Neocortex Synaptic Organization PDF

1 LECTURE 8: SYNAPTIC ORGANIZATION OF THE NEOCORTEX PROF. JUNIOR STEININGER 2 CEREBRAL CORTEX Distinctive lamination 3 phylogenetic categories of cerebral cortex are: Archicortex: “first”, 3-4 layers Paleocortex: “ancient”, 3 layers N...

1 LECTURE 8: SYNAPTIC ORGANIZATION OF THE NEOCORTEX PROF. JUNIOR STEININGER 2 CEREBRAL CORTEX Distinctive lamination 3 phylogenetic categories of cerebral cortex are: Archicortex: “first”, 3-4 layers Paleocortex: “ancient”, 3 layers Neocortex: “new,” 6 layers; detailed perception, learning, intelligence 3 NEOCORTEX: PRIMARY SENSORY/MOTOR AREAS Wilder Penfield’s “Montreal procedure” for treating epileptic seizures: Involved destroying neurons but first stimulated areas to test function and avoid vital areas 4 NEOCORTEX: ASSOCIATION CORTICES Most of the cerebral surface of the human brain Diverse functions of the association cortices are loosely referred to as “cognition” Cognition: ability to attend to external stimuli or internal motivation; to identify the significance of such stimuli; and to make appropriate responses 5 MAMMALIAN CEREBRAL CORTEX In lower mammals the cerebral cortex that overlies the limbic system is a simple, smooth structure. Higher mammals and humans have more folds, and a well defined ‘temporal lobe’ and a bigger frontal lobe. 6 HUMAN CEREBRAL CORTEX: GYRI AND SULCI BRODMANN’S AREAS: 7 FUNCTIONAL LOCALIZATION Brodmann’s Areas (numbered areas) were originally defined based on the cytoarchitectural organization of neurons using Nissl stain Identified with little or no knowledge of functional significance 8 BRODMANN’S AREAS: FUNCTIONAL LOCALIZATION PREFRONTAL CORTEX 9 (PFC) Highest level of cortical hierarchy dedicated to memory, planning and execution of actions Dorsolateral PFC: Right hemisphere – spatial working memory Left hemisphere – language, literacy & logic Ventromedial PFC including Orbitofrontal PFC: Motivated decision making, emotional and autonomic responses, conscience, morality, social skills 10 PFC CHANGES THROUGHOUT LIFESPAN Developmentally, the prefrontal cortex is one of the last structures in the brain to become fully myelinated PFC does not reach functional maturation until ~15-19 yrs After adolescence and through adult life the cytoarchitecture of the human cortex remains relatively stable. In the 7th/8th decade of life -> neuronal “involution” in PFC decrease in size, volume and density of cells (occurs gradually) Shrinkage/disappearance of dendrites 11 PHINEAS GAGE: ORBITOFRONTAL LOBOTOMY 1848 – one of the first demonstrations that certain functions are localized in cerebral cortex Personality transformed from “friendly, conscientious, godfearing” -> “gross, profane and irreverent” = loss of inhibition 12 PHINEAS GAGE: ORBITOFRONTAL LOBOTOMY 1861: Phineas dies, but no autopsy performed on his brain. 1994: from skull, used computer graphics and neural imaging techniques to plot the trajectory of the steel rod bilateral damage to the ventromedial region of the frontal lobes (i.e. decision making, morality, social skills) 13 PFC DURING WORKING MEMORY There is evidence that different PFC regions contribute to different types of working memory (domain specificity theory). 14 DL-PFC AND SPATIAL WORKING MEMORY TASK A dlPFC neuron shows persistent firing during the delay period if the cue had occurred at its preferred direction (e.g. 90˚ in this example) 15 DOMAIN SPECIFICITY THEORY 16 WHAT CAUSES PERSISTENT ACTIVITY? LET’S LOOK AT THE BIOPHYSICS OF NEURONS AND SYNAPSES, AND CIRCUIT CONNECTIVITY WITHIN THE CORTEX FIRST…. 17 TWO PROMINENT PRINCIPLES OF NEOCORTICAL ORGANIZATION: LAYERS AND COLUMNS 18 NEOCORTEX LAMINAR ORGANIZATION LAYER OF GRAY MATTER (UNMYELINATED AXONS) AT THE SURFACE OF THE CEREBRAL HEMISPHERE 19 NEOCORTEX LAMINAR ORGANIZATION Layer of gray matter (unmyelinated axons) at the surface of the cerebral hemisphere Laminar division Six layers 20 NEOCORTEX – HORIZONTAL LAYERS 21 NEOCORTEX – HORIZONTAL LAYERS Layer 1 – consists primarily of dendrites, few cell bodies Layer 2+3 considered together Similar patterns of connectivity Difficult to distinguish experimentally 22 HORIZONTAL LAYERS: REGIONAL DIFFERENCES THICKNESS OF LAYERS CAN VARY BETWEEN CORTICAL REGIONS 23 NEOCORTICAL PRINCIPAL CELLS Principal Cells (~80%): release glutamate 1. Pyramidal Cell Major output cells Large, conical cell body Apical and basal dendrites Axon enters white matter 2. Granule Cell (aka stellate cell) Small, round cell body; found in layer 4 only Major recipient of thalamic input Synapse on output neurons (i.e. pyramidal cells) of cortex 24 NEOCORTICAL GABAERGIC CELLS Local neurons (~20%): release GABA Many different types: bitufted, double bouquet, small basket, large basket, chandelier, undesignated cell (long stringy), neurogliaform 25 NEOCORTEX CANONICAL ORGANIZATION The circuitry of all cortical regions has some common features: Primary source of inputs Primary output target Connections in vertical axis Columnar (or radial) connections Connections in horizontal axis Horizontal (or lateral) connections 26 NEOCORTICAL AFFERENTS (INPUTS)  27 NEOCORTICAL EFFERENTS (OUTPUTS) Majority of axons remain in cortex (~90-99%) Cortico-cortical outputs going to other cortical areas arise from the superficial Layers 2/3 Same or opposite hemisphere Layer 5/6 pyramidal cells send outputs to thalamus and subcortical structures (e.g., basal ganglia, brainstem, spinal cord and superior colliculus, depending on region) 28 CORTICAL MICROCIRCUIT FEEDBACK/RECURRENT EXCITATION BETWEEN PRINCIPAL CELLS 29 CORTICAL MICROCIRCUIT MANY INHIBITORY NEURONS TARGET SOMA/AXON HILLOCK, PROVIDING FEEDFORWARD OR FEEDBACK INHIBITION 30 CORTICAL MICROCIRCUIT AFFERENTS TERMINATE ON BOTH PRINCIPAL CELLS AND INTERNEURONS 31 CANONICAL MODEL OF SENSORY INFORMATION PROCESSING Thalamocortical projection to L4: main route for sensory organs to communication with cortex Information then propagates through cortical column pathway: L4 → L2/3 → L5/6 32 VERTICAL ORGANIZATION - CELL COLUMNS “Barrel” cortex: anatomically distinct regions of primary sensory cortex in rats containing sensory input from each individual whisker 33 VERTICAL ORGANIZATION - CELL COLUMNS 34 CANONICAL MODEL OF SENSORY INFORMATION PROCESSING Similar receptive field cells in LGN are summed to form the receptive field of a “simple cell” in the primary visual cortex Numerous simple cells can then be summed to form the receptive field of a single “complex cell” Hierarchical Processing 35 FUNCTIONAL CORTICAL COLUMNS IN V1 Primary visual cortex also contains “hypercolumns” to analyze discrete regions of the visual field includes complete set of ocular dominance and orientation columns (to represent 360˚) and several “blobs” (discrete regions specific for colour) 36 FUNCTIONAL CORTICAL COLUMNS IN V1 Primary visual cortex also contains “hypercolumns” to analyze discrete regions of the visual field includes complete set of ocular dominance and orientation columns (to represent 360˚) and several “blobs” (discrete regions specific for colour) Recommended: https://www.youtube.com/watch?v=8VdFf3egwfg 37 FUNCTIONAL COLUMNS – BEYOND THE PRIMARY SENSORY AND MOTOR CORTICES It had been thought that a functional columnar organization doesn’t exist beyond the primary sensory and motor cortices. However, there is now evidence to suggest that columnar organization is a feature in associational areas and PFC too. FUNCTIONAL COLUMNS IN 38 INFEROTEMPORAL CORTEX The last stage of the ventral (“what”) stream of visual information processing, important for object and face recognition. Responds to stimuli such as faces and objects with complex features Columns that represent different but related features overlap and constitute a larger scale unit. For instance, there may be a continuous mapping of different views of faces (monkey) along the cortical surface. 39 SUMMARY: FUNCTIONAL COLUMN CIRCUITS 40 SUMMARY: GENERAL PRINCIPLES OF ORGANIZATION Cortical processing is predominantly local and vertical Vertical, functional columns may form a basic, repeating unit of information processing in the cortex, each responsible for representing a small range of stimuli Questions remain as to why the columnar structure exists, as they are absent in some closely related mammalian species. Old World monkeys (e.g. macaques) have ocular dominance columns, but New World monkeys (e.g. marmosets) don’t! Barrels exist in brains of rats, mice, squirrels but are absent in other species which have prominent whiskers, such as cats, dogs, racoons Thus, the columnar organization may not be a necessary feature for information processing 41 NOW, BACK TO OUR QUESTION FROM EARLIER: Can we explain persistent activity in terms of the biophysics of neurons and synapses, and circuit connectivity in the PFC? 42 EXPLANATION OF PERSISTENT ACTIVITY From what we know about the organization of the cortex, the selectivity and sensitivity of persistent activity in PFC neurons to certain directions can be explained in terms of: The existence of discrete functional domains/columns with different “direction memory”  43 EXPLANATION OF PERSISTENT ACTIVITY What kind of a neural network can sustain stimulus-selective persistent activity in the absence of external inputs? Reccurent excitatory loops What type of neural mechanism can be turned on and off rapidly (~100ms)? Local Inhibition 44 PERSISTENT CORTICAL ACTIVITY ALSO SHOWS “UP” AND “DOWN” STATES Cortical “slow oscillation”: a periodic, rhythmic spontaneous activity characterized by periods of sustained depolarization (“up” state) followed by periods of hyperpolarization and silence (“down” state) Function of slow oscillations: cellular correlates of memory maintenance, associated with slow-wave sleep, and may coordinate other sleep rhythms such as spindle and delta waves 45 SEVERAL RECURRENT EXCITATORY LOOPS ARE PRESENT IN THE NEOCORTEX: Layer 6 -> Layer 4 -> Layer 6 Layer 5 -> Layer 5 Layer 5 -> Layers 5 and 2/3 -> Layer 5 46 LOCAL CIRCUIT RECURRENT EXCITATION AS MODEL FOR PERSISTENT ACTIVITY 47 LOCAL CIRCUIT RECURRENT EXCITATION AS MODEL FOR PERSISTENT ACTIVITY 48 INFORMATION PROCESSING IN NEOCORTEX Grandmother cell theory: identity of a stimulus can be determined by considering the activation of a single neuron VS. Distributed coding theory: each neuron in a network is involved in coding more than one familiar thing 49 UPDATED MODEL OF INFORMATION PROCESSING CANONICAL VIEW: THALAMUS → L4 → L2/3 → L5/6  50 UPDATED MODEL OF INFORMATION PROCESSING CANONICAL VIEW: THALAMUS → L4 → L2/3 → L5/6 Anatomy shows Thalamus → L5/6 but connection often ignored Experiment: paired, in vivo recordings from L5/6 neurons and thalamic neurons, in sedated rats Do deep cortical layers respond to thalamic inputs?  51 UPDATED MODEL OF INFORMATION PROCESSING Results Thalamic input reliably caused pyramidal neurons in layer 5 to fire Pharmacological inactivation of L4 had no effect on sensory-evoked synaptic input to L5/6 neurons. L4 is not an obligatory distribution hub for cortical activity Thalamus activates two separate, independent “strata” of cortex in parallel …. But why??  52 UPDATED MODEL OF INFORMATION PROCESSING Results WHY? 1. Redundancy - unlikely (cells behave very differently hooked up to different parts of the brain) 2. Equivalent partners in making behaviors work 3. One is involved in more abstract/contextual representations than the other