BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

Research Report

Neuronal correlates to consciousness. The Hall of Mirrors metaphor describing consciousness as an epiphenomenon of multiple dynamic mosaics of cortical functional modules
Luigi Francesco Agnatia,, 1, 2 , Diego Guidolinb, 1 , Pietro Cortellic, 2 , Susanna Genedanid , Camilo Cela-Condee , Kjell Fuxef
a

Fondazione IRCCS San Camillo, Venezia Lido, Italy Department of Molecular Medicine, University of Padova, Padova, Italy c Department of Neurological Sciences, Alma Mater Studiorum, University of Bologna, Italy d Department of Biomedical Sciences, University of Modena and Reggio Emilia, Via Campi 287, Modena, Italy e University of the Balearic Islands, Palma de Mallorca, Spain f Department of Neuroscience, Karolinska Institutet, 17177 Stockholm, Sweden
b

A R T I C LE I N FO Article history: Accepted 4 January 2012 Available online 11 January 2012 Keywords: Consciousness Cellular sentience Functional module mosaic Mirror neuron Internal theater Hall of Mirrors

AB S T R A C T Humans share the common intuition of a self that has access to an inner theater of mind (Baars, 2003). The problem is how this internal theater is formed. Moving from Cook's view (Cook, 2008), we propose that the sentience present in single excitable cells is integrated into units of neurons and glial cells transiently assembled into functional modules (FMs) organized as systems of encased networks (from cell networks to molecular networks). In line with Hebb's proposal of cell assemblies, FMs can be linked to form higher-order mosaics by means of reverberating circuits. Brain-level subjective awareness results from the binding phenomenon that coordinates several FM mosaics. Thus, consciousness may be thought as the global result of integrative processes taking place at different levels of miniaturization in plastic mosaics. On the basis of these neurobiological data and speculations and of the evidence of mirror neurons the Hall of Mirrors is proposed as a significant metaphor of consciousness. This article is part of a Special Issue entitled: Brain Integration. 2012 Elsevier B.V. All rights reserved.

1.

Introduction

Humans seem to share a common intuition of a self that has access to conscious sensations, inner speech, images and thoughts (self-consciousness). This intuition may be

metaphorically described as a theater of mind (Baars et al., 2003), and conscious events can be defined as those brain activities a subject can accurately report in optimal conditions (Baars et al., 2003). Once the assumption of a chemicophysical basis of consciousness has been accepted, we are

Corresponding author at: Fondazione IRCCS San Camillo, Via Alberoni, 70, Venezia Lido, Italy. Fax: +39 041 731330. E-mail address: luigiagnati@tin.it (L.F. Agnati). 1 These authors equally contributed to the paper. 2 Dedicated to Professor Angelo Pierangeli (19322010) and to Professor Pasquale Montagna (19502010), University of Bologna. 0006-8993/$ see front matter 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2012.01.003

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

confronted with the question of how brain activity leads to conscious experiences, i.e., what are the neuronal correlates of consciousness (NCC). As Chalmers emphasizes, the search for NCCs is the cornerstone of the recent resurgence of the science of consciousness (Chalmers, 2000). NCCs have become important in the wake of the enthusiasm generated by neuroimaging techniques, which correlate specific cognitive tasks with the activation of restricted brain areas, making the relationships between psychology and neurobiology more realistic and fruitful (Monti et al., 2010). This research effort allowed to show that conscious perceptual processing involves the sequential activation of cortical networks at several brain locations with the onset of oscillatory synchronous activity (Gray et al., 1989a, 1989b). Hence, studies on consciousness must explore both a temporal and spatial dimension. As far as the temporal dimension of conscious events is concerned, Pppel and Logothetis (1986) investigated reaction times to visual stimuli, and proposed that perceptual processing operates in basic units of 30 ms, while conscious episodes composing the conscious present can be extended to periods of 2 or 3 s (Pppel, 1994). Studies in humans based on the event-related potential (ERP) paradigm allow an estimate of the temporal dynamics of conscious perceptual processing. ERP measures the temporal location of brain events correlated with conscious processing evoked by stimulus presentation. These studies indicate that the ERP P300 and N400 components are related to working memory and/or attention functions that probably involve conscious processing (Coull, 1998; Knight, 1997). The corresponding brain events occur from 300 to 400 ms after stimulus presentation. According to Pereira and Furlan (2009), these data suggest that 200 ms can be a good estimation of the minimum temporal duration from stimuli presentation to the formation of a conscious percept. Therefore, the neuronal activity required to support conscious processing would need to be sustained from 200 ms to 2/3 s. Consistently, studies on subliminal perception (Murphy and Zajonc, 1993) reveal that a visual stimulus presented for only 5 ms and followed by a mask is not consciously perceived, although it may have unconscious priming effects. This implies that a threshold input firing for perceptual consciousness has not been reached (Pereira and Furlan, 2010). Since such an appropriate temporal dynamics (from 200 ms to 2/3 s) of conscious events is the result of processes linking a large network of brain systems (Buzski, 2007; Laureys, 2005), it follows that the spatial dimension (i.e. the brain's morpho-functional organization) is a key feature to consider in order to derive deductions on the communication modes involved in brain integrative processes leading to conscious percepts. In the present review aspects of the brain's morphofunctional organization corresponding to increasing levels of integration will be addressed (also based on data and hypotheses proposed by our group). In particular, the following issues and their relevance for consciousness formation will be analyzed: 1. Recent findings on the special features of neurons and astrocytes (Agnati et al., 1995; Allman et al., 2005; Pereira and Furlan, 2010; Premack, 2007 ) and the possible existence

of a proto-consciousness phenomenon founded on the mechanism of cell sentience (Cook, 2008; Sevush, 2006); 2. The Volume and Wiring Transmission (VT, WT) modes of communication processes in the brain (Agnati et al., 2005a, 2010a) will be briefly examined since they are the fundamental neurobiological mechanisms that allow the dynamic formation of cell assemblies and the integration of their activity. 3. The concept of mosaics of computational elements will be introduced (Agnati et al., 1982, 1990, 2007a, 2008). In particular, the cellular mosaic (Functional Module, FM) formed by neuronastroglial interactions will be analyzed and proposed to be capable of a first-level integrative sentience; 4. Mechanisms for large-scale integration of FMs into mosaics of higher-order leading to the formation of the neuronal correlates of consciousness allowing the integration of different percepts will be analyzed from the neurobiological perspective. Finally, these aspects will be used to propose a new interpretative metaphor of consciousness, namely the brain as a Hall of Mirrors.

2. Special features of neurons and astrocytes in the human brain

A neurobiological approach to the human capacity for auto-reflection should start at the lower level to clarify the fundamental question of what neural substrates make a human being human (DeFelipe et al., 2002). Plainly, these investigations focused on the amount of neurons and synaptic contacts, the presence of some type of special neurons, the properties of astrocytes and, of major potential interest, a comparison of possible specific features in some transmitteridentified neuronal systems such as monoamine systems (for details see Fuxe et al., 2010).

2.1.

Neuronal aspects specific to the human brain

Let us start by examining some data on the peculiarities of neurons present in the human brain.

2.1.1.

Quantitative evaluations of neural and synaptic densities

A crucial quantitative difference in neuronal density (neurons/ mm3 in layers IVI) distinguishes the human brain from other species. Neuronal density in the cerebral cortex is lower in humans (24 186/mm3) than in rats (54 483/mm3) and mice (12 0315/mm3), whereas the number of synapses per neuron is higher in humans (29 807) than in rats (18 018) and mice (21 133) (DeFelipe et al., 2002). The number of dendritic spines of basal dendrites of layer III pyramidal neurons also differs in mouse and human temporal cortex. The mean number (mean SEM) of spines per 10 m segment is 10.9 0.5 for cells in temporal cortex of mice and 14.2 0.4 with respect to the temporal cortex of humans (Benavides-Piccione et al., 2002). Study of the size of spine heads revealed that the mean area in the temporal cortex of mice was smaller than in humans (mean SEM: 0.37 0.01 m2 and 0.59 0.01 m2, respectively) and the spine necks in the temporal cortex of

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

mice were shorter (0.73 0.01 m) than those of humans (0.94 0.01 m) (Benavides-Piccione et al., 2002). From a functional standpoint, data on the density of spines are in agreement with the observation of a higher number of synapses in humans than in mice or rats. The larger volume of the spine head again indicates a more efficient transfer of information between neurons in the human brain as spine head volume is directly proportional to postsynaptic density, the number of postsynaptic receptors, the pre-synaptic number of docked synaptic vesicles and the rapidly releasable pool of neurotransmitters. In addition, spines with longer necks show longer time constants of calcium compartmentalization than spines with shorter necks. All these data can be interpreted to reflect that the human brain contains the highest density of local circuits (Alonso-Nanclares et al., 2008; DeFelipe et al., 2002) , which are likely involved in memory processes (Douglas et al., 1995; Goldman-Rakic, 1995; Romo et al., 1999; Wang, 2001 see Fig. 1). It should be noticed that a high density of synaptic contacts per neuron allows several alternative pathways in the high density local circuits of the human brain.

2.1.2.

Special types of neurons

The qualitative and quantitative aspects of special types of neurons like the giant Betz nerve cells of the primate motor cortex, and especially the von Economo neurons (or spindle neurons) are worthy of attention. The giant Betz nerve cells of the primate motor cortex are not clustered but evenly distributed in the inner pyramidal cell layer (layer V), whereas clusters of large giganto-cellular nerve cells exist in layer V of the primary motor cortex of giraffe and sheep (Badlangana et al., 2007). This observation is certainly interesting, but for the moment has no clear functional correlate with the peculiar features of motor control in primates. More interesting for their possible functional correlates are the observations on the qualitative and quantitative aspects of the von

Economo neurons (VENs) since they serve to differentiate the human brain from other species (Allman et al., 2005; Premack, 2007). VENs have also been described in the cortex of the elephant (Hakeem et al., 2009) and in the cetacean brain (Butti et al., 2009) hence in animal species that pass the mirror test of being able to recognize themselves in a mirror (de Veer et al., 2003; Gallup, 1970; Plotnik et al., 2006). The VENs are large, bipolar cells located in layer 5 of the anterior cingulate, fronto-insular and dorsolateral (dysgranular) prefrontal cortex (Fajardo et al., 2008). They are distinguished from pyramidal cells because VENs have only a single large basal dendrite, whereas pyramidal cells have an array of smaller basal dendrites extending from the cell body. Among the morphological and functional human differences, Premack (2007) cites evidence showing that our species has many more and larger VENs than apes. In particular, the perimeters of VENs are far wider: an average of 51 m in humans compared with 36 m in chimpanzees and in monkeys. This increase is produced by an enlarged dendritic tree and higher density of synapses, leading to an increased density of local circuits. As mentioned above, human VENs are located in only two parts of the brain, the anterior cingulate cortex and the fronto-insular cortex. These areas appear to be involved in socially crucial tasks, such as empathy, feelings of guilt, and embarrassment. In addition, human columns (see below) in the left planum temporale an area involved in language and perhaps music are organized differently from those of chimpanzees and rhesus monkeys. However, they may alternatively represent a structural plasticity process to cope with the demands of achieving integration and communication in very large brains. It is interesting to note that VENs develop late in ontogeny. They first appear in very small numbers in the 35th week of gestation, and the adult number is attained only at 4 years of

Fig. 1 Schematic example showing how increasing the number of synaptic contacts per neuron also increases the number of potentially available reverberant circuits. As illustrated, a higher number of synapses per neuron allows several alternative pathways to be exploited for signal reverberation.

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

age. In all great apes and in postnatal human brains, VENs are around 30% more numerous in the fronto-insular cortex of the right hemisphere compared with their number in the corresponding cortex of the left side. It may therefore be surmised that VEN predominance in the right hemisphere in the postnatal period is related to the right hemisphere specialization in humans, at least to the social emotions that are of basic importance for self-consciousness. VENs may also be involved in the fast intuitive assessment of complex situations (Allman et al., 2005).

2.2.

Astrocytic aspects specific to the human brain

Astrocytes are also excitable cells and play important roles in information processing as can be deduced from evolutionary data showing increasing numbers and a more complex organization of astrocytes in the human brain. The astrocyte/ neuron ratio increases in evolution from 1:25 in the leech to 3:2 in the human brain (Pereira and Furlan, 2010). These data likely implicate these cells in the evolution of increasingly complex brain functions.

2.2.1.

Special types of astrocytes

Of major interest are recent findings by Oberheim et al. (2006, 2009) indicating the existence of different types of astrocytes, besides the classical protoplasmic and fibrous astrocytes, the so-called varicose projection astrocytes that seem exclusive to the human brain. These cells present more spiny processes than exhibited by typical protoplasmic astrocytes and extend one to five, essentially unbranched, millimeter long fibers within the deep layers of the cortex. The evenly spaced varicosities suggest specialized structures or compartmentalization of cellular elements along the great distance of the fibers. These astrocytes are connected to each other and with protoplasmic astrocytes by gap junctions, forming a brain-wide network allowing long-distance communication across cortical layers or even between gray and white matter (Oberheim et al., 2009).

2.3. Possible existence of a proto-consciousness founded on single-cell sentience

According to Cook (2008) and Sevush (2006), a single-neuron theory of consciousness can be proposed. In particular, Cook analyses the capability of neurons, as excitable cells, of sensing the extracellular environment during action potential generation. In particular, these authors suggest that a sort of neuronal sentience emerges as a consequence of the momentary openness of the plasma membrane to the external world during the action potential. This is not feeling in the sense of a human being having emotional experiences, but it is feeling at the cellular level: the neuron in isolation senses its extracellular fluid by making contact with its surrounding world (e.g. by absorbing a small part of the local ionic charge) as it goes about its cognitive business of synaptic communications (Cook, 2008). Thus, in a very simplified way, Cook's proposal can be summarized by stating that while the synaptic interconnections in a neuronal network are the biological substrate of the proto-phenomenon of cognition, the ion-flows

during the action potential are the biological substrate of the proto-phenomenon of sentience. It can also be surmised that the sensing capability of neurons is likely broader than that proposed above since these cells can derive information on their external environment not only by ion-fluxes but also via extrasynaptic receptors that decode a variety of volume transmission signals (see below and Agnati et al., 2010a). Similar concepts could apply to astrocytes as well. In fact, they can sense their environment, converting such an information into detailed wavelike patterns of calcium ions and ATP signaling that can be directly communicated to other astrocytes and neurons (Pereira and Furlan, 2010). As a consequence, a peculiar combination of cognition and sensitivity is present in the nervous tissue (Guidolin et al., 2011), not only in neurons but also in glial cells, probably representing, in agreement with Cook's proposal, a prerequisite for the emergence of the highest functions performed by the CNS, such as subjectivity and consciousness. To further illustrate this point, other systems in which the above two characteristics are not simultaneously present can briefly be considered. On the one side, it is well known that networks of artificial devices with fixed activation thresholds, and connected by simple excitatory or inhibitory synapses are able to perform any cognitive task that can be adequately defined (Koblauch et al., 2010; McCulloch and Pitts, 1943; Penrose and Gardner, 1999). Thus, at least in principle, all forms of information processing (i.e. cognition) could be performed by such a circuitry. However, whether something like subjectivity and consciousness could emerge from such a computational system is far from being obvious (Freeman, 1997; Werner, 2007). On the other side, there are excitable cells, such as cardiomyocytes, sharing with neurons the possibility to sense their external environment, since they are opened to ion-fluxes during their electrical activity. However, the lack of adequate connectivity schemes, primarily the lack of both excitatory and inhibitory connections, prevents the emergence of either cognitive functions or subjectivity. Summing up, a comment and three considerations can be added to Cook's interesting proposal (see Fig. 2): a.) The comment: to be properly tuned and to have longlasting effects the sentience of a neuron needs a term of comparison and likely has to impinge on a core neuronal structure. Thus, we suggest two environments should be considered: the extracellular fluid as the classical internal milieu and the cytoplasm. The Energide (formed by the nucleus, microtubules and other satellite structures) is the basic unit of living organisms (see Baluka et al., 2006). Thus, the Energide (i.e., the core structure of the cell) interacts with the cytoplasm that, in turn, interacts with the interstitial fluid, and hence with the medium classically known as the internal milieu (Agnati et al., 2009a). In neurons the cytoplasm around the Energide may represent the term of comparison for neuronal sentience. It could also be surmised that in view of the fundamental functions carried out by the Energide, especially by the nucleus, long-lasting effects on a neuron's sentience are the result of the modulatory

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

Fig. 2 Schematic representation of a neuron. Signal transmission is the essential function of the neuron, but what flows down the axon is a sudden process of adjustment of the membrane potential. It is associated with a material flow of ions orthogonal to the direction of signal transmission. According to Cook (2008) these two cellular processes could represent the protophenomenon of cognition and sentience respectively. The scheme points out the key role that the comparison between the extra-cellular fluid (i.e., the classical internal milieu) and the cytoplasm, the private cellular internal environment around the Energide (see Baluka et al., 2006) may have to realize such a cell sentience. Furthermore, the possible contribution of Volume Transmission (VT) signals of both chemical and physical nature (as, for instance, the pressure waves caused by the arterial blood pressure pulses in cerebral arteries) to the process is indicated. For further details, see text.

actions of the private internal medium (i.e., the cytoplasm around the nucleus) on the Energide. b.) First consideration: the presence of receptors for transmitters outside the synapse and not only at soma and dendritic level, but also at axonal levels (Aoki et al., 1998; Riad et al., 2000; Semyanov and Kullmann, 2002) points to the possible important role of volume transmission (see below and Agnati et al., 2010a) in the sentience of single neurons by modulating not only the composition of the cytoplasm and the activation of molecular networks, but also the generation of action potentials at soma-dendritic level and hence the biochemical features of the cytoplasm around the Energide. c.) Second consideration: according to the Tide Hypothesis (Agnati et al., 2005a), the piston-like movements of the entire brain towards the occipital foramen cause pulsatory distortions of the cell membranes. Thus, mechanosensitive ion channels present at neuronal (Zurborg et al., 2007) and astrocyte (Ostrow and Sachs, 2005) membrane level are stimulated. Their activation can affect the ion movements across the plasma membranes and hence both the generation of action potentials (i.e., sentience of neurons) and the integrative action of astrocyte networks. d.) Third consideration: visceral homeostatic processes can affect brain function via neuronal visceral (especially vagal) afferences or endocrine signals. It should also be noted that astrocytic networks are in a privileged location to respond to blood and cerebrospinal fluid signaling mediated by endothelial and ependymal cells respectively. These signals include small molecules like hormones and neuropeptides. All these aspects should be entertained in investigating the integrative actions of the brain since they underpin the global sentience that represents subjective awareness.

3. Wiring and Volume Transmission and the supra-cellular organization of the cortex
A basic feature of all processes leading to consciousness is the existence of communication modes between cells leading to the formation of supracellular forms of organization allowing the integration of information.

3.1. 3.1.1.

Communication modes in the brain Wiring Transmission (WT)

This is basically a signaling process along a physically welldelimited channel like a wire. Two types of WT have been studied in detail, chemical synaptic transmission and gapjunctions (for an updated discussion see Agnati et al., 2010a). WT allows interneuronal connections in the millisecond range and, in some instances, it even allows a sort of transient syncytial functional organization of the CNS (Agnati et al., 2007b; Hameroff, 2010). However, as mentioned above, astrocyte networks are also connected by WT since gap junctions between chains of astrocytes have been described. Thus, whereas neurons are interconnected through chemical and electrical synapses and their excitable response are basically electrical signals, astrocytes are interconnected via gap junctions and their excitable responses are basically intracellular calcium waves (Oberheim et al., 2009; Pereira and Furlan, 2010). Furthermore, the existence of three-dimensional molecular networks has been proposed, mainly made of proteins and carbohydrates, which can be organized in a large Global Molecular Network (GMN) pervading the intra- and extracellular environment of the central nervous system. The GMN might transmit signals and connect different compartments through interactions between extra- and intracellular molecular networks. Thus, the GMN may be involved in the integrative actions of

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

the brain also in view of its plastic structure, in fact its extracellular part is continuously under the remodeling action of the matrix metalloproteinases (Agnati et al., 2006).

3.1.2.

Volume Transmission (VT)

This is a diffuse mode of signaling process that has similarities with radio broadcasting (for an updated discussion see Agnati et al., 2010a). Thus, diffusible chemical signals can affect entire brain regions. In this context the electromagnetic fields (EMFs) could also be mentioned. The brain's endogenous electromagnetic fields are generated by the fields induced by neuron firing and the fields generated by the movement of ions into and out of cells and within extracellular channels (McFadden, 2002; Pockett, 2000). Pressure waves due to arterial pulses in the brain arteries may also operate as VT signals and/or affect VT signal diffusion and mechano-sensitive ion channels in the plasma membranes of neurons and astrocytes (Agnati et al., 2005a). Thus, this type of signaling can give a beat-to-beat connection between cardiac activity and the functional state of entire brain areas.

3.2. Supra-cellular organization of cortical neurons and astrocytes

In 1938 Lorente de N (1938) suggested that the cortex is organized in cortical cylinders composed of vertical chains of neurons crossing all cortical layers and having specific afferent fibers as their axis. These cylinders represent units of operation (DeFelipe et al., 2002; Lorente de N, 1938). This concept was further developed by Mountcastle's investigations demonstrating that neurons are arranged vertically (or radially in the convoluted cerebrum) in the form of columns spanning the width of the primate somatosensory cortex and responding to a single receptive field in the periphery (Mountcastle et al., 1957; Rakic, 2008). As stated by Rakic (2008), cortical columns can be considered functional units consisting of an array of iterative neuronal groups extending radially across the cortical cellular layers.

Rakic points out that it has been assumed that neurons within a given column are stereotypically interconnected in the vertical dimension, share extrinsic connectivity, and hence act as basic functional units subserving a set of common static and dynamic cortical operations that include not only sensory and motor areas but also association areas subserving the highest cognitive functions. Even though there is no consensus on a single anatomical columnar entity (da Costa and Martin, 2010), the concept of column is still supported by experimental evidence (Hubel and Wiesel, 1977). Thus, when a physiological investigation is carried out through the cortex of primates, ungulates and carnivores in a trajectory perpendicular to its surface, it is generally possible to detect a remarkable constancy in the receptive field properties of the neurons regarding one set of stimulus features (see Fig.3. and Carreira-Perpin and Goodhill, 2002; da Costa and Martin, 2010; Miyawaki et al., 2008). Of special relevance is the evidence of ontogenetic columns and the proven validity of the radial unit hypothesis as the basis for understanding the evolutionary expansion of the cortex (Rakic, 2008). However, some authors claim that the term column does not actually correspond to any single structure within the cortex. Thus, it is impossible to find a canonical microcircuit corresponding to the classical cortical column (Crick and Koch, 2005; Rockland, 2010). The concept could be clarified, on the basis of anatomical findings (Peters and Sethares, 1996), by distinguishing minicolumns (diameter of about 50 m) from macrocolumns (diameter in the range of 300500 m). A further theoretical development of this important distinction can be found in a recent paper by Rinkus (2010) who proposes that minicolumns do have a generic functionality, which only becomes clear when seen in the context of the function of the higher level, subsuming unit, the macrocolumn. Thus, Rinkus proposes that a macrocolumn's functions are to store sparse distributed representations of its inputs and to recognize those inputs; while the generic function of the minicolumn is to enforce macrocolumnar code sparseness.

Fig. 3 Example of the organization of the visual cortex in functional columns (see Carreira-Perpin and Goodhill, 2002). A. Schematic view of an area of visual cortex representing a particular place in the visual field. It appears organized as a pinwheel of columns each reacting to a specific orientation of the edges at that visual field position. B. The surface of the visual cortex appears as a topologically mapped mosaic. The different shades of gray indicate patches that have different orientation preferences. Topologically arranged columns can similarly be identified for other characteristics of the visual field, such as color and spatial frequency (see Miyawaki et al., 2008).

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

Astrocytes can play an important role in the functional organization of the cerebral cortex from specific interactions with single synaptic contacts to modulatory interactions with entire neuronal networks (Oberheim et al., 2009; Pereira and Furlan, 2010). The concept of tripartite synapse has been introduced since the extremity of a protoplasmic astrocyte process wraps the synaptic cleft in most glutamatergic central synapses. It should be noted that astrocytes express membrane receptors to neurotransmitters and can release their own chemical messengers (gliotransmitters). Thus, they establish a cross-talk with both pre- and post-synaptic neurons. It is important to emphasize that VT and WT have a specialized point of contact at the level of tripartite synapses hence it is possible to assess that astrocyte networks regulate neural functions and vice versa (Giaume et al., 2010). Several astrocytes participate in this functional unit (i.e., the tripartite synapse) and are coupled with each other by gap junctions forming a network that can support largescale integrative functions in the brain via a continuous cross-talk with neurons. Thus, neuronastroglial networks do exist (see Agnati and Fuxe, 1984 the concept of complex cellular networks) controlling not simply dynamic glucose delivery (Rouach et al., 2008), but also cognitive information processing (Robertson, 2002). In particular, astrocytes connected via gap junctions represent not only a pathway for direct intercellular exchange of ions, nutrients and signaling molecules (Parpura et al., 2004), but can even be an active computing mechanism. This proposal is supported by findings demonstrating that gap junction channels are regulated by extra- and intracellular signals and allow exchange of

information, so that astrocyte networks can operate as a Turing B-like device since the gap junctions can be set in the pass-mode or in the interrupt mode (Agnati and Fuxe, 2000; Giaume et al., 2010).

4. The concept of functional module and its integrative sentience

Altogether the abovementioned data give some evidence for the existence of elementary processing units functionally located between and hence bridging single cells and system levels (Cutsuridis et al., 2009; Graybiel and Grillner, 2006). They can be called functional modules (FM) and can be defined as micro-circuits in which nerve cells (neurons and astrocytes) are organized into specific patterns to carry out a specific processing activity (see Shepherd, 2011). As discussed above, the functional module concept is basically in agreement with both Rinkus' (Rinkus, 2010) proposal and Buxhoeveden and Casanova's (Buxhoeveden and Casanova, 2002) assumption that cortical morphofunctional units (columns) have no fixed anatomical borders but only functional boundaries since the precise combination of inhibition and excitation creates units of different size and function. Following our group's previous proposal of the brain as a system of nested networks (Agnati and Fuxe, 1984; Agnati et al., 2007a, 2007b, 2008), it may also be surmised that within a FM the information processing could occur at different levels of miniaturization from the cell network level down to the molecular network level (see Fig. 4). It follows that the final, integrated, activity of the FM would emerge from the complex

Fig. 4 Schematic representation of a functional module (FM). As illustrated, in the CNS it is possible to distinguish a horizontal organization (mosaic pattern) and a vertical (hierarchical) organization, following a Russian doll pattern (Agnati and Fuxe, 1984).

10

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

dynamics of this hierarchical system of relations. In this respect, of particular interest for the present discussion are the so called synaptic clusters (SC), in which multiple synapses act cooperatively to modulate their strength (Golding et al., 2002; Shepherd, 1979). Each FM can have one or more SC, often organized around the dendritic spines and partially isolated from the surrounding environment by glial cells (Cutsuridis et al., 2009; Golding et al., 2002). SC were shown to be very plastic entities from both the structural (Holtmaat et al., 2005) and the functional (Welzel et al., 2010) point of view. As a matter of fact, it has been reported that plastic changes induced by Long Term Potentiation (LTP) at one synaptic contact lower the threshold for the induction of LTP at neighboring synapses at a stimulation strength that did not cause any plastic changes under control conditions. The extent of this sensitized plasticity zone spans about 10 m of dendrite and lasts for 10 min and is presumably due to the diffusional spread of the Ca2 + activated small GTPase Ras to neighboring spines (Harvey and Svoboda, 2007). These characteristics allow the SC to act as a sort of intelligent layer between the activity at the cellular level and the integrative functions performed at the molecular level by supramolecular complexes such as those formed at the cell membrane by G protein-coupled receptors, owing to direct receptorreceptor interactions (see Agnati et al., 2010b; Fuxe et al., 2007a; Kenakin et al., 2010 for reviews). To summarize, we could illustrate the general structure of a FM as a hierarchical assemblage of mosaics of different miniaturization, where the term mosaic (see Agnati et al., 2009b) is defined as in figurative art, i.e. as the process of making pictures by inlaying small bits of colored stones (tesserae). Thus, it indicates how from a given set of elements (cells, synapses or molecules in this case) it is possible to achieve different patterns endowed with different emergent properties (e.g. of computational type). It may be surmised that recurrent (Hebb, 1949) WT- and/or VT-based communication processes within the different levels of miniaturization may also help in keeping information on-line and hence allow those integrative processes requiring a certain time interval (see above) to reach the threshold of consciousness. With reference to the abovementioned Cook's proposal (Cook, 2008, see Section 2.3), this schematic representation of the integrative processes occurring within each FM indicates that FM could represent the first integrative level of proto-cognition and proto-consciousness, i.e. a structure able to convert the incoming fragments of sensation into a particular fragment of perception (John, 2002). As pointed out by some authors (see Cutsuridis et al., 2009; Graybiel and Grillner, 2006), examples of such FM include cortical minicolumns, glomeruli in the olfactory systems, networks for the storage and recall of memories in the hippocampus and the prefrontal cortex, and neural microcircuits generating different aspects of motor behavior. The presence of this organization of the human brain can be discussed in the frame of Jacob's (1977) proposal on evolution working not as an engineer but as a tinkerer. Jacob claims that evolution tinkers together contraptions in a natural selection process that acts by adding direction to changes, orienting chance, and slowly and progressively producing more complex structures, new organs, and new species. Thus, novelties come from previously unseen associations of already available material.

In agreement with this proposal, it is suggested that FMs in the human brain can be seen as the result of a tinkering process carried out at different time-scales: a.) long-term scale, by evolution b.) intermediate-term scale, by life-long individual experience c.) short-term scale, by the moment-to-moment external and/or internal inputs impinging on each individual human brain. Evolution by natural selection gave rise to a human brain having, at least in some areas, special FMs thanks to the particular features of VENs, the high density and peculiar biochemical characteristics of synaptic contacts, and the presence of varicose projection astrocytes (see Section 2.2). Moreover, it can be conjectured that special functional rules for recruitment of computational elements and information handling are present in the human brain and even if they are genetically determined, they may differ to some extent from subject to subject and can be made more efficient by appropriate educational training.

5. Fundamental mechanisms for large-scale integration in the brain

The set of FMs corresponding to sensory inputs generates a fractured representation of the world. The issue of perceptual unit needs mechanisms that allow these different sensory components to be gathered into one global image. In other words, to provide the fine texture of consciousness and the global nature of a momentary cognitive instant of experience, a cooperative process is required. In line with the classical Hebb hypothesis on the possible existence of cell assemblies interconnected via reverberating circuits (Hebb, 1949; Wang, 2001), we propose that different FMs can be transiently interconnected to form a higher-order mosaic, representing the further integrative step in the chain leading to cognition and consciousness. It has been suggested (John, 2002) that not only a binding phenomenon, but also a background tone (both exploiting the communication processes of WT and VT) is needed to integrate the information handled by the FMs.

5.1. Formation of the neuronal correlates of consciousness: the background tone and the binding phenomenon
Classical paradigms relate neural activity to controlled sensory stimuli, to the motor responses following stimulation of the motor system or, more generally, to physiological responses in controlled cognitive conditions. Besides these results, several studies have investigated the temporally coherent activity in cortical areas in the absence of overt goal-directed behavior. In humans, this resting state has been suggested not simply to represent noise, but rather to implicate spontaneous and transient processes involved in task-unrelated imagery and thought. The resting state networks not associated with sensory or motor regions, such as the medial prefrontal, parietal and posterior and anterior cingulate cortices, seem to be most engaged when persons are not involved in overt goal-directed behavior. Thus, these networks have been thought to underlie

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

11

certain aspects of conscious introspection, being specific to humans (Ghosh et al., 2008). However, other authors have recently shown that the spatiotemporal patterns of the resting state also exist in anesthetized monkeys, thus demonstrating that they do not reflect a state of consciousness (Shmuel and Leopold, 2008). These dynamic resting states have been considered to manifest the intrinsic characteristics of the underlying brain structure, being useful for keeping the system in a highly competitive state between different sub-networks that are later used during different tasks (Deco et al., 2009). Following this view, it can be suggested that the resting state might represent the background tone on which binding mechanisms (i.e. mechanisms capable of integrating different perceptions, emotions, thoughts and memories in an unified conscious experience) operate. Different binding mechanisms have been proposed to solve the problem of how our brain integrates information distributed among billions of spatially separated neurons to generate the unity of conscious experience. In the late 1980s, Wolf Singer and colleagues (Gray et al., 1989a, 1989b) found a specific, phase-synchronized EEG in cats' visual cortex which was strongly correlated to a particular visual stimulation. The phase synchrony they found in the gamma frequency band (from 30 to 90 Hz) of the EEG became known as coherent 40 Hz. Subsequent studies have shown gamma synchrony in various brain locations correlating with conscious perception. This synchrony has been regarded as the electrophysiological marker of the binding of different unconscious components in a unified conscious percept. Location and distribution of gamma synchrony within the brain can change dynamically, shifting on timescales of hundreds of milliseconds or faster (Hameroff, 2010). However, this mechanism has been criticized as an explanation of the crucial marker of the binding phenomenon. For example, gamma band synchrony occurs in brains of both conscious and unconscious animals in response to visually presented objects (Canales et al., 2007). Large-scale integration, or binding, has then been suggested to involve fluctuations around the background state (Edelman and Tononi, 2000; John, 2002). What leads to a binding of the fragmented information into a coherent process would be the identification (by neuronal coincidence detectors in the cortex) of synchrony within and coherence among brain regions which deviate from the ground state. In this respect, oscillations of local field potentials (in the beta, alpha and theta/delta frequencies) also play an important role in facilitating synchronization and coherence (John, 2002). In this context, a potentially important contribution of astrocytes to the binding process was emphasized by (Pereira and Furlan (2009). According to their theory, the transfer of information patterns embodied in local field potentials to astrocytic calcium waves would facilitate a binding of spatially distributed patterns into unitary conscious episodes. According to the concept of FM illustrated in Section 4, a view could also be proposed, in which the background tone and the binding phenomenon are considered as different aspects of the collective dynamics of FMs. In fact, it can be surmised that the background tone simply reflect reverberating activity playing the leading role of continuously assembling

pools of FMs. In the absence of a salient internal and/or external input this process results as a spontaneous chaotic activity (recorded as a noise) present in many brain regions. Under internal and/or external inputs triggering a binding phenomenon, a self-organization occurs and this basal state moves towards an attractor, eventually leading to a dynamic synchronization and hence to the functional assemblage of several FMs into specific high-order mosaics. A comment by Werner (2007) may be in line with this hypothesis: On the horizon I note promising new actors to count with in future: I draw attention to computational simulations of radically novel features of neural microcircuits which function more like liquids responding to perturbations with ripples of waves, rather than like digital gates () If proven real, such microcircuits would adopt a system dynamics at the boundary region of ordered and chaotic behavior. They would, thus, belong to the class of natural systems operating at the edge of chaos, which are known for their capacity for critical self-organization. Summing up, the existence of global synchrony in the brain indicates the operation of a tuning mechanism accounting for the coordination of local circuits and orienting some FMs to form transient higher-order mosaics, which may represent the NCCs. The assemblage is not the result of any conscious process, but on the contrary, is a largely unconscious process that depends on genetically coded as well as learned interaction rules, which integrate incoming inputs with short and longterm memories stored in the FMs. In a recent interesting study by Buzski (2010) possible mechanisms allowing the identification and organization of cell assemblies are extensively reviewed and discussed. FMs are recruited according to different spatial patterns (locations of the FMs with respect to each other) and temporal patterns (time sequences of their activation). These two patterns also regulate the mosaic networks of different miniaturization inside each FM. An adaptation is, therefore, possible by shaping and activating the mosaics to fit the specific task to fulfill.

5.2. The neuroanatomical bases of large-scale brain integration

A crucial aspect is the characterization of the neuronal systems selecting FMs and linking them together into the abovementioned higher-order mosaics representing the NCCs. Two of these, namely the thalamo-cortical and the brainstemsubcortical/cortical interconnections will be briefly described, since they are widely acknowledged to play a key role in the formation of conscious events. Some comment will be also deserved to the claustrum and to its proposed (Crick and Koch, 2005) special integrative role.

5.2.1.

Thalamo-cortical interconnections

Thalamo-cortical reverberating circuits are a highly relevant topic for consciousness processes. Several lines of clinical experience suggest that if we lack the thalamus, the cortex is useless, leaving the patient in a state close to total coma. Thus, as emphasized by Llinas and Ribary (2001), consciousness needs a continuous dialog between the thalamus and the cortex to arise. As Llinas et al. (1998) pointed out, the thalamus represents a sort of hub from which any site in the cortex can

12

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

communicate with other sites. The corticothalamic pathway, therefore, can selectively mediate coherence and synchronicity of activity between selected groups of interconnected cortical and thalamic neurons during particular functional states. A second organizing principle, however, may be equally important. It is based on the temporal rather than spatial relationships among groups of neurons and the thalamo-cortical iterative recurrent activity certainly plays a key role in such a process. In fact, as proposed by Llinas et al. (1998) two complementary loops between the thalamus and the cerebral cortex work in conjunction to subserve temporal binding: - a specific system formed by sensory or motor nuclei projecting to layer IV of the cortex. It produces cortical oscillations by direct activation and feed-forward inhibition via inhibitory interneurons. These oscillations re-enter the reticular nucleus of the thalamus via layer-VI pyramidal cells; - a non-specific system, in which intralaminar nonspecific thalamic nuclei project to cortical layers I and V and to the reticular nucleus. Layer-V pyramidal cells return oscillations to the reticular nucleus and intralaminar nuclei. It is apparent from literature that neither of these circuits alone can generate cognition: damage of the non-specific system leads to deep disturbances of consciousness, while damage of the specific system produces loss of some particular cognitive modality. These observations led to the hypothesis (Llinas and Ribary, 2001) that the specific system would provide the content relating to the external world, and the non-specific system would give rise to the temporal conjunction, or the context (on the basis of a more interoceptive background concerned with alertness). Together they would generate a single cognitive experience. An interesting structural feature of the interconnections between cortex and thalamus is that a very large percentage of this connectivity is recurrent, and that much of its activity is related to such intrinsic connectivity not necessarily related to the immediacy of sensory input. In particular, the thalamic input from the cortex is larger than that from the peripheral sensory system (Llinas et al., 1998). It is, therefore, reasonable to assume that the brain is essentially a closed system capable of self-generated activity based on the intrinsic electrical properties of its component neurons and their connectivity. In this respect Llinas et al. (1998) proposed an analogy with the spinal cord based on the pioneering work of Brown (1914) on locomotion. Brown demonstrated that sensory inputs are mostly modifiers of the intrinsic activity of the spinal cord since locomotion was still present after bilateral dorsal root deafferentation, hence after the total removal of sensory inputs. On that basis, Brown proposed that the complex motor output required for locomotion is a property of the spontaneous activity of the neuronal circuits in the spinal cord and brainstem. Similarly, Llinas et al. (1998) suggested that consciousness might be the result of the intrinsic activity of brain circuits. As such, consciousness can be thought of as an oneiric-like internal functional state modulated, rather than generated, by the senses. This proposal is in line with the classical view that perception is a model in the brain (Blakemore, 1976) and such a model is built up with inborn circuits and

hence is present at birth, and is fine-tuned later on during normal maturation (see also Agnati et al., 2007c). In other words, the CNS is a reality-emulating system in which only some parameters of such reality are delineated by the senses. Such a view of the brain as a closed system capable of an autonomous creation of reality even in the absence of sensory inputs, resembles the more recent proposal by Hobson (2009) that during rapid eye movement (REM) sleep (i.e., in the absence of sensory inputs) the brain may create a virtual reality model of the world. Such a model could be of functional use in the development and maintenance of waking consciousness. Available experimental data comparing awake and REM sleep state (Llinas and Par, 1991) provide support to this idea. They suggest that we do not perceive the external world during REM sleep because the intrinsic activity of the corticothalamic systems does not place sensory input in the context of the functional state being generated by the brain. In other words, the dreaming condition appears as a state of hyperattentiveness to intrinsic activity in which sensory input cannot access the machinery generating conscious experience. In agreement with the crucial role played by corticothalamic interconnections in consciousness formation are also the neuropathological data on familial fatal insomnia (FFI) (Montagna et al., 2003). Autopsy verification in FFI patients disclosed atrophy of the mediodorsal and anterior ventral thalamic nuclei. In these patients worsening of sleep and autonomic disturbances are associated with the onset of peculiar oneiric behaviors whereby patients, especially if left to themselves, fall into a hallucinatory state displaying motor gestures related to the content of their dreams.

5.2.2.

The cortical and subcortical projections from the brainstem

Moruzzi and Magoun (1949) obtained evidence of how distinct nerve cell populations in the midbrain and pons could produce a global activation of the cerebral cortex. Lesions to these so-called reticular activating systems resulted in a sleep-like state. The discovery of monosynaptic DA, NA and 5-HT nerve cell projections from the pons and midbrain to the cerebral cortex and subcortical regions provided the structural and neurochemical correlate to these pioneering physiological observations (Andn et al., 1964, 1966; Chalmers, 2000; Dahlstrm and Fuxe, 1964; Fuxe, 1965; Fuxe et al., 2007b, 2010; Thierry et al., 1973). In particular, it has been shown that the locus coeruleus ascending NA projections from the pons to the cerebral cortex play a significant role in tonic arousal (Jouvet, 1972; Lidbrink and Fuxe, 1973). Furthermore, the ascending 5-HT projections from the mesencephalic raphe nuclei to cortical and subcortical regions play a role, among others, in the maintenance of slowwave sleep and in preventing melancholy (Fuxe et al., 2007b; Jouvet, 1972; Kiianmaa and Fuxe, 1977). The meso-limbiccortical DA neurons from the ventral tegmental area (Andn et al., 1966; Dahlstrm and Fuxe, 1964; Thierry et al., 1973) innervate the subcortical limbic forebrain, especially the nucleus accumbens core and shell, and many cortical regions namely the limbic cortex and prefrontal cortex. They play a major role in reward and reward prediction, attention, working memory and modulating the transfer of emotional information from the subcortical limbic forebrain to the cerebral cortex, especially the prefrontal cortex (Fuxe et al., 2007b).

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

13

The impact of these subcortical and cortical monoamine projections for the performance of the FMs in the cerebral cortex and, hence, for higher consciousness also becomes clear from the fact that disturbances in these ascending monoamine systems contribute to the development of schizophrenia, attention deficit and hyperactivity disorders (Fuxe et al., 2007b). Furthermore, hallucinogens of the indolalkylamine type like d-LSD produce their hallucinations by activating distinct subtypes of 5-HT receptors, the 5-HT2A subtype located in the cerebral cortex (Fuxe et al., 2009). These ascending monosynaptic monoamine projections from the pons and midbrain innervating the cerebral cortex and subcortical regions serve as important examples of the substantial impact the lower brainstem afferents exert on the function of the cerebral cortex directly or indirectly via innervations of subcortical regions like the striatum, and thus on consciousness. Summing up these data on the functional meaning of the monoamine ascending systems, it can be stated that the integration of the various functional cortical modules can no longer develop well-tuned high brain function without their global modulation and fine-tuning by cortical and subcortical DA, NA and 5-HT nerve terminal networks acting mainly via VT (see below and Agnati et al., 2007b; Fuxe et al., 2010). The cortical cholinergic nerve terminal systems arising from the basal forebrain probably have a similar role in consciousness and can operate via both VT and WT (Descarries, 1998; Perry et al., 1999). Thus, while it is clear that the evolutionary recent six-layered cerebral cortex plays a fundamental role in the higher level consciousness, this evolutionarily recent structure remains under the strong control of phylogenetically ancient brainstem systems that are, in fact, essential for higher consciousness. As an intriguing example of such, it has been proposed that a primary consciousness may exist in children born without a cerebral cortex. Also, goal-directed behavior has been observed in mammals after experimental decortication (Merker, 2007).

dorsal extent, thereby synchronizing different perceptual, cognitive and motor modalities. Furthermore, they pointed out that this postulated intra-claustrum integrative action differs from that of the thalamus, which also has widespread reciprocal relations with most cortical regions but does not possess any obvious mechanism to directly link its various constitutive nuclei.

6. From NCCs to conscious episodes: the need for metaphors

It is not yet possible to move from the NCCs to conscious episodes introspectively obtained and some philosophers even claim that some aspects of consciousness, such as subjectivity, might be inherently inexplicable. As pointed out by Baars (1998), however, we cannot know today whether or not we will eventually understand problems like that, although they might become clearer as more plausible hypotheses are tested. In this respect, a useful thinking tool is to suggest metaphors giving hints of some aspects of the NCCs-to-consciousness journey. A metaphor can be defined as the application of a word or phrase to an object or concept it does not literally denote, suggesting comparison to that object or concept (Webster's College Dictionary, 1995). How far can we go along this journey by means of metaphors? It is worth mentioning Werner's (2007) advice on the use of metaphors in neuroscience of cognition and consciousness. Werner discusses the strict limits within which the use of metaphors can illuminate a target domain in cognitive neuroscience. Thus, Werner delineates the risk of metaphors since they tend to carry with them the style of reasoning of the source domain which may be (and often is) quite inappropriate for the target; thus entailing the risk of tacitly contaminating the target with erroneous styles of reasoning (Werner, 2007). Criteria for productive metaphors, however, have been defined by Baars (1998), who states: productive metaphors should help organize existing evidence, yield testable hypotheses and suggest conceptual clarifications (Baars, 1998). Any cognitive metaphor should, therefore, be considered from Werner's and Baar's points of view. Thus, metaphors should be used as instruments to grasp some aspects of the still obscure physical processes relating NCCs to consciousness, and not as a scientific description of consciousness itself. This is the case, thus far, to give any account of internal, conscious experiences such as the so-called qualia. No scientific description can be given. However, the way in which the neurobiological pre-conscient issues are integrated reaching subjective experiences of the self can benefit from a metaphoric approach. Many metaphors have been proposed to illustrate certain features of consciousness. They basically reflect a common theme that can be labeled the theater metaphor (Baars, 1988, 1997; Crick, 1984; Dennett and Kinsbourne, 1992) and imply both convergence of input and divergent dissemination of the integrated content. We would like to introduce a new one, the Hall of Mirrors, to explain reverberating activity between and within the FMs. Our metaphor suggests that the mosaic of FMs, i.e., the NCCs, can be viewed as a transient assembled Hall of Mirrors (Fig. 5) where each mirror reflects the images of other mirrors.

5.2.3.

The claustrum

As discussed above, it is widely accepted that there is no single cortical area where it all comes together to produce the conscious content. The elements of a coalition implicated in the NCCs are widely distributed over both the back and the front of the brain. Thus, effectively, they bind by interacting in a widespread manner. However, claustrum, that according to many authors should be considered the seventh layer of the cortex in the insular region (Tanne-Gariepy et al., 2002), has been proposed to play a special integrative role in view of its vast interconnectivity with most allocortical and neocortical regions (Crick and Koch, 2005). Important claustrum interconnections have been described with the frontal lobe including the cingulated cortex, and the temporal, parietal and entorhinal cortex. As mentioned by Crick and Koch (2005), the claustrum also projects to the hippocampus, amygdala and caudate nucleus. It follows that it could have the possibility to bind disparate events into a single percept experienced at one point in time (Crick and Koch, 2005). Based on these data, Crick and Koch (2005) suggested that it might be important to investigate the functional role of the claustrum, namely whether it contains specialized mechanisms capable of integrating the information that travels widely within its anteriorposterior and ventral

14

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

Fig. 5 Consciousness as a virtual space and a virtual personage created by the existence of multiple interacting mirrors. The metaphor of mirrors is simply a heuristic hypothesis to depict the metaphor of the internal theater in neurobiological terms. A temptative clarification of the term mirror is given in Section 6.1 for the moment let us discuss their fundamental feature, namely the reflection process. The reflection process of a mirror is not achieved passively, but by filtering and enriching the images with the memories that it may store, and according to the peculiar integrative capabilities with which the reflection is endowed. The final result of these multiple reciprocal reflections among mirrors is the creation of a virtual space (the set of the internal theater) and of a virtual personage (the self) who lives within this theater. These are dynamic operations that allow the stage of the theater and the attitudes and feelings of the personage to be continuously rebuilt. Of particular interest can be the discussion of the suggestions proposed by the metaphor from the standpoint of the abovementioned criteria indicated by Baars (1998). Some hints to give a morpho-functional correlate to the Mirror expression and hence to the Hall of Mirrors metaphor could be deduced from recent studies on the parcellation of the cortex, aimed to characterize its functionalanatomical organization (Knsche and Tittgemeyer, 2011). Parcellation leads to the subdivision of the cortical surface into compact areas, which are internally relatively homogeneous and distinct from one another, with respect to the considered structural and/or functional criteria. It could be surmised that sometimes a Mirror is a mosaic of FMs within one of these relatively homogenous and compact areas. As a matter of fact, the importance of the cortex parcellation has been proved also very useful for studying the organizational principles of the brain and its ontogenetic and phylogenetic development (Bystron et al., 2008; Rakic, 2009). Given that brain regions frequently maintain characteristic connectivity profiles and the functional repertoire of a cortical area is closely related to its anatomical connections, longrange connectivity may be used not only to define segregated cortical areas (Knsche and Tittgemeyer, 2011) but also the Hall of Mirrors associated to a certain function. This aspect could be also discussed in the frame of Mesulam's proposal (Mesulam, 2005) that many neurological and psychiatric disorders are likely to be associated with altered anatomical connectivity. Thus, it could be possible to investigate neurological and psychiatric disorders at least at three different integrative levels: Functional Module, Mirror and Hall of Mirrors. As discussed in the previous sections, this view is consistent with existing experimental evidence of peculiar columnar mappings of the sensory receptive fields onto the somatosensory (Mountcastle et al., 1957; Rakic, 2008; Woolsey and Van der Loos, 1970), visual (Hubel and Wiesel, 1977), auditory (Hromdka and Zador, 2009), piriform (Gottfried, 2010) and primary taste (Chen et al., 2011) brain cortex. Similar forms of morphofunctional organization also exist outside the cortical areas, such as in the brainstem (Erzurumlu et al., 2010), the superior colliculus and the cerebellum (Arenz et al., 2009). As proposed by Llinas et al. (1998) the activity involving resonant columns could be the basis for cognitive events. According to

6.1.

Hall of Mirrors: existing evidence

The proposed metaphor shares with all the metaphors illustrating the integrated brain activity as an internal theater the view that the overall function of consciousness is to provide very widespread access to unconscious brain regions (the audience of the theater). As demonstrated by a number of neuropsychological studies (Buchwald, 1974; Kosslyn, 1994; Shiffrin et al., 1981) such access is needed for coordination and control (see Baars, 1998 for a thoughtful discussion). According to our proposal in a top-down summing up, a rather stable organization of a fundamental set of mirrors (leading to the self) variably interconnected with a plastic set of mirrors (building up a changeable Hall of Mirrors, i.e. the stage of the internal theater) provides the tools for such a coordination. Consciousness, therefore, appears to emerge from this dynamic mirror effect, as well as from the integrative actions inside of the FMs involved. Thus, the present metaphor suggests that specific morphofunctional features of cerebral cortex structures can integrate, display, disseminate (reflect) contents to other brain structures and receive feedback from them.

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

15

our model, these resonant columns are made up of FMs at cortical level. FMs are microcircuits organized as nested mosaics of different miniaturization (see Fig. 4). The columns (the Mirrors) dynamically assemble to form a higher-order mosaic where reverberating activity occurs along some of the available wiring interconnections (such as the corticothalamic pathways), leading to the formation of a NCC (Hall of Mirrors). This view is in agreement with Edelman and Tononi's (2000) proposal on the existence of a dynamical core formed by shifting assemblies of spiking neurons throughout the forebrain that are stabilized using massive re-entrant feedback connections (Edelman and Tononi, 2000). As also pointed out by Crick and Koch (2003), competitions among rival cell assemblies, i.e., potentially differently formed mosaics, occur as a dynamic process (Crick and Koch, 2003). The winning coalition is the one that makes the largest contribution to what we are conscious of. An interesting approach to investigate our hypothesis could be based on the following question: what is it that makes the human Hall of Mirrors (i.e., the mosaics of FMs) so peculiar when compared to other mammalian species, despite quite similar types of macro-anatomical cortical regions? As mentioned in Section 2, the human brain possesses not only special types of neurons (VEN) and astrocytes (varicose projection astrocytes) but also a higher density of synaptic contacts per neuron and a higher astrocyte/neuron ratio. Thus, data on neurons suggest more reverberating circuits, while the data on astrocytes seem to indicate more extended astrocyte networks and hence more complex neuronastroglial interactions. Altogether they open the possibility for the formation of more complex FMs. Furthermore, it may also be surmised that the hierarchical dimension of the integrative processes occurring within human FMs could be more complex. In this respect it would be of particular interest to investigate whether the human brain has special protein mosaics, e.g., special receptor mosaics (Agnati et al., 2005b, 2007d).

6.2.

Hall of Mirrors: testable hypotheses

The Hall of Mirrors metaphor yield testable hypotheses that will be the focus of the present Section. A common view in the cognitive neuroscience is that brain areas are highly selective and exhibit considerable specialization, each responding to a set of inputs and contributing primarily to a single cognitive domain. The here proposed metaphor would suggest a different view in which the same basic modules (the FMs) can be variably associated to realize a large spectrum of NCCs, leading to the testable hypothesis that single brain regions (even fairly small regions) could contribute to multiple cognitiveemotional tasks. It is noteworthy that over the past years increasing evidence emerged pointing to this direction. Examples are provided by studies on the Broca's area (see Poldrack, 2006), showing that current evidence for the notion that Broca's area is a language region is fairly weak, since it was more frequently activated by non-language tasks than by language-related ones. Similarly, a meta-analysis by Anderson et al. (2010) demonstrated that most regions of the brain appear to be activated by multiple tasks across diverse task categories. The results reported in that study suggested that the brain achieves its variety of functions

by putting the same regions together in different patterns of functional cooperation. In this context, a cognitive task of particular interest is the so-called mirror neuronal episodes. According to current neurobiological research, a mirror neuron is a neuron that fires both when an animal acts and when the animal observes the same action performed by another subject (Rizzolatti and Craighero, 2004). Thus, the neuron mirrors the behavior of the other, as though the observer were acting directly. Brain activity consistent with that of mirror neurons has been found in the premotor cortex and the inferior parietal cortex in humans (Ferrari et al., 2006, 2009; Rizzolatti et al., 2009). Such neurons have been observed in primates, and are believed to occur in other species including birds. Functional magnetic resonance imaging (fMRI) studies in humans suggest that a much wider network of brain areas shows mirror properties in humans than was previously thought. These additional areas include the somato-sensory cortex and are thought to make the observer feel what it feels like to move in the observed way (Gazzola and Keysers, 2009). fMRI experiments suggest that rather than mirror neurons the human brain areas have mirror neuron systems (Iacoboni et al., 1999). Based on the Hall of Mirrors metaphor, we suggest the testable hypothesis that mirror systems are an epiphenomenon of a more fundamental feature of the functional organization of the brain, i.e., the existence of computational units (mosaics of FMs) acting like mirrors. These mosaics mirror a suitable input, processing it according to their intrinsic functional characteristics, namely connectivity and more generally to memory stores, hence their history. The first set of mirrors is the FM mosaics reflecting the relevant cues of the environment in an analogic code that is the most useful code for the task the brain should tackle in that instancea motor code in the case of a movement, an emotional code in the case of a feeling. This assumption is not only in agreement with the data on mirror systems but also with the imagery processes: to produce an imagery movement the brain activates, at least in part, networks in the motor area in order to produce an imagery movement. Motor Imagery (MI) has been shown to involve the conscious internal representation of movement, without overt motor performance (Decety and Grezes, 1999; Fleming et al., 2010). Similarities in brain activation between MI and actual movement have also been demonstrated by fMRI (Gerardin et al., 2000). In particular, fMRI studies have shown that during imagined and executed finger movements, common areas of brain activation can be observed in the premotor cortex, supplementary motor area and parietal cortex, with activation peaks slightly more rostral in frontal areas and slightly more superior and caudal in parietal areas during MI compared with movement execution (Gerardin et al., 2000). Both the right and left parietal cortices show greater blood oxygen level-dependent signals during imagery than during motor execution. In addition, the greater activation of the superior parietal cortex during imagery than during preparation for movement, indicates that MI is more than simply readiness to move (Stephan et al., 1995). The mirror episodes stress the importance of reverberating phenomena in the brain. However, emphasis should be placed on the difference between the passive process of reflection of a beam of light by a mirror and the reverberation of neuronal activity between two FM mosaics, which transform

16

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

the incoming information not only according to their structural organization but also according to their history. The hall of mirror metaphor also suggests that the same logic can be at the basis of a particularly important task such as self-awareness. In other words, the continuous building of a virtual personage (the self) living in the internal theater would be the result of the mirror effect within some set of FMs. An interesting possibility to test such an hypothesis may come from the fundamental study by Craig (2009) where he proposes that the anterior insular cortex contains interoreceptive representations of all subjective feelings from the body and likely also emotional awareness, therefore being able to play a fundamental role for self-awareness. Thus, FMs that are key for high-level sentience may be located in the insula and anterior cingulate cortex. Experiments on self-recognition seemed to confirm the important role of these regions in the generation of an abstract representation of oneself. In other words, these regions appear to be deeply involved in the creation of the virtual sentient personage within the internal theater. As mentioned above, it is noteworthy that these two regions exhibit an extraordinary concentration of VENs, which are present only in few species (namely those capable of passing the mirror test). These neurons could be part of the circuitry supporting human social networks. In particular, Allman et al. (2005) proposed that the VENs relay an output from the fronto-insular and anterior cingulate cortex to the parts of the frontal and temporal cortex associated with theory-of-mind, where fast intuitions are merged with slower, deliberative judgments (Allman et al., 2005). In agreement with such a view, it has been shown that the loss of emotional awareness and self-consciousness in patients with fronto-temporal dementia correlates with the degeneration of VENs (Seeley et al., 2006).

6.3.

Hall of Mirrors: suggested conceptual issues

As described in the foregoing Sections, the basic idea concerning the morphofunctional organization of the brain suggested by the Hall of Mirrors metaphor is a view of the cerebral cortex as composed of FMs, i.e. microcircuits structured as a hierarchical series of networks (computational mosaics) of different miniaturization. These basic units are able to perform a firstlevel integrative function by allowing the conversion of incoming fragments of sensation into particular fragments of perception. They, in turn, can be dynamically linked to form mosaics of increasing order, leading to the NCCs. A block diagram (see Fig. 6) could indicate aspects of the main system components involved in the present view A piano keyboard could also serve as a simple analogy to illustrate the FM articulation in the cerebral cortex. In fact, FMs can be represented by the keys of the piano activated by interoceptive and/or external stimuli but also by imagery. The complex mechanism behind a piano key can be thought as the complex hierarchical articulation of a FM. With the given set of piano keys a great many different assemblies of sounds can be produced, since the keys can be touched according to different spatial and temporal patterns. Each specific melody, in other words, depends on the set of keys used (spatial pattern) and the temporal sequence in which these keys are activated (temporal pattern). Similarly, the cognitive value of the mosaic of FMs depends on the selection of the activated FMs, and hence on their spatial and temporal pattern, and on the possibility that the partial and separated cognitive elements are properly integrated. It follows that, according to the proposed Hall of Mirrors metaphor, a key concept in the formation of the internal theater is that of reuse of the same basic circuits to build the many different patterns of activity that lead to conscious episodes. In this respect, the present neurobiological view

Fig. 6 Possible merging of different available hypotheses in a unique schematic integrated view: (1) Coalitions of shifting assemblies of brain cells stabilized by re-entrant feed-backs (Edelman and Tononi, 2000) represent the NCCs. According to the hypothesis here presented the basic element (tessera) of the assembly (mosaic) is a FM characterized by a Russian doll structure; (2) Special FMs could be found at the level of the insula and the anterior cingulate cortex (Craig, 2009) giving the body an emotional awareness; (3) as proposed by Llinas (Llinas et al., 1998) thalamo-cortical interconnections should represent the main system selecting and binding the FMs to form the NCCs; (4) the claustrum could contribute to the process by acting as a conductor giving the proper emphasis to each FM (Crick and Koch, 2003).

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

17

shows consistency with some general cognitive theories that appeared in the last 5 years (Anderson, 2010; Dehaene, 2009; Gallese, 2008; Hurley, 2008) suggesting that low-level neural circuits are used and reused for various purposes in different cognitive and task domains. The massive redeployment hypothesis (Anderson, 2007a, 2007b), for instance, explains the observed patterns of scattered regions corresponding to different cognitive tasks (such as language, visual perception, attention) with the suggestion that local circuits may have low-level computational workings that can be put to many different higher-level cognitive uses. As pointed out by Anderson (2010) this represents a novel concept, something about the brain that we are just now beginning to notice. It offers a distinct perspective on several general topics (such as the evolution and development of the brain, the degree of modularity in brain organization and the degree of localization of cognitive functions) and could also have some practical implications in terms of rehabilitative medicine.

only to imagine skies above us but even non-existing skies (Pessoa, 1982). As a final remark, the provided considerations clearly delineate the restricted use of the Hall of Mirrors metaphor here proposed. It is suggested only as a pictorial or literary image of the mechanisms building the set of our internal theater. In the meantime, we believe that the negative analogy of such a metaphor can be thought of as a useful way of illustrating the complexity of the neurobiological problems involved and the neurobiological investigations that should be carried out, namely regarding: The rules for recruitment of the mosaics of different miniaturization forming single FMs and those underlying the logical operations carried out at each level. The processes leading to the recruitment of the FMs forming the higher-order mosaics and the NCCs. The functional organization of the FMs: boundaries and communication processes inside each FM and among FMs. In particular, the different functional meaning of the background tone versus the binding phenomenon.

7.

Final comments

A basic assumption of the present paper is the building up of mosaic networks made of elements (the FMs) that are themselves hierarchically organized (like a Russian doll), i.e. encasing mosaic networks each made of other mosaic networks of a higher miniaturization. The relevance of this vertical morphofunctional organization for the integrative actions of the brain can be better appreciated when observing that such an arrangement allows an enormous number of possible configurations for each FM, providing it with an extraordinary potential capability to process and store information. The integrative power of the brain is further enhanced if the above discussed mechanism of neural reuse (Anderson, 2010) is in operation. In fact, starting from a given set of FMs, it allows the realization of a wide spectrum of different assemblies, leading to the emergence of NCCs. As mentioned above, parcellation leads to the subdivision of the cortical surface into compact areas, which internally contain relatively similar FMs. It is suggested that these FMs can independently participate in the formation of different mosaics. Thus, it can also be surmised that, by involving different mosaics of FMs, the same area can be simultaneously redeploymented for different tasks. As pointed out in Section 5, the formation of these conscious episodes could result from the interplay of two basic components of the cerebral activity: the first one internal and withdrawn, the second one, being responsible for sensorimotor actions, open to the external world. As suggested by Llinas, in principle one can see how the intrinsic activity of neurons, which reflect a closed reference system, may be the stage on which our image of the external world is ultimately generated (Llinas, 1988). In conclusion, everyone lives in his own internal theater as assessed by Chamfort: C'est l proprement l'homme; l se borne son empire. Tout le reste lui est tranger (see Auguis, 18241825). The internal theater, however, cannot be compared to a prison since, as stated by Pessoa, there is an extraordinary opening in its thick walls: our imagination which allows us not

Acknowledgments
This work was supported by grants from IRCCS San Camillo, Italy. In addition, C. Cela-Conde's contribution was made possible by project research grant HUM2007-64086/FISO awarded by the Direccin General de Investigacin del Ministerio de Educacin y Ciencia (Spain).

REFERENCES

Agnati, L.F., Fuxe, K., 1984. New concepts on the structure of the neuronal networks: the miniaturization and hierarchical organization of the central nervous system. Biosci. Rep. 4, 9398. Agnati, L.F., Fuxe, K., 2000. Volume Transmission as a key feature of information handling in the central nervous system possible new interpretative value of the Turing's B-type machine. Prog. Brain Res. 125, 319. Agnati, L.F., Fuxe, K., Zoli, M., Rondanini, C., Ogren, S.O., 1982. New vistas on synaptic plasticity: the receptor mosaic hypothesis of the engram. Med. Biol. 60, 183190. Agnati, L.F., Zoli, M., Merlo Pich, E., Benfenati, F., Fuxe, K., 1990. Aspects of neural plasticity in the central nervous system. VII. Theoretical aspects of brain communication and computation. Neurochem. Int. 16, 479500. Agnati, L.F., Cortelli, P., Pettersson, R., Fuxe, K., 1995. The concept of trophic units in the central nervous system. Prog. Neurobiol. 46, 561574. Agnati, L.F., Fuxe, K., Ferr, S., 2005a. How receptor mosaics decode transmitter signals. Possible relevance of cooperativity. Trends Biochem. Sci. 30, 88193. Agnati, L.F., Genedani, S., Lenzi, P.L., Leo, G., Mora, F., Ferr, S., Fuxe, K., 2005b. Energy gradients for the homeostatic control of brain ECF composition and for VT signal migration: introduction of the tide hypothesis. J. Neural Transm. 112, 4563. Agnati, L.F., Zunarelli, E., Genedani, S., Fuxe, K., 2006. On the existence of a global molecular network enmeshing the whole central nervous system: physiological and pathological implications. Curr. Protein Pept. Sci. 7, 315.

18

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

Agnati, L.F., Agnati, A., Mora, F., Fuxe, K., 2007a. Does the human brain have unique genetically determined networks coding logical and ethical principles and aesthetics? From Plato to novel mirror networks. Brain Res. Rev. 55, 6877. Agnati, L.F., Genedani, S., Leo, G., Rivera, A., Guidolin, D., Fuxe, K., 2007b. One century of progress in neuroscience founded on Golgi and Cajal's outstanding experimental and theoretical contributions. Brain Res. Rev. 55, 167189. Agnati, L.F., Guidolin, D., Fuxe, K., 2007c. The brain as a system of nested but partially overlapping networks. Heuristic relevance of the model for brain physiology and pathology. J. Neural Transm. 114, 319. Agnati, L.F., Guidolin, D., Leo, G., Fuxe, K., 2007d. A boolean network modelling of receptor mosaics relevance of topology and cooperativity. J. Neural Transm. 114, 7792. Agnati, L.F., Guidolin, D., Carone, C., Dam, M., Genedani, S., Fuxe, K., 2008. Understanding neuronal molecular networks builds on neuronal cellular network architecture. Brain Res. Rev. 58, 379399. Agnati, L.F., Baluka, F., Barlow, P.W., Guidolin, D., 2009a. Mosaic, self-similarity logic, and biological attraction principles: three explanatory instruments in biology. Commun. Integr. Biol. 2, 552563. Agnati, L.F., Fuxe, K., Baluska, F., Guidolin, D., 2009b. Implications of the Energide concept for communication and information handling in the central nervous system. J. Neural Transm. 116, 10371052. Agnati, L.F., Guidolin, D., Guescini, M., Genedani, S., Fuxe, K., 2010a. Understanding wiring and Volume Transmission. Brain Res. Rev. 64, 137159. Agnati, L.F., Guidolin, D., Leo, G., Carone, C., Genedani, S., Fuxe, K., 2010b. Receptor-receptor interactions: a novel concept in brain integration. Prog. Neurobiol. 90, 157175. Allman, J.M., Watson, K.K., Tetreault, N.A., Hakeem, A.Y., 2005. Intuition and autism: a possible role for von economo neurons. Trends Cogn. Sci. 9, 367373. Alonso-Nanclares, L., Gonzalez-Soriano, J., Rodriguez, J.R., DeFelipe, J., 2008. Gender differences in human cortical synaptic density. Proc. Natl. Acad. Sci. U. S. A. 105, 1461514619. Andn, N.E., Carlsson, A., Dahlstrm, A., Fuxe, K., Hillarp, N.A., Larsson, K., 1964. Demonstration and mapping out of nigro-neostriatal dopamine neurons. Life Sci. 3, 523530. Andn, N.E., Dahlstrm, A., Fuxe, K., Larsson, K., Olson, L., Ungerstedt, U., 1966. Ascending monoamine neurons to the telencephalon and diencephalon. Acta Physiol. Scand. 67, 313326. Anderson, M.L., 2007a. Evolution of cognitive function via redeployment of brain areas. Neuroscientist 13, 1321. Anderson, M.L., 2007b. Massive redeployment, exaptation, and the functional integration of cognitive operations. Synthese 159, 329345. Anderson, M.L., 2010. Neural reuse: a fundamental organizational principle of the brain. Behav. Brain Sci. 33, 245313. Anderson, M.L., Brumbaugh, J., Suben, A., 2010. Investigating functional cooperation in the human brain using simple graph-theoretic methods. In: Chaovalitwongse, A., P.M. Pardalos, V., Xanthopoulos, P. (Eds.), Computational Neuroscience. Springer, New York, pp. 3142. Aoki, C., Venkatesan, C., Go, C.-G., Forman, R., Kurose, H., 1998. Cellular and subcellular sites for noradrenergic action in the monkey dorsolateral prefrontal cortex as revealed by the immunocytochemical localization of noradrenergic receptors and axons. Cereb. Cortex 8, 269277. Arenz, A., Bracey, E.F., Margrie, T.W., 2009. Sensory representations in cerebellar granule cells. Curr. Opin. Neurobiol. 19, 445451. Auguis, P.R., 18241825. Chamfort, Oeuvres compltes, tome II. Chamerot, Paris.

Baars, B.J., 1988. A Cognitive Theory of Consciousness. Cambridge University Press, Cambridge. Baars, B.J., 1997. In the Theater of Consciousness. The Workspace of the Mind. Oxford University Press, Oxford. Baars, B.J., 1998. Metaphors of consciousness and attention in the brain. Trends Neurosci. 21, 5862. Baars, B.J., Ramsoy, T.Z., Laureys, S., 2003. Brain, conscious experience and the observing self. Trends Neurosci. 26, 671675. Badlangana, N.L., Bhagwandin, A., Fuxe, K., Manger, P.R., 2007. Observations on the giraffe central nervous system related to the corticospinal tract, motor cortex and spinal cord: what difference does a long neck make? Neuroscience 148, 522534. Baluka, F., Volkmann, D., Barlow, P.W., 2006. Cellcell channels and their implication for cell theory. In: Baluka, F., Volkmann, D., Barlow, P.W. (Eds.), CellCell Channels. Landes Bioscience, Georgetown, pp. 118. Benavides-Piccione, R., Ballesteros-Yez, I., DeFelipe, J., Yuste, R., 2002. Cortical area and species differences in dendritic spine morphology. J. Neurocytol. 31, 337346. Blakemore, C., 1976. Mechanics of the Mind. Cambridge University Press, Cambridge. Brown, G., 1914. The intrinsic factors in the act of progression in the mammal. Proc. R. Soc. Lond. 84, 308319. Buchwald, J.S., 1974. Operant conditioning of brain activity an overview. In: Chase, M.H. (Ed.), Operant Conditioning of Brain Activity. University of California Press, Los Angeles, pp. 1243. Butti, C., Sherwood, C.C., Hakeem, A.Y., Allman, J.M., Hof, P.R., 2009. Total number and volume of Von Economo neurons in the cerebral cortex of cetaceans. J. Comp. Neurol. 515, 243249. Buxhoeveden, D.P., Casanova, M.F., 2002. The minicolumn hypothesis in neuroscience. Brain 125, 935951. Buzski, G., 2007. The structure of consciousness. Nature 446, 267. Buzski, G., 2010. Neural syntax: cell assemblies, synapsembles, and readers. Neuron 68, 362385. Bystron, I., Blakemore, C., Rakic, P., 2008. Development of the human cerebral cortex: Boulder Committee revisited. Nat. Rev. Neurosci. 9, 110122. Canales, A.F., Gmez, D.M., Maffe, C.R., 2007. A critical assessment of the consciousness by synchrony hypothesis. Biol. Res. 40, 517519. Carreira-Perpin, M.A., Goodhill, G.J., 2002. Are visual cortex maps optimized for coverage? Neural Comput. 14, 15451560. Chalmers, D.J., 2000. What is a neural correlate of consciousness? In: Metzinger, T. (Ed.), Neural Correlates of Consciousness: Empirical and Conceptual Questions. MIT Press, Cambridge, MA, pp. 1740. Chen, X., Gabitto, M., Peng, Y., Ryba, N.J.P., Zuker, C.S., 2011. A gustotopic map of taste qualities in the mammalian brain. Science 333, 12621266. Cook, N.D., 2008. The neuron-level phenomena underlying cognition and consciousness: synaptic activity and the action potential. Neuroscience 153, 556570. Coull, J.T., 1998. Neural correlates of attention and arousal: insights from electrophysiology, functional neuroimaging and psychopharmacology. Prog. Neurobiol. 55, 343361. Craig, A.D., 2009. How do you feel now? The anterior insula and human awareness. Nat. Rev. Neurosci. 10, 5970. Crick, F., 1984. Function of the thalamic reticular complex: the searchlight hypothesis. Proc. Natl. Acad. Sci. U. S. A. 81, 45864590. Crick, F., Koch, C., 2003. A framework for consciousness. Nat. Neurosci. 6, 119126. Crick, F., Koch, C., 2005. What is the function of the claustrum? Philos. Trans. R. Soc. Lond. B Biol. Sci. 360, 12711279. Cutsuridis, V., Wennekers, T., Graham, B.P., Vida, I., Taylor, J.G., 2009. Microcircuits: their structure, dynamics and role for brain function. Neural Netw. 22, 10371038. da Costa, N.M., Martin, K.A., 2010. Whose cortical column would that be? Front. Neuroanat. 4, 110.

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

19

Dahlstrm, A., Fuxe, K., 1964. Evidence for the existence of monoamine-containing neurons in the central nervous system. I. Demonstration of monoamines in the cell bodies of brain stem neurons. Acta Physiol. Scand. Suppl. 232, 155. de Veer, M.W., Gallup Jr., G.G., Theall, L.A., van den Bos, R., Povinelli, D.J., 2003. An 8-year longitudinal study of mirror self-recognition in chimpanzees (Pan troglodytes). Neuropsychologia 41, 229234. Decety, J., Grezes, J., 1999. Neural mechanisms subserving the perception of human actions. Trends Cogn. Sci. 3, 172178. Deco, D., Jirsa, V., McIntosh, A.R., Sporns, O., Ktter, R., 2009. Key role of coupling, delay, and noise in resting brain fluctuations. Proc. Natl. Acad. Sci. U. S. A. 106, 1030210307. DeFelipe, J., Alonso-Nanclares, L., Arellano, J.I., 2002. Microstructure of the neocortex: comparative aspects. J. Neurocytol. 31, 299316. Dehaene, S., 2009. Reading in the Brain. Viking, New York. Dennett, D.C., Kinsbourne, M.J., 1992. Time and the observer: the where and when of consciousness in the brain. Behav. Brain Sci. 15, 183247. Descarries, L., 1998. The hypothesis of an ambient level of acetylcholine in the central nervous system. J. Physiol. 92, 215220. Douglas, R.J., Koch, C., Mahowald, M., Martin, K.A., Suarez, H.H., 1995. Recurrent excitation in neocortical circuits. Science 269, 981985. Edelman, G.M., Tononi, G., 2000. A Universe of Consciousness. Basic Books, New York. Erzurumlu, R.S., Murakami, Y., Rijli, F.M., 2010. Mapping the face in the somatosensory brainstem. Nat. Rev. Neurosci. 11, 252263. Fajardo, C., Escobar, M.I., Buritic, E., Arteaga, G., Umbarila, J., Casanova, M.F., Pimienta, H., 2008. Von Economo neurons are present in the dorsolateral (dysgranular) prefrontal cortex of humans. Neurosci. Lett. 435, 215218. Ferrari, P.F., Visalberghi, E., Paukner, A., Fogassi, L., Ruggiero, A., Suomi, S.J., 2006. Neonatal imitation in rhesus macaques. PLoS Biol. 4, 15011508. Ferrari, P.F., Bonini, L., Fogassi, L., 2009. From monkey mirror neurons to primate behaviours: possible direct and indirect pathways. Philos. Trans. R. Soc. Lond. B Biol. Sci. 364, 23112323. Fleming, M.K., Stinear, C.M., Byblow, W.D., 2010. Bilateral parietal cortex function during motor imagery. Exp. Brain Res. 201, 499508. Freeman, W.J., 1997. Three centuries of category errors in studies of the neural basis of consciousness and intentionality. Neural Netw. 7, 11751183. Fuxe, K., 1965. Evidence for the existence of monoamine neurons in the central nervous system. IV. Distribution of monoamine nerve terminals in the central nervous system. Acta Physiol. Scand. Suppl. 64 (Suppl. 247), 3985. Fuxe, K., Canals, M., Torvinen, M., Marcellino, D., Terasmaa, A., Genedani, S., Leo, G., Guidolin, D., Diaz-Cabiale, Z., Rivera, A., Lundstrom, L., Langel, U., Narvaez, J., Tanganelli, S., Lluis, C., Ferre, S., Woods, A., Franco, R., Agnati, L.F., 2007a. Intramembrane receptorreceptor interactions: a novel principle in molecular medicine. J. Neural Transm. 114, 4975. Fuxe, K., Dahlstrom, A., Hoistad, M., Marcellino, D., Jansson, A., Rivera, A., Diaz-Cabiale, Z., Jacobsen, K., Tinner-Staines, B., Hagman, B., Leo, G., Staines, W., Guidolin, D., Kehr, J., Genedani, S., Belluardo, N., Agnati, L.F., 2007b. From the golgicajal mapping to the transmitter-based characterization of the neuronal networks leading to two modes of brain communication: Wiring and Volume Transmission. Brain Res. Rev. 55, 1754. Fuxe, K., Marcellino, D., Woods, A.S., Leo, G., Antonelli, T., Ferraro, L., Tanganelli, S., Agnati, L.F., 2009. Integrated signaling in heterodimers and receptor mosaics of different types of GPCRs

of the forebrain: relevance for schizophrenia. J. Neural Transm. 116, 923939. Fuxe, K., Dahlstrm, A.B., Jonsson, G., Marcellino, D., Guescini, M., Dam, M., Manger, P., Agnati, L.F., 2010. The discovery of central monoamine neurons gave Volume Transmission to the wired brain. Prog. Neurobiol. 90, 82100. Gallese, V., 2008. Mirror neurons and the social nature of language: the neural exploitation hypothesis. Soc. Neurosci. 3, 317333. Gallup Jr., G.G., 1970. Chimpanzees: self-recognition. Science 167, 8687. Gazzola, V., Keysers, C., 2009. The observation and execution of actions share motor and somato-sensory voxels in all tested subjects: single-subject analyses of unsmoothed fMRI data. Cereb. Cortex 19, 12391255. Gerardin, E., Sirigu, A., Lehericy, S., Poline, J.B., Gaymard, B., Marsault, C., Agid, Y., Le Bihan, D., 2000. Partially overlapping neural networks for real and imagined hand movements. Cereb. Cortex 10, 10931104. Ghosh, A., Rho, Y., McIntosh, A.R., Ktter, R., Jirsa, V.K., 2008. Noise during rest enables the exploration of the brain's dynamic repertoire. PLoS Comput. Biol. 4, e1000196. Giaume, C., Koulakoff, A., Roux, L., Holcman, D., Rouach, N., 2010. Astroglial networks: a step further in neuroglial and gliovascular interactions. Nat. Rev. Neurosci. 11, 8799. Golding, N.L., Staff, N.P., Spruston, N., 2002. Dendritic spikes as a mechanism for cooperative long-term potentiation. Nature 418, 326331. Goldman-Rakic, P.S., 1995. Cellular basis of working memory. Neuron 14, 477485. Gottfried, J.A., 2010. Central mechanisms of odour object perception. Nat. Rev. Neurosci. 11, 628641. Gray, C.M., Engel, A.K., Knig, P., Singer, W., 1989a. Stimulus-dependent neuronal oscillations in cat visual cortex. Receptive field properties and feature dependence. Eur. J. Neurosci. 2, 607 619. Gray, C.M., Knig, P., Engel, A.K., Singer, W., 1989b. Oscillatory responses in cat visual cortex exhibit inter-columnar synchronization which reflects global stimulus properties. Nature 338, 334337. Graybiel, A.M., Grillner, S., 2006. Microcircuits: The Interface between Neurons and the Global Brain Function. MIT Press, Boston. Guidolin, D., Albertin, G., Guescini, M., Fuxe, K., Agnati, L.F., 2011. Central nervous system and computation. Q. Rev. Biol. 86, 265285. Hakeem, A.Y., Sherwood, C.C., Bonar, C.J., Butti, C., Hof, P.R., Allman, J.M., 2009. Von Economo neurons in the elephant brain. Anat. Rec. (Hoboken) 292, 242248. Hameroff, S., 2010. The conscious pilot- dendritic synchrony moves through the brain to mediate consciousness. J. Biol. Phys. 36, 7193. Harvey, C.D., Svoboda, K., 2007. Locally dynamic synaptic learning rules in pyramidal neuron dendrites. Nature 450, 11951200. Hebb, D.O., 1949. The Organization of Behavior. Wiley, New York. Hobson, J.A., 2009. Rem sleep and dreaming: towards a theory of protoconsciousness. Nat. Rev. Neurosci. 10, 803813. Holtmaat, A.J., Trachtenberg, J.T., Wilbrecht, L., Shepherd, G.M., Zhang, X., Knott, G.W., Svoboda, K., 2005. Transient and persistent dendritic spines in the neocortex in vivo. Neuron 45, 279291. Hromdka, T., Zador, A.M., 2009. Representations in auditory cortex. Curr. Opin. Neurobiol. 19, 430433. Hubel, D.H., Wiesel, T.N., 1977. Functional architecture of macaque monkey visual cortex. Proc. R. Soc. Lond. B 198, 159. Hurley, S.L., 2008. The shared circuits model (SCM): how control, mirroring, and simulation can enable imitation, deliberation, and mindreading. Behav. Brain Sci. 31, 158. Iacoboni, M., Woods, R.P., Brass, M., Bekkering, H., Mazziotta, J.C., Rizzolatti, G., 1999. Cortical mechanisms of human imitation. Science 286, 25262528.

20

BR A IN RE S EA RCH 1 47 6 (2 0 1 2 ) 3 21

Jacob, F., 1977. Evolution and tinkering. Science 196, 11611166. John, E.R., 2002. The neurophysics of consciousness. Brain Res. Rev. 39, 128. Jouvet, M., 1972. The role of monoamines and acetylcholinecontaining neurons in the regulation of the sleepwaking cycle. Ergeb. Physiol. 64, 166307. Kenakin, T., Agnati, L.F., Caron, M., Fredholm, B., Guidolin, D., Kobilka, B., Lefkowitz, R.W., Lohse, M., Woods, A., Fuxe, K., 2010. International Workshop at the Nobel Forum, Karolinska Institutet on G protein-coupled receptors: finding the words to describe monomers, oligomers, and their molecular mechanisms and defining their meaning. Can a consensus be reached ? J. Recept. Signal Transduct. Res. 30, 284286. Kiianmaa, K., Fuxe, K., 1977. The effects of 5,7dihydroxytryptamine-induced lesions of the ascending 5-hydroxytryptamine pathways on the sleep wakefulness cycle. Brain Res. 131, 287301. Knight, R.T., 1997. Distributed cortical network for visual attention. J. Cogn. Neurosci. 9, 7591. Knsche, T.R., Tittgemeyer, M., 2011. The role of long-range connectivity for the characterization of the functionalanatomical organization of the cortex Front. Syst. Neurosci. 5, 113. Koblauch, A., Palm, G., Sommer, F.T., 2010. Memory capacities for synaptic and structural plasticity. Neural Comput. 22, 289341. Kosslyn, S.M., 1994. Image and Mind. Harvard University Press, Harvard. Laureys, S., 2005. The neural correlate of (un)awareness: lessons from the vegetative state. Trends Cogn. Sci. 9, 556559. Lidbrink, P., Fuxe, K., 1973. Effects of intracerebral injections of 6-hydroxydopamine on sleep and waking in the rat. J. Pharm. Pharmacol. 25, 8487. Llinas, R., 1988. The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous system function. Science 242, 16541664. Llinas, R., Par, D., 1991. Of dreaming and wakefulness. Neuroscience 44, 521535. Llinas, R., Ribary, U., 2001. Consciousness and the brain. The thalamocortical dialogue in health and disease. Ann. N Y Acad. Sci. 929, 166178. Llinas, R., Ribary, U., Contreras, D., Pedroarena, C., 1998. The neuronal basis for consciousness. Philos. Trans. R. Soc. Lond. B Biol. Sci. 353, 18411849. Lorente de N, R., 1938. Architectonics and structure of the cerebral cortex. In: Fulton, J.F. (Ed.), Physiology of the Nervous System. Oxford University Press, New York, pp. 291330. McCulloch, W.S., Pitts, W.H., 1943. A logical calculus of the ideas immanent in nervous activity. Bull. Math. Biophys. 5, 115133. McFadden, J., 2002. Hebbian reverberations in emotional memory electromagnetic field: evidence for an electromagnetic theory of consciousness. J. Conscious. Stud. 9, 2350. Merker, B., 2007. Consciousness without a cerebral cortex: a challenge for neuroscience and medicine. Behav. Brain Sci. 30, 63134. Mesulam, M., 2005. Imaging connectivity in the human cerebral cortex: the next frontier? Ann. Neurol. 57, 57. Miyawaki, Y., Uchida, H., Yamashita, O., Sato, M.A., Morito, Y., Tanabe, H.C., Sadato, N., Kamitani, Y., 2008. Visual image reconstruction from human brain activity using a combination of multiscale local image decoders. Neuron 60, 915929. Montagna, P., Gambetti, P., Cortelli, P., Lugaresi, E., 2003. Familial and sporadic fatal insomnia. Lancet Neurol. 2, 167176. Monti, M.M., Vanhaudenhuyse, A., Coleman, M.R., Boly, M., Pickard, J.D., Tshibanda, L., Owen, A.M., Laureys, S., 2010. Willful modulation of brain activity in disorders of consciousness. N. Engl. J. Med. 362, 579589. Moruzzi, G., Magoun, H.W., 1949. Brain stem reticular formation and activation of the EEG. Electroencephalogr. Clin. Neurophysiol. 1, 455473.

Mountcastle, V.B., Davies, P.W., Berman, A.L., 1957. Response properties of neurons of cats somatic sensory cortex to peripheral stimuli. J. Neurophysiol. 20, 374407. Murphy, S.T., Zajonc, R.B., 1993. Affect, cognition, and awareness: affective priming with optimal and suboptimal stimulus exposures. J. Pers. Soc. Psychol. 64, 723739. Oberheim, N.A., Wang, X., Goldman, S., Nedergaard, M., 2006. Astrocytic complexity distinguishes the human brain. Trends Neurosci. 29, 547553. Oberheim, N.A., Takano, T., Han, X., He, W., Lin, J.H., Wang, F., Xu, Q., Wyatt, J.D., Pilcher, W., Ojemann, J.G., Ransom, B.R., Goldman, S.A., Nedergaard, J., 2009. Uniquely hominid features of adult human astrocytes. Neuroscience 29, 32763287. Ostrow, L.W., Sachs, F., 2005. Mechanosensation and endothelin in astrocyteshypothetical roles in CNS pathophysiology. Brain Res. Rev. 48, 488508. Parpura, V., Scemes, E., Spray, D.C., 2004. Mechanisms of glutamate release from astrocytes: gap junction hemichannels, purinergic receptors and exocytotic release. Neurochem. Int. 45, 259264. Penrose, R., Gardner, M., 1999. The Emperor's New Mind: Concerning Computers, Minds and the Laws of Physics. Oxford University Press, Oxford. Pereira, A., Furlan, F.A., 2009. On the role of synchrony for neuronastrocyte interactions and perceptual conscious processing. J. Biol. Phys. 35, 465480. Pereira, A., Furlan, F.A., 2010. Astrocytes and human cognition: modeling information integration and modulation of neuronal activity. Prog. Neurobiol. 92, 405420. Perry, E.K., Walker, M., Grace, J., Perry, R., 1999. Acetylcholine in mind: a neurotransmitter correlate of consciousness? Trends Neurosci. 22, 273280. Pessoa, F., 1982. Livro de desassossego por Bernardo Soares. Atica, Lisboa. Peters, A., Sethares, C., 1996. Myelinated axons and the pyramidal cell modules in monkey primary visual cortex. J. Comp. Neurol. 365, 232255. Plotnik, J.M., de Waal, F.B., Reiss, D., 2006. Self-recognition in an Asian elephant. Proc. Natl. Acad. Sci. U. S. A. 103, 1705317057. Pockett, S., 2000. The Nature of Consciousness: A Hypothesis. Writers Club Press, Lincoln. Poldrack, R.A., 2006. Can cognitive processes be inferred from neuroimaging data? Trends Cogn. Sci. 10, 5963. Pppel, E., 1994. Temporal mechanisms in perception. Int. Rev. Neurobiol. 37, 185202. Pppel, E., Logothetis, N., 1986. Neuronal oscillations in the human brain. Discontinuous initiations of pursuit eye movements indicate a 30-Hz temporal framework for visual information processing. Naturwissenschaften 73, 267268. Premack, D., 2007. Human and animal cognition: continuity and discontinuity. Proc. Natl. Acad. Sci. U. S. A. 104, 1386113867. Rakic, P., 2008. Confusing cortical columns. Proc. Natl. Acad. Sci. U. S. A. 105, 1209912100. Rakic, P., 2009. Evolution of the neo-cortex: a perspective from develop-mental biology. Nat. Rev. Neurosci. 10, 724735. Riad, M., Garcia, S., Watkins, K.C., Jodoin, N., Doucet, E., Langlois, X., El Mestikawy, S., Hamon, M., Descarries, L., 2000. Somatodendritic localization of 5-HT1A and preterminal axonal localization of 5-HT1B serotonin receptors in adult rat brain. J. Comp. Neurol. 417, 181194. Rinkus, G.Y., 2010. A cortical sparse distributed coding model linking mini- and macrocolumn-scale functionality. Front. Neuroanat. 4, 113. Rizzolatti, G., Craighero, L., 2004. The mirror-neuron system. Annu. Rev. Neurosci. 27, 169192. Rizzolatti, G., Fabbri-Destro, M., Cattaneo, L., 2009. Mirror neurons and their clinical relevance. Nat. Clin. Pract. Neurol. 5, 2434.

BR A IN RE S EA RCH 1 4 76 (2 0 1 2 ) 3 21

21

Robertson, J.M., 2002. The astrocentric hypothesis: proposed role of astrocytes in consciousness and memory formation. J. Physiol. Paris 96, 251255. Rockland, K.S., 2010. Five points on columns. Front. Neuroanat. 4, 110. Romo, R., Brody, C.D., Hernndez, A., Lemus, L., 1999. Neuronal correlates of parametric working memory in the prefrontal cortex. Nature 399, 470473. Rouach, N., Koulakoff, A., Abudara, V., Willecke, K., Giaume, C., 2008. Astroglial metabolic networks sustain hippocampal synaptic transmission. Science 322, 15511555. Seeley, W.W., Carlin, D.A., Allman, J.M., Macedo, M.N., Bush, C., Miller, B.L., Dearmond, S.J., 2006. Early frontotemporal dementia targets neurons unique to apes and humans. Ann. Neurol. 60, 660667. Semyanov, A., Kullmann, D.M., 2002. Kainate receptor-dependent axonal depolarization and action potential initiation in interneurons. Nat. Neurosci. 4, 718723. Sevush, S., 2006. Single-neuron theory of consciousness. J. Theor. Biol. 238, 704725. Shepherd, G.M., 1979. The Synaptic Organization of the Brain. Oxford University Press, New York. Shepherd, G.M., 2011. The microcircuit concept applied to cortical evolution: from three-layer to six-layer cortex. Front. Neuroanat. 5, 30. Shiffrin, R.M., Dumais, S.T., Schneider, W., 1981. Characteristics of automatism. In: Long, J., Baddeley, A. (Eds.), Attention and Performance IX. Erlbaum, Hillsdale (NJ), pp. 223240.

Shmuel, A., Leopold, D.A., 2008. Neuronal correlates of spontaneous fluctuations in fMRI signals in monkey visual cortex: implications for functional connectivity at rest. Hum. Brain Mapp. 29, 751761. Stephan, K.M., Fink, G.R., Passingham, R.E., Silbersweig, D., Ceballos-Baumann, A.O., Frith, C.D., Frackowiak, R.S., 1995. Functional anatomy of the mental representation of upper extremity movements in healthy subjects. J. Neurophysiol. 73, 373386. Tanne-Gariepy, J., Boussaoud, D., Rouiller, E.M., 2002. Projections of the claustrum to the primary motor, premotor, and prefrontal cortices in the macaque monkey. J. Comp. Neurol. 454, 140157. Thierry, A.M., Stinus, L., Blanc, G., Glowinski, J., 1973. Some evidence for the existence of dopaminergic neurons in the rat cortex. Brain Res. Rev. 50, 230234. Wang, X.-J., 2001. Synaptic reverberation underlying mnemonic persistent activity. Trends Neurosci. 24, 455463. Webster's College Dictionary, 1995. Random House. Welzel, O., Tischbirek, C.H., Jung, J., Kohler, E.M., Svetlitchny, A., et al., 2010. Synapse clusters are preferentially formed by synapses with large recycling pool sizes. PLoS One 5, e13514. Werner, G., 2007. Perspectives on the neuroscience of cognition and consciousness. Biosystems 87, 8295. Woolsey, T.A., van der Loos, H., 1970. The structural organization of layer IV in the somatosensory region (SI) of mouse cerebral cortex. Brain Res. 17, 205242. Zurborg, S., Yurgionas, B., Jira, J.A., Caspani, O., Heppenstall, P.A., 2007. Direct activation of the ion channel TRPA1 by Ca2 +. Nat. Neurosci. 10, 277279.