Automated Reconstruction of Neural Tissue and the

Authors: James Kozloski

Publish Date: 2011/05/03

Volume: 9, Issue: 2-3, Pages: 133-142

PDF Link

Abstract

The brain implements a myriad of global brain functions to support adaptive behaviors Despite their seeming innumerability these emerge from combinations of lower level functions implemented by a relatively small set of brain tissues Evidence from brain imaging studies shows that spatiotemporal patterns of activations across different brain tissues correlate with brain function and hence with an organism’s behavior To support a diversity of global functions gross connections between brain tissues while structurally static must undergo modulation The strength of this modulation can define functional boundaries and interfaces between brain tissues wherever functional relationships between brain regions are highly modulated tissue boundaries occurTissuelevel functions while also diverse are more stereotyped than global brain functions Similar to spatiotemporal modulation and recombination of tissue activation variation and recombination of familiar structural elements of the brain neurons and their connections synapses generate tissuelevel functions Unlike other organs’ gross morphological specializations of single tissues eg muscle bone brain specialization yields distinct tissues derived from stationary statistical combinations of a variety of neuron and synapse types in space which we define as microcircuitry Measurable consistent patterning of microcircuitry across a tissue and in different organisms ie stereotypy further defines a tissue’s boundaries wherever patterning changes abruptly one tissue ends and another beginsShepherd defined microcircuits abstractly and independent of neural tissues based on simple computations they might implement1 Defining stereotyped microcircuitry as a stationary combination of neuron and synapse types within a specific tissue restricts strong synaptic plasticity to its boundaries Where plasticity is strongest stationary circuit components are recombined to serve underlying tissuelevel functions for example learning and memory2 Observations that strong departures from stereotypy in developing vertebrate tissue arise where neural competition dominates supports this view3 We therefore define microcircuitry circumscribed by strong plasticity as a microcircuit which is then iterated to create a tissueFor example cerebellar tissue derives from a microcircuit iterated millions of times4 Boundaries between components occur at highly plastic parallel fiber synapses onto Purkinje cells Similarly neocortex derives from a microcircuit with stereotypical properties along its radial axis iterated many millions of times within the cortical plane5 Highly plastic lateral connections between microcircuitry delineate the columnar cortical microcircuit smaller and distinct from functional cortical columns that are characterized by intrinsic variability in receptive fields and connections2How do we attack the problem of analyzing the functions of neural tissues and synthesize a theory of global brain function One option is to first map microcircuits in these tissues then use maps to constrain functional simulations aimed at modeling and explaining function Long underway6 mapping approaches change as new techniques are developed to attack the problem7 8 9Typically approaches study functional connectivity between neurons using physiological recording techniques10 or reconstruct and analyze tissue structure at the level of neurons and their connections by determining threedimensional locations of tissue components and their relationships 11 The purpose of this commentary is to consider how and the degree to which high throughput reconstruction might transform mapping microcircuits in the brain both in technical execution and in its application to elucidating global brain functionHigh throughput neural tissue reconstruction depends first on treating a tissue to reveal its histological structure12 The usability of structural data is determined first by the resolution of the light microscope Small caliber fibers for example axons found in all microcircuits typically lie near diffraction limits of resolution such that only experimental fluorescent microscopic techniques promise to achieve sensitivity necessary to resolve sparsely stained tissues13 Second data usability depends on resolution of relationships between fibers in densely stained tissue This requirement which we term relationship determination depends on but is not equivalent to fiber resolutionTo illustrate relationship determination consider two fibers that originate and terminate at separate resolvable points The proximity of their component sections may change making them impossible to resolve at some intermediate point When this degradation occurs relationships between the unresolved fiber components and their relationships to all subsequent components become uncertain Thus relationship determination remains degraded even when component resolution recovers

Keywords: