Opinion
A cellular mechanism for cortical associations: an organizing principle for the cerebral cortex

https://doi.org/10.1016/j.tins.2012.11.006Get rights and content

A basic feature of intelligent systems such as the cerebral cortex is the ability to freely associate aspects of perceived experience with an internal representation of the world and make predictions about the future. Here, a hypothesis is presented that the extraordinary performance of the cortex derives from an associative mechanism built in at the cellular level to the basic cortical neuronal unit: the pyramidal cell. The mechanism is robustly triggered by coincident input to opposite poles of the neuron, is exquisitely matched to the large- and fine-scale architecture of the cortex, and is tightly controlled by local microcircuits of inhibitory neurons targeting subcellular compartments. This article explores the experimental evidence and the implications for how the cortex operates.

Introduction

The cortex remains an enigmatic structure, at once beautifully simple and yet mysterious. After more than a century of concerted investigation, both the purpose and operating principles of the cerebral cortex are hotly debated 1, 2, 3, 4. It is still deeply puzzling how neurons in different regions, sometimes many centimeters apart, can be linked with each other and act in concert to form single conscious percepts [5]. But even basic questions such as why the cortex is layered [6] or made up of 70–80% pyramidal neurons [7] remain unanswered. Most computational models of cortical function still treat neurons as simple single-compartment units even though the potential power of single neurons in information processing has long been understood [8]. Here, the view is presented that both the cellular properties and architecture of the cortex are tightly coupled, suggesting a powerful operating principle of the cortex.

At first glance, the architecture of the cortex seems bizarre. Long-range connectivity in the cortex follows the basic rule that sensory input (i.e., the feed-forward stream) terminates in the middle cortical layers, whereas information from other parts of the cortex (i.e., the feedback stream) tends to project to the outer layers 9, 10, 11, 12, 13 (Figure 1). This also applies to projections from the thalamus, a structure that serves as both a gateway for feed-forward sensory information to the cortex and as a hub for feedback interactions between cortical regions (Box 1) 14, 15, 16. This wiring feature is mysterious because the principal targets for such feedback connections in L1 are a handful of interneurons [17] and the very distal tuft dendrites of pyramidal neurons. Referred to as the ‘crowning mystery’ of the cortex by David Hubel [18], only ∼10% of the synaptic inputs to L1 come from nearby neurons and the missing 90% from long-range feedback connections 19, 20.

A naïve interpretation of this architecture would lead to the conclusion that feedback information is relatively inconsequential compared to the feed-forward stream. However, it is clear from multiple lines of research that the feedback information stream is in fact vitally important for cognition 21, 22, 23, 24, 25 and conscious perception 26, 27, 28, 29. This has led to the suggestion that the cortex operates via an interaction between feed-forward and feedback information 30, 31, 32. In this scenario, feedback provides context or predictive information for modulating neural activity in a given area 33, 34, 35, and also provides a mechanism for the cortex to attend to particular features [36].

Because feedback targets the apical dendrites of cortical pyramidal neurons in L1, several authors have proposed an important role for these dendrites 16, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47. However, all these theories must contend with the fact that the bulk of cortical feedback inputs arrive at the most electrically remote region of the pyramidal neuron distal tuft dendrites, where they have the least influence on spike generation in the axon 48, 49. Thus, understanding the properties of these dendrites becomes central to explaining the influence of feedback connectivity in the brain. The aim of this article is to set out the evidence for considering the pyramidal neuron to be an associative element and how this property relates to cortical function.

Section snippets

The calcium-spike initiation zone in L5 pyramidal neurons

With regard to the remoteness of the tuft dendrite in L1, a key finding was the discovery of a second initiation zone for broad calcium action potentials (‘Ca2+ spikes’) near the apical tuft of layer 5 (L5) pyramidal neurons 50, 51, 52, 53, 54. Feedback inputs to the tuft are therefore much closer to this action-potential initiation zone than to the one in the axon at the other end of the cell. Moreover, the calcium spike is a tremendously explosive engine, driving the L5 pyramidal cell to fire

Backpropagation-activated coupling: an associative mechanism within each cell

The discovery of the dendritic Ca2+ spike did not immediately solve the problem of remote input to the tuft. It still remained to be explained how distal synaptic input could overcome the threshold for evoking such dendritic spikes because even input directly to the tuft has little effect on the apical Ca2+ initiation zone 49, 63. The conceptual breakthrough emerged from the demonstration that the Na+ and Ca2+ spike initiation zones can influence each other 59, 69 (Figure 2). This occurs via

Inhibitory control of BAC firing

It is now abundantly clear that associative pairing can be very effectively blocked by specific inhibitory neurons in the cortical microcircuit, placing special significance on dendrite targeting inhibition [73]. This powerful suppression of dendritic plateau potentials has been observed in vitro 69, 74, 75, 76, 77 and in vivo 72, 78. Anesthesia – which is associated with elevated inhibition [79] – also dampens dendritic calcium spikes in vivo [80]. Although the activation of both GABAA and GABA

BAC firing and cortical information processing

To this point, this article has dealt with the biophysical evidence for the existence of an associative firing mechanism in pyramidal neurons and its influence on the input/output function. This degree of integration between the micro- and macroarchitecture, as well as inbuilt complexity at the cellular level, invites speculation about whether and how the whole system utilizes this feature. The importance of this mechanism conceptually is that the pyramidal neuron is able to detect coincident

Evidence for BAC firing in vivo and the relationship to conscious perception

By their nature, the distal dendrites of pyramidal neurons are extremely difficult to investigate in living animals. However, advances with fiber-optic approaches and two-photon microscopy are starting to reveal evidence for the existence of the BAC firing mechanism in vivo. Dendritic calcium spikes have been recorded in vivo 57, 84, 85 that correlate to behavior 78, 86. Using a fiber-optic imaging technique, which enabled recordings specifically from populations of tuft dendrites of L5

Concluding remarks

The BAC firing hypothesis presented here offers a cellular mechanism that addresses a number of questions about the cortex. It suggests that the pyramidal neuron cell type is an associative element which carries out the same essential task at all cortical stages: that of coupling feed-forward and feedback information at the cellular level. This mechanism succinctly explains the advantage of the cortical hierarchy with its structured terminations in different cortical layers, and also offers a

Acknowledgments

This work was supported by the NeuroCure Excellence Cluster, Berlin, SystemsX.ch (NeuroChoice) and the Swiss National Science Foundation (31003A_130694).

References (116)

  • K. Rauss

    Top-down effects on early visual processing in humans: a predictive coding framework

    Neurosci. Biobehav. Rev.

    (2011)
  • B. Noudoost

    Top-down control of visual attention

    Curr. Opin. Neurobiol.

    (2010)
  • S. Sevush

    Single-neuron theory of consciousness

    J. Theor. Biol.

    (2006)
  • D. Laberge et al.

    The apical dendrite theory of consciousness

    Neural Netw.

    (2007)
  • R. Yuste

    Ca2+ accumulations in dendrites of neocortical pyramidal neurons: an apical band and evidence for two functional compartments

    Neuron

    (1994)
  • J.E. Lisman

    Bursts as a unit of neural information: Making unreliable synapses reliable

    Trends Neurosci.

    (1997)
  • E.M. Izhikevich

    Bursts as a unit of neural information: selective communication via resonance

    Trends Neurosci.

    (2003)
  • R. Miles

    Differences between somatic and dendritic inhibition in the hippocampus

    Neuron

    (1996)
  • E. Pérez-Garci

    The GABAB1b isoform mediates long-lasting inhibition of dendritic Ca2+ spikes in layer 5 somatosensory pyramidal neurons

    Neuron

    (2006)
  • G. Silberberg et al.

    Disynaptic inhibition between neocortical pyramidal cells mediated by Martinotti cells

    Neuron

    (2007)
  • B.J. He et al.

    The fMRI signal, slow cortical potential and consciousness

    Trends Cogn. Sci.

    (2009)
  • J. Bullier

    The role of feedback connections in shaping the responses of visual cortical neurons

    Prog. Brain Res.

    (2001)
  • E.M. Callaway

    Feedforward, feedback and inhibitory connections in primate visual cortex

    Neural Netw.

    (2004)
  • M.H. Herzog et al.

    The role of feedback in learning a vernier discrimination task

    Vision Res.

    (1997)
  • J.C. Horton et al.

    The cortical column: a structure without a function

    Philos. Trans. R. Soc. Lond B: Biol. Sci.

    (2005)
  • S.D. Van Hooser

    Similarity and diversity in visual cortex: is there a unifying theory of cortical computation?

    Neuroscientist

    (2007)
  • K.S. Rockland et al.

    Some thoughts on cortical minicolumns

    Exp. Brain Res.

    (2004)
  • R.J. Douglas et al.

    Neuronal circuits of the neocortex

    Annu. Rev. Neurosci.

    (2004)
  • R. Nieuwenhuys

    The neocortex. An overview of its evolutionary development, structural organization and synaptology

    Anat. Embryol. (Berl.)

    (1994)
  • C. Koch et al.

    The role of single neurons in information processing

    Nat. Neurosci.

    (2000)
  • D.J. Felleman et al.

    Distributed hierarchical processing in the primate cerebral cortex

    Cereb. Cortex

    (1991)
  • T.A. Coogan et al.

    Conserved patterns of cortico-cortical connections define areal hierarchy in rat visual cortex

    Exp. Brain Res.

    (1990)
  • L.J. Cauller

    Backward cortical projections to primary somatosensory cortex in rats extend long horizontal axons in layer I

    J. Comp. Neurol.

    (1998)
  • V.C. Wimmer

    Dimensions of a projection column and architecture of VPM and POm axons in rat vibrissal cortex

    Cereb. Cortex

    (2010)
  • P. Rubio-Garrido

    Thalamic input to distal apical dendrites in neocortical layer 1 is massive and highly convergent

    Cereb. Cortex

    (2009)
  • R.R. Llinás

    Temporal binding via cortical coincidence detection of specific and nonspecific thalamocortical inputs: a voltage-dependent dye-imaging study in mouse brain slices

    Proc. Natl. Acad. Sci. U.S.A.

    (2002)
  • H. Markram

    Interneurons of the neocortical inhibitory system

    Nat. Rev. Neurosci.

    (2004)
  • D.H. Hubel

    Cortical neurobiology: a slanted historical perspective

    Annu. Rev. Neurosci.

    (1982)
  • Y. Shu

    Turning on and off recurrent balanced cortical activity

    Nature

    (2003)
  • A. Soltani et al.

    Visual saliency computations: mechanisms, constraints, and the effect of feedback

    J. Neurosci.

    (2010)
  • I.R. Olson

    Contextual guidance of attention – human intracranial event-related potential evidence for feedback modulation in anatomically early, temporally late stages of visual processing

    Brain

    (2001)
  • A. Pascual-Leone et al.

    Fast backprojections from the motion to the primary visual area necessary for visual awareness

    Science

    (2001)
  • H. Tomita

    Top-down signal from prefrontal cortex in executive control of memory retrieval

    Nature

    (1999)
  • J.J. van Boxtel

    Consciousness and attention: on sufficiency and necessity

    Front. Psychol.

    (2010)
  • E.R. John et al.

    The anesthetic cascade: a theory of how anesthesia suppresses consciousness

    Anesthesiology

    (2005)
  • M. Boly

    Preserved feedforward but impaired top-down processes in the vegetative state

    Science

    (2011)
  • R.P. Rao et al.

    Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects

    Nat. Neurosci.

    (1999)
  • M. Bar

    A cortical mechanism for triggering top-down facilitation in visual object recognition

    J. Cogn. Neurosci.

    (2003)
  • H.T. Chang

    Dendritic potential of cortical neurons produced by direct electrical stimulation of the cerebral cortex

    J. Neurophysiol.

    (1951)
  • J.C. Eccles

    Evolution of consciousness

    Proc. Natl. Acad. Sci. U.S.A.

    (1992)
  • Cited by (514)

    • Why birds are smart

      2024, Trends in Cognitive Sciences
    View all citing articles on Scopus
    View full text