form of regress. (Domasio; Convergence Zone, 70)

Oscillations

Brain oscillators should be considered as a system of oscillators with an intricate relationship between the various rhythmic components. (Buzsáki - Rhythms of the Brain, 119)

The amplitude of the EEG power spectrum increases as the frequency decreases. This inverse relationship is expressed as the "one over f" power spectrum (also called "pink" noise). (Buzsáki - Rhythms of the Brain, 119)

Long-range interneurons provide the necessary conduit for synchronizing distantly operating oscillators and allow for coherent timing of large numbers of neurons that are not directly connected with each other. (Buzsáki; Rhythms of the Brain, 71)

Synchronization

Intralaminar nuclei of the thalamus send diffuse projections to most areas of the cerebral cortex and help to synchronize its overall level of activity. (Edelman; Universe of Consciousness, 107)

Intralaminar nuclei of the thalamus, which send diffuse connections to most areas of the cortex, helps to synchronize thalamocortical responses and regulate the overall levels of activity in these multiple reentrant systems. (Edelman; Wider than the Sky, 55)

Perhaps music creates a resonance in the brain between neurons firing in synchrony with a sound wave and a natural oscillation in the emotion circuits? (Pinker; How the Mind Works, 538)

We can conclude that cell assemblies are synchronized within the time window of gamma oscillation, because this allows an assembly to exert the largest possible impact on their downstream targets. (Buzsáki; Rhythms of the Brain, 163)

Synchronization by gamma oscillations results in not only perceptual binding but, inevitably, modification of connections among the neurons involved. (Buzsáki; Rhythms of the Brain, 247)

One mechanism for dynamic binding is the precise synchronization of neuronal responses that occurs when neuronal populations engage in well synchronized oscillatory activity in the beta and gamma frequency range. (Singer; Neuronal Synchronization, 45)

Synchronized oscillations are strongly reduced or missing when the brain is in states that are incompatible with conscious processing. (Singer; Neuronal Synchronization, 46)

Gestalt features of the stimulus strongly modulate synchronization. (Revonsuo; Inner Presence, 273)

Wolfgang Singer suggests that synchronization reflects Gestalt criteria for perceptual grouping (proximity, similarity, continuity, common fate). (Revonsuo; Inner Presence, 273)

Information is measured as the reduction of uncertainty among a number of alternatives. Neural processes underlying consciousness must be highly differentiated to be informative. (Edelman; Universe of Consciousness, 125)

While the net is active, new memory can be added to it through the creation of new associations in accord with the principle of synchronous convergence. (Fuster; Memory in Cerebral Cortex, 113)

Spike-timing-dependent plasticity highlights the essential role of spike timing in modifying network connectivity, a fundamental brain mechanism. (Buzsáki; Rhythms of the Brain, 247)

Theoretical requirement is that the firing of the neurons in each set should be strongly correlated with each other, while at the same time firing of neurons in different sets should be weakly correlated, or not at all. (Crick; Astonishing Hypothesis, 211)

Mutual reentrant interactions, for a brief time, link various neuronal groups in each map to those of others to form a functioning circuit. Neurons that yield such circuits fire more or less in phase with each other, or synchronously. (Edelman; Wider than the Sky, 44)

Binding through Synchronization

Neural network oscillations and any functionality of synchronization needs to be understood. Do neural oscillations and synchronization have any function in cognition and emotion in any sensory modality, e.g. in music? Are neural oscillations and apparent synchronization merely an associated feature of neural network activity? The experts are not consistent in their current interpretations.

The term ‘synchronization’ should be interpreted broadly to mean ‘coherence of neuronal assemblies’ at the group level, not necessarily synchronized individual neuronal firing.

The wide range in the power spectrum observed in the EEG can be considered the result of reentry and recursion among nested hierarchical levels in the neural network. The synchronizations observed are the result of coherence in the neural computations rather than a cause of the network functionality.

Synchronous activity among distant brain areas is often an indication of rapid functional clustering. (Edelman; Universe of Consciousness, 124)

Activity of large populations of neurons can be highly synchronized over a short period. (Edelman; Universe of Consciousness, 124)

Synchronized activity across brain regions at around 40 Hz; a signature of wakefulness, provides a mechanism by which the contents of consciousness can be bound into a unified whole. (Zeman; Consciousness, 301)

Listening to music caused a cascade of brain regions to become activated in a particular order. The cerebellum and basal ganglia were active throughout, presumably supporting the processing of rhythm and meter. (Levitin; Your Brain on Music, 187)

There is little additional information, to the great deal available in the firing rates, from any stimulus-dependent cross-correlations of synchronization that may be present. (Rolls & Deco; Noisy Brain, 62)

Stimulus dependent synchronization might only be useful for grouping different neuronal populations, and would not easily provide a solution to the binding problem in vision. (Rolls & Deco; Noisy Brain, 62)

Oscillation as a linking mechanism -- synchronization by oscillation is effective even through a few and weak links. (Buzsáki; Rhythms of the Brain, 239)

Research results are consistent with the hypothesis that feature binding is implemented by neurons that respond to features in the correct relative spatial locations, and not by temporal synchrony and attention. (Rolls; Memory, Attention, and Decision-Making, 718)

Neurons entrain to each other's rhythms via synchronized action potentials as they communicate with one another. (Schneck & Berger; Music Effect, 140)

Consequence of this dynamic process is the widespread synchronization of the activity of widely distributed neuronal groups. Binds their functionally segregated activities into coherent output. (Edelman; Wider than the Sky, 41)

The time from V1 to inferior temporal cortex takes about 50 ms. There is insufficient time for a return projection from IT to reach V1, influence processing there, and in turn for V1 to project up to IT to alter processing there. Nevertheless, back projections are a major feature of cortical connectivity. (Rolls & Treves; Neural Networks, 240)

Reentry solves the binding problem. Through reentry, for example, the color, orientation and movement of a visual object can be integrated. (Edelman; Wider than the Sky, 41)

Features processed in separate parts of the cortex by different sets of neurons are bound into a complex representation in a matter of 200 ms or so. (Buzsáki; Rhythms of the Brain, 232)

Oscillation as a linking mechanism -- synchronization by oscillation is effective even through a few and weak links. (Buzsáki; Rhythms of the Brain, 239)

Neurons several millimeters apart in the same or different stages of the visual system, and even across the two cerebral hemispheres, have been shown to come together in time transiently by gamma frequency synchronization. (Buzsáki; Rhythms of the Brain, 240)

Although the binding problem was originally formulated in terms of visual object recognition, the idea of various attributes making up a whole is quite general and should apply to all modalities. (Buzsáki; Rhythms of the Brain, 243)

It is unlikely that the processing time in each cortical area is sufficiently long for a stochastic iterative process, or for temporal encoding and synchronization of multiple different populations of neurons. (Rolls & Treves; Neural Networks, 177)

Gamma Oscillation

The cell assembly time-window also matches the time period of gamma frequency oscillations. (Buzsáki; Rhythms of the Brain, 163)

Gamma oscillation is not ubiquitous but localized temporarily to areas engaged in a particular operation. (Buzsáki; Rhythms of the Brain, 244)

Purposeful mental effort causes activity in the beta (15-25 Hz) and the gamma (>30 Hz) bands. (Koch; Quest for Consciousness, 39)

Theta oscillations (5-8 Hz) and Gamma oscillations (30-80 Hz) (Rolls; Memory, Attention, and Decision-Making, 71)

The 40-Hz oscillation is a candidate mechanism to produce temporal conjunction of rhythmic activity over a large ensemble of neurons. (Llinás; I of the Vortex, 124)

Global patterns of organized bioelectrical activity that constitutes the phenomenal level change in about the rate of once in 100 to 300 ms. (Revonsuo; Inner Presence, 345)

Synchronous oscillations on which the phenomenal level is based at the underlying microlevels are probably within the gamma range (20 -- 80 Hz). (Revonsuo; Inner Presence, 345)

We remain agnostic with the respective the relevance of gamma oscillations to conscious perception. (Crick & Koch; Consciousness and Neuroscience, 46)

Temporal Coding for Consciousness-related Binding

The temporal coding hypothesis has been found to be the most promising candidate to account for consciousness-related binding. (Revonsuo; Inner Presence, 281)

Researchers Engle and Singer proposed their overall view of the neural basis of consciousness as a hierarchy of neural assemblies bound together by neural interactions in different frequency bands. (Revonsuo; Inner Presence, 278)

Rodolfo Llinás (1994) proposed that 40 Hz oscillations subserving temporal binding are generated in thalamocortical loops. (Revonsuo; Inner Presence, 278)

Is the information that must be stored present in the firing rates or is it present in the synchronized firing of subsets of neurons? (Rolls; Memory, Attention, and Decision-Making, 659)

Synchronization to implement syntactic binding has a number of disadvantages and limitations. (Rolls; Memory, Attention, and Decision-Making, 325)

When stimulus-dependent temporal synchronization has been rigorously tested with information theoretic approaches, it is then found that most of the information available is in the number of spikes, with rather little, less than 5% of the total information, in stimulus-dependent synchronization. (Rolls; Memory, Attention, and Decision-Making, 325)

Complexity

Scale-free dynamics generate complexity, whereas oscillations allow for temporal predictions. (Buzsáki; Rhythms of the Brain, 135)

Self-Organization

Rest and sleep are the best examples of self-organized operations within neuronal circuits and systems. (Buzsáki; Rhythms of the Brain, 175)

Integrate-and-Fire Neurons

The time constant of pyramidal cells is perhaps most important, because this is the window that determines a cell's integration ability. (Buzsáki; Rhythms of the Brain, 163)

For a hippocampal population of neurons, the optimal window for cell assembly was found to be between 10 and 30 ms. This time scale may be of particular functional significance because many physiological variables share this time window. (Buzsáki; Rhythms of the Brain, 163)

Thalamocortical System

Thalamocortical system does not contain loops so much as highly connected layered local structures with massively reentrant connections. (Edelman; Bright Air, 117)

Thalamic afferents as the crucial determining factor of cortical functions. (Fuster; Memory in Cerebral Cortex, 47)

Functional specificity may not be determined only by thalamocortical afferents but by corticocortical connections as well. (Fuster; Memory in Cerebral Cortex, 47)

The thalamus is no longer viewed as a gigantic array of independent relays, but as a large communication hub that assists in linking large cortical areas in a flexible manner. The principal mechanism of the cortical-thalamic-cortical flow of activity is self-sustained oscillations. (Buzsáki; Rhythms of the Brain, 186)

A very large part of the thalamocortical circuits do not have much to do with primary sensory information. There are important inputs from the cerebellum and the basal ganglia, but the bulk of the afferents are supplied by the neocortex. (Buzsáki; Rhythms of the Brain, 177)

Thalamocortical circuits are associated with thalamic divisions "higher order" nuclei, as opposed to the "first-order" nuclei with specific sensory motor information. (Buzsáki; Rhythms of the Brain, 178)

In partnership with the cortex, the reciprocal excitatory connections of the thalamus are prone to oscillation, and such a mechanism is perfectly poised to mix thalamocortical information. (Buzsáki; Rhythms of the Brain, 179)

Memory Retrieval

Knowledge retrieval would be based on relatively simultaneous, attended activity in many early cortical regions, engendered over several recursions. Separate activities in early cortices would be the basis for reconstructed representations. (Domasio; Convergence Zone, 72)

Distributed Representation

An important characteristic of images is that they have spatially and temporally organized patterns, and in the case of visual, somatosensory, and auditory images, those patterns are topographically organized. (Damasio; Making Images, 20)

Topographic representations can be committed to memory in the form of nontopographically organized dispositional representations, and can be stored in dormant form in both cortical regions or subcortical nuclei. (Damasio; Making Images, 20)

Subsequent reactivation of those dormant dispositional representations, followed by signaling from their storage sites back to early sensory cortices, can regenerate topographically organized representations. (Damasio; Making Images, 20)

Retroactivation process uses the rich connectional patterns of feed-forward and feedback that characterize the architecture of cortical regions and subcortical nuclei.

Topographic representations can arise in turn as a result of signals external to the brain in the perceptual process, or in the process of recall from signals inside the brain, coming from memory records held in dispositional representation form. (Damasio; Making Images, 20)

Gestalt Laws

Gestalt psychologists have long known that the whole is often faster recognized than its parts, indicating that object recognition is not simply representation of elementary features, but the result of bottom-up and top-down interactions, in harmony with the architectural organization of the cerebral cortex. (Buzsáki; Rhythms of the Brain, 243)

Brains perform inferences using parallel activity of millions of neurons, -- Perception can elegantly integrate both bottom-up and top-down information. (Thagard; Brain and the Meaning of Life, 71)

Importance of both bottom-up and top-down processes is exemplified by Ullman's model of information flow in the visual cortex for object recognition. (Niebur & Koch; Computational Architectures for Attention, 167)

Counterstreams sequence-seeking model is a bidirectional search performed by top-down and bottom-up streams seeking to meet. (Ullman; Sequence Seeking Counterstreams, 270)

Bayesian Brain and Attractor Dynamics

Amalgamate and reconcile the ideas of Doya, Bayesian Brain with Rolls, Noisy Brain.

Brain operates as a reality emulator

Brain operates as a reality emulator. (Llinás; I of the Vortex, 13)

The brain constructs an explicit, multilevel, symbolic interpretation of parts of its environment. (Koch and Crick; Neuronal Basis, 109)

Internal generation of a world-analog. (Tononi; Sleep and Dreaming, 100)

Brain must combine inputs from seven different sensory mode modalities and then generate a complete internal representation that does not depend exclusively on any one input. (Kandel; Search of Memory, 308)

Memory Systems in the Brain

Short-term and long-term memory share much of the same cortical substrate and simply reflect different activation states of that substrate. (Fuster; Memory in Cerebral Cortex, 4)

Brain's multiple memory systems are mediated by different brain structures and systems. (Eichenbaum; Neuroscience of Memory, v)

One short-term memory system is in dorsolateral prefrontal cortex, area 46. This is involved in remembering the locations of spatial responses. (Rolls & Treves; Neural Networks, 246)

Another short-term memory system is implemented in the inferior temporal visual cortex. This memory is for whether a particular visual stimulus (such as the face) has been seen recently. (Rolls & Treves; Neural Networks, 246)

Another short-term memory system is human auditory-verbal short-term memory, which appears to be implemented in the left hemisphere as a junction between the temporal, parietal, and occipital lobes. (Rolls & Treves; Neural Networks, 247)

Working memory depends on sustained activation of portions of long-term memory, which must be construed in a broad sense. Perceiving, recognizing, understanding, and rehearsing are all abilities that result from continual refinements involving long-term memory. (Postle; Activated Long-Term Memory, 333)

Short-term retention of information in working memory is supported by sustained activity in the same nonPFC brain regions that process this information in situations that do not require memory. (Postle; Activated Long-Term Memory, 333)

Broad classification of brain regions and memory types, allocating working memory to bilateral regions in frontal, as well as parietal and temporal lobes. (Halford; Relational Processing, 262)

Genetics, Randomness, Selection — Development of Neural Network

Theory of Neuronal Group Selection (TNGS) - (1) Developmental selection, (2) Experimental selection, (3) Reentry (Edelman; Remembered Present, 43)

Epigenesis by Selective Stabilization, (1) growth, (2) transient redundancy, (3) selective stabilization, - (diagram) (Changeux; Neuronal Man, 228)

Minicolumns as Cortical Functional Unit

The column, or minicolumn, could be conceptualized as the ubiquitous and irreducible functional unit of the cortex. (Fuster; Memory in Cerebral Cortex, 49)

Minicolumn contains about 110 cells. This figure is almost invariant between different neocortical areas and different species of mammals, except for the striate cortex of primates, where it is 260. (Mouthcastle; Organizing Cerebral Function, 37)

Since the total volume of the human neocortex is about 1000 cm³, and assuming an average thickness of 2,500 µm, neocortex of the human brain has a surface area of about 4,000 cm² and contains about 600 million minicolumns and on the order of 50 billion neurons. (Mouthcastle; Organizing Cerebral Function, 37)

A cortical column corresponds to the mass of neural tissue lining under a 1 mm x1 mm square area along the surface of the cortex and spanning the 1.5- to 2.0-mm thickness of the cortical sheet. (LaBerge; Attention, the Triangular Circuit, 292)

The number of neurons contained within the typical cortical column is said to be on the order of 100,000, while a typical column in the area V1 contains nearer to 180,000 neurons. (LaBerge; Attention, the Triangular Circuit, 293)

The cortical column is sometimes subdivided into minicolumns of various sizes, having widths of about 400, 200, and 30 µ wide, containing about 23,000, 5500, and 140 neurons, respectively. (LaBerge; Attention, the Triangular Circuit, 293)

Vertical groups of neurons are centered around the long vertical dendrites of layer 5 neurons, forming a cylinder-like volume having a diameter of approximately 30 µ.(LaBerge; Attention, the Triangular Circuit, 293)

The minicolumn, which contains about 140 neurons, has been proposed as a functional unit of the cortex. (LaBerge; Attention, the Triangular Circuit, 293)

Network Structure, Building Blocks

Connections among cells are not precisely specified in the genes of the animal. (Edelman; Bright Air, 23)

Long reentrant connective loops have been identified that originate in frontal areas, course through the basal ganglia and the lateral thalamus, and return to frontal areas. (Fuster; Memory in Cerebral Cortex, 77)

Precise point-to-point wiring cannot occur; the variation is too great for the information stored in the genome. (Edelman; Bright Air, 25)

Neurons generally send branches of their axons out in diverging arbors that overlap with those of other neurons. (Edelman; Bright Air, 25)

In some regions of the developing nervous system up to 70% of the neurons die before the structure of the region is completed. In general, uniquely specified connections cannot exist. (Edelman; Bright Air, 25)

Neurons have treelike arbors that overlap and ramify in myriad ways. (Edelman; Bright Air, 69)

Any connection at any neuron can be a member of many different networks, of many different memories. (Fuster; Memory in Cerebral Cortex, 98)

Dynamic interaction between perception and memory, which conforms to the almost complete overlap of processing and representational networks in the cortex. (Fuster; Memory in Cerebral Cortex, 89)

In memory networks the connectivity up the hierarchies include convergence as well as divergence every step of the way and, and addition, there is recurrence and backflow. (Fuster; Memory in Cerebral Cortex, 94)

Cortical Columns, Structured Neuronal Groups, Single-Neuron Functionality

Structured neuronal groups containing up to 10,000 neurons are formed during embryogenesis and development. Intrinsic connections within a group and extrinsic connections among groups are specified by gene programming and synaptic selection. (Edelman; Group Selection and Reentrant Signaling, 94)

Understanding cognitive functions such as object recognition, memory recall, attention, and decision-making requires single neuron data to be closely linked to the computational models of how the interactions between large numbers of neurons in many networks of neurons allow these cognitive problems to be solved. (Rolls; Memory, Attention, and Decision-Making, 3)

Whether a neural symbol is best thought of as a scalar (one neuron) or a vector (a group of closely associated neurons as in population coding in the superior colliculus) is a difficult question. (Koch; Neuronal Substrate of Visual Consciousness, 248)

Stochastic Neuronal Behavior, Populations of Neurons

Because single neurons have small and uncertain effects on other neurons, the cortical description must be carried out in terms of neuronal populations of cells rather than at the level of individual cells. (Stevens; Cortical Theory, 242)

The overall stochastic nature of neuronal behavior suggests that the physiologically meaningful signal from cortex should be the average firing rates of a population of perhaps 100 to 1000 neurons near a particular cortical site. (Stevens; Cortical Theory, 243)

The behavior of cortex at a particular point is described by the firing in a population of neurons. The total firing that represents this population would be determined by a weighted average of the appropriate neurons in the cortical region that surrounded the point, perhaps with weights that are described by a spatial Gaussian. Moving from one cortical location to an adjacent one, the variables describing cortical state would vary continuously with cortical position. (Stevens; Cortical Theory, 243)

Modeling Neural Network Functionality

Modeling studies emphasize the necessity of considering the collective action of neural assemblies in information processing. (Motter; Neurophysiology, 53)

Understanding how neurons represent information is fundamental for understanding how neurons and networks of neurons read the code from other neurons, and the actual nature of the computation that could be performed in a memory network. (Rolls; Memory, Attention, and Decision-Making, 3)

Neurocomputational models enable the single neuron level of analysis to be linked to the level of large-scale neuronal networks and the interactions between them, so that large-scale processes such as memory retrieval, object recognition, attention, and decision-making can be understood. (Rolls; Memory, Attention, and Decision-Making, 3)

Correlation and Synchronization of Neural Network Activity Patterns

Attentional mechanism makes use of correlated firing -- what matters is not just the average rate of firing of a neuron, but the exact moments at which each neuron fires. (Crick; Astonishing Hypothesis, 211)

Neural Network Activity Patterns and Relations

The ability to process relations is at the core of analogical reasoning and of many other higher cognitive processes. (Halford; Relational Processing, 262)

Prefrontal regions are involved in the retrieval and monitoring of relational information, parietal regions are involved in the maintenance of relational structures (i.e., the explicit dimensions over which the relations are defined), and temporal regions are involved in the formation of bindings between related items. (Halford; Relational Processing, 263)

Recurrent Colatterals in Neural Networks

The main factors that determine the maximum number of memories that can be stored in an autoassociative network are thus the number of connections on each neuron devoted to the recurrent collaterals, and the sparseness of the representation. (Rolls; Memory, Attention, and Decision-Making, 568)

Memory capacity arises from combinations of synaptic alterations in reentrant circuits. (Edelman; Universe of Consciousness, 105)

The effect of inhibition in cortical circuits is not simply an antagonistic balancing act versus excitation. Inhibition acts as a trimming damper on the explosive growth of positive feedback gain of excitatory cortical circuits. Relatively small amounts of inhibition provided at the correct time can shape the amplification of information. (Motter; Neurophysiology, 53)

Recurrent collaterals from pyramids usually follow a horizontal or oblique course and spread around as much is 2 to 3 mm. (Fuster; Memory in Cerebral Cortex, 48)

Oscillatory activity of various frequencies may be the expression of the recurrent activity between cortex and thalamus that has been postulated to be essential for the regulation of all sleep stages. (Fuster; Memory in Cerebral Cortex, 286)

The master loop networks are characterized by an architecture of recurrent connectivity. That special architecture supports closed loops of coherent bioelectrical activity. The loops extend "horizontally" across the cortex (like the feedback sweep in the ventral visual stream) and also "vertically" between the cortex and the thalamus ("specific" and "nonspecific" thalamocortical loops). (Revonsuo; Inner Presence, 319)

Competitive Selection of Neural Networks

Executive networks grow in the prefrontal cortex and are essentially selective; they select neurons and circuits among those that have been overproduced in earliest stages of development, while other neurons and terminals undergo regression and disappearance. (Fuster; Prefrontal Cortex, 18)

"Selective stabilization" is the result of competition for inputs on the part of neurons and terminals. (Fuster; Prefrontal Cortex, 18)

Consolidation is a competitive process in which some aspects of memory for the original event are forgotten, while those that remain are strengthened. (Squire; Memory and Brain, 205)

In the hierarchy of cortical stages, convergence and competition are key aspects of the processing. (Rolls; Memory, Attention, and Decision-Making, 31)

Attention is an emergent property of slow competitive interactions that work in parallel across the visual field. (Webster; Neuroanatomy of Visual Attention, 30)

Voluntary action is probabilistically determined by resolution of conflict, in frontal cortex, between competing neural assemblies of diverse origin. (Fuster; Memory in Cerebral Cortex, 296)

Reflexive (bottom-up) attention happens independently in the two hemispheres, while voluntary attention involves hemispheric competition, with control preferentially lateralized to the left hemisphere. (Gazzaniga; Human, 292)

Intermingling of past and present requires both long-range intermodal and short range intramodal competition. (Baars; Neuronal Mechanisms of Consciousness, 271)

Reticular nucleus of the thalamus has inhibitory connections with the specific nuclei and can select or gate various combinations of their activity. (Edelman; Universe of Consciousness, 107)

Nucleus reticularis thalami (nRt) seems like an ideal device for achieving global input competition, being composed solely of inhibitory cells and also having remarkable lateral conductivity similar to that of the outer plexiform layer in the retina. (Baars; Neuronal Mechanisms of Consciousness, 271)

Conscious content reflects an endless interplay of competition and cooperation between possible inputs and top-down influences. (Baars; Neuronal Mechanisms of Consciousness, 272)

The evolution of complex systems is based upon cooperation and competition among its parts, and in the process certain constituents gain dominance over the others. This dominance is called the "attractor" property in chaos theory. (Buzsáki; Rhythms of the Brain, 14)

Degeneracy

Rolls’s circuit models must be configured to have the property of degeneracy, a property pointed out by Edelman. Graceful degradation is a feature of degeneracy. Here are a sample of paraphrases related to degeneracy, redundancy, and associative neuronal groups.

Countless examples of degeneracy in the brain. The complex meshwork of connections in the thalamocortical system assures that a large number of different neuronal groups can similarly affect the output. A consequence of degeneracy is that certain localized neurological lesions may often appear to have little effect. (Edelman; Universe of Consciousness, 87)

In the next time period, different neurons and neuronal groups may form a structurally different circuit, which nevertheless has the same output. (Edelman; Wider than the Sky, 45)

These different circuits are degenerate - they are different in structure, but they yield similar outputs to solve the binding problem. (Edelman; Wider than the Sky, 45)

As a result of reentry, the properties of synchrony and coherency allow more than one structure to give a similar output. (Edelman; Wider than the Sky, 45)

All memory is essentially associative. (Fuster; Memory in Cerebral Cortex, 11)

The neural network is an amalgam of many associative neuronal groups selected by prior experience. When offered a new input signal, many subsets of neuronal connections within the network will recognize (closely conform to) a signal similar to ones that previously have fashioned the network (accumulated pattern of synaptic strengths via Hebbian processes). Many neuronal groups (synaptically-connected patterns) can respond more or less well to an input signal. The tolerance for variation in the signal (degeneracy), ensures wide distribution of information and a degree of stimulus constancy -- i.e. equivalents of response to variants of the same stimulus. (Fuster; Memory in Cerebral Cortex, 90)

Two fundamental features of cortical processing and memory -- (1) probabilistic nature of cortical processing, and (2) robustness of the memory. (Fuster; Memory in Cerebral Cortex, 86)

Redundancy in the system; conscious awareness of amygdala activity can come about in several ways. (LeDoux; Emotional Brain, 285)

Perhaps the most important and useful property of autoassociation networks is that they complete an incomplete input vector allowing recall of a whole memory from a fraction of it. (Rolls; Memory, Attention, and Decision-Making, 564)

Memory is robust, since altering a few connections will usually not alter its behavior very much. (Crick; Astonishing Hypothesis, 184)

Degeneracy implies an approximate or highly probable fit between the structure of the network, in connective terms, and the structure of the external Gestalt in relational terms. (Fuster; Cortex and Mind, 92)

Edelman's Neural Darwinism shares with Hayek's model the merits of distribution of information, probabilistic response, and robustness. (Fuster; Memory in Cerebral Cortex, 91)

Edelman’s ‘Remembered Present’ and Working Memory

Working memory is closely related to consciousness.

Episodic buffer of working memory is a limited capacity temporary storage system that integrates information from a number of sources across space and time. (Martin; Cognitive Neuropsychology, Working Memory, 184)

Episodic buffer is assumed to be dissociable from LTM, but interacts with (introduces information into, and retrieves information from), long-term memory. (Martin; Cognitive Neuropsychology, Working Memory, 184)

Episodic buffer serves as an interface to integrate representations from a number of systems using, 'common multidimensional code.' (Martin; Cognitive Neuropsychology, Working Memory, 185)

It is proposed that the episodic buffer serves as an interface between memory and conscious awareness. (Martin; Cognitive Neuropsychology, Working Memory, 185)

Emotions

Emotions and the limbic system are topics of intensive ongoing research. The limbic structures should include the orbitofrontal cortex and anterior cingulate gyrus as well as subcortical structures. A major challenge for understanding the limbic system is to understand the brain’s emotional and cognitive response to music.

Emotions are among the very oldest of our brain properties. (Llinás; I of the Vortex, 156)

Some sort of basic emotional state is present whenever you are conscious. (Greenfield; Private Life of Brain, 16)

Classified the emotions-proper in three tiers: (1) background emotions, (2) primary emotions, and (3) social emotions. (Damasio; Looking for Spinoza, 43)

Emotions are sandwiched between the basic survival kit (regulation of metabolism; simple reflexes; motivations; biology of pain and pleasure) and the devices of high reason, but are still very much a part of the hierarchy of life regulation devices. (Damasio; Feeling of What Happens, 54)

The basic program routines of emotions are stereotypical at all body levels at which they are executed -- external motions; visceral changes in the heart, lungs, gut, and skin; and endocrine changes. (Damasio; Self Comes to Mind, 123)

Background emotions can be distinguished from moods, which refer to the sustaining of a given emotion over long periods of time. (Damasio; Looking for Spinoza, 43)

Emotion and motivation are linked by the property that both involve rewards and punishers. (Rolls; Emotion Explained, 1)

Value systems - memories of reward and punishment acquired during past behavior. (Edelman; Universe of Consciousness, 109)

Orbitofrontal cortex has a key role in representing primary reinforcers, and learning and rapidly changing associations between stimuli and primary reinforcers, and thus is important in many types of emotional and motivational behavior. (Rolls; Memory, Attention, and Decision-Making, 183)

Neural Network Activity and Conscious Experience

Activity of a group of neurons can contribute directly to conscious experience if it is part of a functional cluster, characterized by strong mutual interactions among a set of neuronal groups over a period of hundreds of milliseconds. (Edelman; Universe of Consciousness, 139)

The number of spikes in a fixed time window over which a postsynaptic neuron could integrate information is a more realistic code. This time might be of the order of 20 ms for single receiving neuron, or much longer if the receiving neurons are connected by recurrent collateral associative synapses and so can integrate information over time. (Rolls; Memory, Attention, and Decision-Making, 703)

A rapid readout of information from any one stage is important, for the ventral visual system is organized as a hierarchy of cortical areas, and the neuronal response latencies are approximately 100 ms in inferior temporal visual cortex, and 40-50 ms in the primary visual cortex, allowing only 50-60 ms of processing time for V1 -- V2 -- V4 -- inferior temporal cortex. (Rolls; Memory, Attention, and Decision-Making, 702)

Orbitofrontal face responsive neurons tend to respond with longer latencies (130-220 ms typically) than temporal lobe neurons (80-100 ms). They also convey information about which face is being seen, by having different responses to different faces. (Rolls; Memory, Attention, and Decision-Making, 165)

Link to — Consciousness Subject Outline