advertisement

EGR 183 Bow Tie Presentation

33 %
67 %
advertisement
Information about EGR 183 Bow Tie Presentation
Health & Medicine

Published on March 20, 2009

Author: jme17

Source: slideshare.net

Description

Long (teaching, novice) version powerpoint summarizing research paper file 'Paper EGr 183 Modeling Neural Networks in silico'.
advertisement

“ How Did Nature Solve the information processing problems through the development of neural networks as well as the subsequent training and coordination with other networks?” Daniel Carlin, Daniel Cook, Joshua Mendoza-Elias Macro Micro Nano Macro Micro The Physical and Chemical Limitations that Nature Overcame were: solubility, dispersibility, surface area , conduction, signal modulation The Problem Nature HAS in Disease is: the severing of a neural network: paralysis, loss of sensation. Nano LTP-KO: NMDA Receptor Agonist CPP No change synaptic efficacy No Learning Loss of memory

“ How Did Nature Solve the information processing problems through the development of neural networks as well as the subsequent training and coordination with other networks?” Daniel Carlin, Daniel Cook, Joshua Mendoza-Elias Macro Micro Nano Macro Micro The Physical and Chemical Limitations that Nature Overcame were: solubility, dispersibility, surface area , conduction, signal modulation The Problem Nature HAS in Disease is: the severing of a neural network: paralysis, loss of sensation. Nano

Nervous System: FUNCTION Receptors collect information through mechanical or chemical changes to protein structure. Receptors cause neurons to fire. Intensity is modulated by frequency. Firing are “additive.” The aggregate of neural signalling determines the output of a network. The subnetworks may interact with larger networks. Large network come to control many subnetworks -e.g. Homeostasis, visual pathways, audition, vestibular system, motor control, visceral motor system. Cognition: Short terms memory, Long term memory, emotional regulation

Receptors collect information through mechanical or chemical changes to protein structure.

Receptors cause neurons to fire. Intensity is modulated by frequency.

Firing are “additive.” The aggregate of neural signalling determines the output of a network.

The subnetworks may interact with larger networks.

Large network come to control many subnetworks

-e.g. Homeostasis, visual pathways, audition, vestibular system, motor control, visceral motor system.

Cognition: Short terms memory, Long term memory, emotional regulation

Function continued Evolutionarily, the nervous system is intended to: (i.) regulate physiological processes (ii.) collect information about an organisms’s environment in 5 dimensions (5 senses). (iii.) prompt innate behaviors to react to environment  programmed behavior (iv.) prompt learned behaviors to interact with the environment  cognition (v.) Override control with other “scripts”: Breathing, reflexes, Fight/Flight

Evolutionarily, the nervous system is intended to:

(i.) regulate physiological processes

(ii.) collect information about an organisms’s environment in 5 dimensions (5 senses).

(iii.) prompt innate behaviors to react to environment  programmed behavior

(iv.) prompt learned behaviors to interact with the environment  cognition

(v.) Override control with other “scripts”:

Breathing, reflexes, Fight/Flight

Component Design Nervous System by function CNS (encased in bone) PNS Brain Spinal Chord Somatic System Autonomic System Parasympathetic Enteric Sympathetic Movement Coordination Receive external stimuli 8 cervical 12 thoraic 5 lumbar 5 sacral 11 coccygeal Cerrvical Spinal: C1-C4 Brachial Plexus: C5-T1 Lateral Chord: C5-C6 Posterior Chord: C6-C8 Medial Chord: C7-T1 Rest Relaxation Digestion Flight/Fight

Component Design Continued

Component Design continued

Material Selection Bone encasement for CNS -Neurons!!! Different classes/morphologies are generated through growth factors and stimulus -Myelin Sheath: For insulation and propagation of action potentials. 80% lipid 20% protein Myelin basic protein (MBP) Myelin Oligodendrocyte glycoprotein (MOG) Transmission electron micrograph of a myelinated neuron. Generated at the Electron Microscopy Facility at Trinity College, Hartford, CT

Bone encasement for CNS

-Neurons!!! Different classes/morphologies are generated through growth factors and stimulus

-Myelin Sheath: For insulation and propagation of action potentials.

80% lipid 20% protein

Myelin basic protein (MBP)

Myelin Oligodendrocyte glycoprotein (MOG)

Material Selection: Neurons Sensory Neuron - Converts external stimuli to electrical signals - Chemoreceptors (e.g. olfactory signals) - Mechanoreceptors (e.g. joint position detection) Motor Neuron: - Stimulated by interneurons (small feedback loops or from ANS/PNS) - Activates effectors (glands, muscles, ...) Interneuron: - Data processing, stimulated by: - sensory neurons - other interneurons or both. - many unknown types remain A: Cortical pyramidal cell - primary excitatory neurons of cerebral cortex B: Purjinke cell of cerebellum. Transmit output of cerebral cortex C: Stellate cell - provides inhibitory input to cerebral cortex Example neurons from the brain: In Out Process Basic building blocks of nervous system are neurons. Hundreds of different types, many uncatalogued. Three main categories:

Material Selection: Glia in the CNS - Structural & metabolic support (feed neurons) - Transmitter reuptake: express transporters for neurotransmitters - Regulate ion concentrations (potassium) - Act as immune cells of the nervous system - Responsible for myelin sheathing of axons - Single oligodendrocyte myelinates 10-15 axons - Modulates axon conduction speed Astrocytes Microglial cells Oligodendrocyte Approx. 3:1 ratio of glial cells to neurons in the brain Modulate signal propagation and neurotransmitter uptake at the synaptic cleft Provide scaffold for neural development, help in injury recovery

Facts & Figures on Neurons in Brains Neocortex Number of neocortical neurons (males) = 22.8 billion (Pakkenberg et al., 1997; 2003) Average number of neocortical glial cells (young adults ) = 39 billion (Pakkenberg et al., 1997; 2003) Brain Average number of neurons in the brain = 100 billion Average number of glial cells in brain = 10-50 times the number of neurons Cerebral cortex: Total number of synapses in cerebral cortex = 60 trillion (yes, trillion) * (from G.M. Shepherd, The Synaptic Organization of the Brain, 1998, p. 6). However, C. Koch lists the total synapses in the cerebral cortex at 240 trillion (Biophysics of Computation. Information Processing in Single Neurons, New York: Oxford Univ. Press, 1999, page 87). http://faculty.washington.edu/chudler/facts.html Brain requirements: Brain utilization of total resting oxygen = 20% Blood flow from heart to brain = 15-20% Blood flow through whole brain (adult) = 750-1000 ml/min Fault tolerance: 1 neuron dies each second in the brain Functioning humans Brain Brain Composition Whole Brain (%) Water 77 to 78 Lipids 10 to 12 Protein 8 Carbohydrate 1 Soluble organic substances 2 Inorganic salts 1 processing elements 10 14 synapses element size 10 -6 m energy use 30W processing speed ~100Hz

Neocortex

Number of neocortical neurons (males) = 22.8 billion (Pakkenberg et al., 1997; 2003)

Average number of neocortical glial cells (young adults ) = 39 billion (Pakkenberg et al., 1997; 2003)

Brain requirements:

Brain utilization of total resting oxygen = 20% Blood flow from heart to brain = 15-20% Blood flow through whole brain (adult) = 750-1000 ml/min

Fault tolerance:

1 neuron dies each second in the brain

Functioning humans

Materials Performance http://vadim.oversigma.com/MAS862/Project.html Number of neurons (adult)* 20,000,000,000 - 50,000,000,000 Number of neurons in cerebral cortex (adult)* about 20,000,000,000 (some sources have incorrect number 8,000,000) Number of synapses (adult) 1014 (2,000-5,000 per neuron) Weight Birth 0.3 kg, 1 y/o 1 kg, puberty 1.3 kg, adult 1.5 kg Power consumption (adult) 20-40 Watts (0.5-4 nW/neuron) Percentage of body 2% weight, 0.04-0.07% cells, 20-44% power consumption Genetic code influence 1 bit per 10,000-1,000,000 synapses Atrophy/death of neurons 50,000 per day (between ages 20 and 75) Sleep requirement (adult) average 7.5 hours/day or 31% Normal operating temperature 37 ± 2°C Maximum firing frequency of neuron 250-2,000 Hz (0.5-4 ms intervals) Signal propagation speed inside axon 90 m/s sheathed, <0.1 m/s unsheathed Processing of complex stimuli 0.5s or 100-1,000 firings

Design Specs System connections interacting in subnetworks to produce aggregate outputs from inputs New data is stored in forms of neural pathways, instead of a hard drive or RAM.

System connections interacting in subnetworks to produce aggregate outputs from inputs

New data is stored in forms of neural pathways, instead of a hard drive or RAM.

Production Method In humans, neuralation begins in the blastoderm stage Basic neuralation complete by 2 weeks The anterior segment of neural tube gives rise to: forebrain , midbrain , and the hindbrain . Structures patterned by Hox genes . Ion pumps increase the fluid pressure within the tube and create a bulge. Brain regions further divide into subregions. -The hindbrain divides into different segments called rhombomeres . -Neural crest cells form ganglia above each rhombomere. -The neural tube becomes the germinal neuroepithelium and serves as a source of new neurons during brain development. The brain develops from the inside-out.

In humans, neuralation begins in the blastoderm stage

Basic neuralation complete by 2 weeks

The anterior segment of neural tube gives rise to: forebrain , midbrain , and the hindbrain .

Structures patterned by Hox genes .

Ion pumps increase the fluid pressure within the tube and create a bulge.

Brain regions further divide into subregions.

-The hindbrain divides into different segments called rhombomeres .

-Neural crest cells form ganglia above each rhombomere.

-The neural tube becomes the germinal neuroepithelium and serves as a source of new neurons during brain development. The brain develops from the inside-out.

Prototype With the advent of verterbrates and chordates, fish are among those with the most ancestral brains Evolutionary changes of insect guts

With the advent of verterbrates and chordates, fish are among those with the most ancestral brains

“ How Did Nature Solve the information processing problems through the development of neural networks as well as the subsequent training and coordination with other networks?” Daniel Carlin, Daniel Cook, Joshua Mendoza-Elias Macro Micro Nano Macro Micro The Physical and Chemical Limitations that Nature Overcame were: solubility, dispersibility, surface area , conduction, signal modulation The Problem Nature HAS in Disease is: the severing of a neural network: paralysis, loss of sensation. Nano

Design Methodology: Define the function component to carry a load Material Selection Component Design Tentative component design Approximate stress analysis Tentative choice of material Assemble Materials Data Analysis of Materials Performance iterate Detailed Specifications and Design Choice of Production Methods Prototype Testing Establish Production Further Development iterate iterate iterate

Function of Neurons Two flavours: Electrical Synapse & Chemical Synapse Electrical Synapses are much faster. Not well characterized. Chemical Synapses well studied for 50 years. Well characterized. Neurochemicals Na + , K + , Cl - Absorption Reuptake Secretion Transcytosis Processing (reset) Establish Connections Reinforce connections that used “ Trim-out” connections that are weak Synthesis

Two flavours: Electrical Synapse & Chemical Synapse

Electrical Synapses are much faster. Not well characterized.

Chemical Synapses well studied for 50 years. Well characterized.

Function of Neurons Continued - ES & CS

Functions of Neurons Continued - ES Gap Junctions

Functions of Neurons continued - CS

Material Selection Carbohydrates Lipids Amino Acids

Carbohydrates

Lipids

Amino Acids

Assemble Materials OK, we need: Ion gradients to generate electric potential Insulation We need a pH gradient Mad vesicles everywhere!!! Neurochemical regulation (degradation/synthesis) Accessory Proteins All this encoded and regulated by Genome

OK, we need:

Ion gradients to generate electric potential

Insulation

We need a pH gradient

Mad vesicles everywhere!!!

Neurochemical regulation (degradation/synthesis)

Accessory Proteins

All this encoded and regulated by Genome

Component Design Nature said: -Need electric potentials  Ion Pumps -Need neurochemicals out and about:  Exocytosis/Endocytosis/Transcytosis  Intracellular trafficking network/  pH Insulation  Myelin Sheath Need long wires  axons (up to 2 meters in length) Need huge power source  many mitochondria

Nature said:

-Need electric potentials  Ion Pumps

-Need neurochemicals out and about:

 Exocytosis/Endocytosis/Transcytosis

 Intracellular trafficking network/  pH

Insulation  Myelin Sheath

Need long wires  axons (up to 2 meters in length)

Need huge power source  many mitochondria

Tentative Component Design OK, we need: High surface area at synapse Receptors of ligand binding Receptors for targeting to membrane Mechanism for loading/unloading neurochemicals Processing enzymes Empirical Method for assessing Performance and levels of neuroreceptors and neurochemicals.

OK, we need:

High surface area at synapse

Receptors of ligand binding

Receptors for targeting to membrane

Mechanism for loading/unloading neurochemicals

Processing enzymes

Empirical Method for assessing Performance and levels of neuroreceptors and neurochemicals.

Analysis of Materials Performance Well, it works for us. But assuming ligand depletion is neglible, the following mass balance equation can be written for surface receptors: Equilibrium Receptor Recycling Receptor Synthesis

Well, it works for us.

But assuming ligand depletion is neglible, the following mass balance equation can be written for surface receptors:

Established Micro Action Potential Model: Hodgkin-Huxley Set of non-linear differential equations describing the voltage dependance of the impedance of the neural membrane Impedance given by: Solving gives the voltage trace in time or space along a neuron axon This also gives the maximum firing rate, determined by the refractory period

Set of non-linear differential equations describing the voltage dependance of the impedance of the neural membrane

Impedance given by:

Solving gives the voltage trace in time or space along a neuron axon

This also gives the maximum firing rate, determined by the refractory period

Established Macro “Model”: Coupled Firing Problem: Only a few neurons can be observed at a time, not the whole ensemble Are the firing of two neurons related? Look at “firing trains,” the number of firings in a given time period represents a function of the neuron activity Cross-correlation of these function gives an idea of whether or not the neurons are related, and if they are in phase or not

Problem: Only a few neurons can be observed at a time, not the whole ensemble

Are the firing of two neurons related?

Look at “firing trains,” the number of firings in a given time period represents a function of the neuron activity

Cross-correlation of these function gives an idea of whether or not the neurons are related, and if they are in phase or not

Detailed Specifications & Design - Neurochemical Transport Cell membrane Small molecules Large molecules Active transport Targeting AA, peptide frags, Ions, H 2 O FA, aggregate complexes Clathrin COPI/II  pH Signal Transduction ATP Kinase Endocytosis Secretion Processing Diffusion

Choice of Production Methods Cells arise from previous cells. Genome encoded functions: -juxtracrine & paracrine signaling Germline eventually develops tissues and cellular domains. Neural tissues develop from ectoderm Ideally, we would like to work with neural precursor cells that have undetermined cell fates. .

Prototype Testing Metazoans - Eumetazoa e.g. Cnidaria, Ctnophora, Anthozoa, etc. The simplest neural networks may be found in the jellyfish. In terms of a simple system from humans, afferent and efferent innervation. For the purposes of neural networks, this system demonstrates the interaction of the many subnetworks: -CNS-PNS -Sensory-PNS -Motor-CNS

Metazoans - Eumetazoa

e.g. Cnidaria, Ctnophora, Anthozoa, etc.

The simplest neural networks may be found in the jellyfish.

In terms of a simple system from humans, afferent and efferent innervation.

For the purposes of neural networks, this system demonstrates the interaction of the many subnetworks:

-CNS-PNS

-Sensory-PNS

-Motor-CNS

“ How Did Nature Solve the information processing problems through the development of neural networks as well as the subsequent training and coordination with other networks?” Daniel Carlin, Daniel Cook, Joshua Mendoza-Elias Macro Micro Nano Macro Micro The Physical and Chemical Limitations that Nature Overcame were: solubility, dispersibility, surface area , conduction, signal modulation The Problem Nature HAS in Disease is: the severing of a neural network: paralysis, loss of sensation. Nano

Introduction Synapse Video:

Neurotransmitters Release Function Functions of Neurotransmitters Neurotransmitters are chemicals that are used to relay, amplify and modulate signals between a neuron and another cell. According to the prevailing beliefs of the 1960s, a chemical can be classified as a neurotransmitter if it meets the following conditions: (i.) It is synthesized endogenously, that is, within the presynaptic neuron; (ii.) It is available in sufficient quantity in the presynaptic neuron to exert an effect on the postsynaptic neuron; (iii.) Externally administered, it must mimic the endogenously-released substance; and (iv.) A biochemical mechanism for inactivation must be present. Functions of Vesicles A vesicle is a relatively small and enclosed compartment, separated from the cytosol by at least one lipid bilayer. If there is only one lipid bilayer, they are called unilamellar vesicles; otherwise they are called multilamellar. Vesicles store, transport, or digest cellular products and waste. Functions of Protein Pumps Ion protein pumps establish the electrochemical gradient required for the transmission of action potentials. Functions of Receptors Receptors are responsible for binding neurochemicals and causing propogation of the signal being transmitted by the synapse. Receptors are also responsible for the upregulation of Cellular Pathways that form the basis of “learning.” Functions of Cellular Pathways in LTP and LTD Once a receptor binds a neurochemical, many changes begin to occur within the neuron that determines if the connection will be reinforced or weakened. The timing, frequency, and strength of stimulus will determine which genetic programs are activated. In terms of learning, genes associated with long-term potentiation (LTP) “learning” or long-term depression “unlearning” will be activated.

Synapse Function - Protein Pumps/Channels Types of Proteins Establishing Ion Gradients: Ion transporters use energy (ATP) to establish electrochemical gradients. Ion Channels use diffusion down chemical gradients. Ion balance will “prime” the neuron and also determine parameters of time as firing cannot occur until the neuron is reset. Resetting results in a return in ion balance and neurotransmitter levels.

Synapse Function - Neurotransmitters Types of Neurotransmitters: Acetylcholine - voluntary movement of the muscles Norepinephrine - wakefulness or arousal Dopamine - voluntary movement and motivation, &quot;wanting&quot; Serotonin - memory, emotions, wakefulness, sleep and temperature regulation GABA (gamma aminobutyric acid) - inhibition of motor neurons Glycine - spinal reflexes and motor behaviour Neuromodulators - sensory transmission-especially pain *Neurotransmitters achieve neurotransmission by binding neuroreceptors.

Synapse Function - Vesicles Types of Macromolecules in Vesicles: V and T Snares: Proteins that facililtate vesicle fusion and release COPI/COPII/Clathrin: Protein responsible for targeting of vesicles. Neurotransmitters are loaded onto vesicles for release into the synaptic cleft.

Receptors are any membrane bound protein that binds a neurochemical and facilitates synaptic transmission. There is a large repertoire of receptors that bind neurochemicals in different ways and illicit different effects in the neuron. This is achieved by differences in: (i.) affinity for the neurochemical (ii.) specialized receptors that activate specific secondary mechanisms and/or cellular pathways. Synapse Function - Receptors A B C

Long-term potentiation (LTP) is the mechanism by which neural networks are “trained”. LTP: High fast peak rise in Ca+2  increase efficacy  Activates LTP second messengers, phosphorylation pathways (short-term)  increrase Ampa receptors  genetic pathways e.g. thicker axons (long-term) *All proteins involved in LTP have not been fully characterized. Synapse: The Process of Learning - LTP and LTD

Repeated stimulation causes changes and activation of LTP causes reinforcement of learning. Synapse: The Process of Learning cont’d - LTP and LTD

Synapse: The Process of Learning cont’d - LTP and LTD Long-term depression (LTD) is The mechanism by which neurons are “detrained.” LTD: Slow rise peak in Ca+2  decrease synaptic efficacy  Activated LTP secondary messengers (phosphatases) (short-term)  decrease Ampa Receptors  genetic pathways e.g. thinner axons (long-term). Similar short-term and long-term changes are observed as in LTP. *All of the proteins involved involved in LTD have not been fully characterized.

Timing is important for LTP and LTD. -Connections that are established by LTP are specific . -Connections may be associative in that “neurons that fire together, fire together.” Synapse: The Process of Learning cont’d - Spike Driven Plasticity

“ How Did Nature Solve the information processing problems through the development of neural networks as well as the subsequent training and coordination with other networks?” Daniel Carlin, Daniel Cook, Joshua Mendoza-Elias Macro Micro Nano Macro Micro The Physical and Chemical Limitations that Nature Overcame were: solubility, dispersibility, surface area , conduction, signal modulation The Problem Nature HAS in Disease is: the severing of a neural network: paralysis, loss of sensation. Nano LTP-KO: NMDA Receptor Agonist CPP No change synaptic efficacy No Learning Loss of memory

NMDA Receptor NMDAr works in concert with NMDA receptors and plays a critical role in LTP and LTD. Glutamate binding to the NMDAr relieves the Mg +2 block and allows Ca +2 to enter the postynaptic cleft.

NMDAr works in concert with NMDA receptors and plays a critical role in LTP and LTD.

Glutamate binding to the NMDAr relieves the Mg +2 block and allows Ca +2 to enter the postynaptic cleft.

NMDAr in LTP NMDAr (i.) Binds glutamate (ii.) Relieves Mg+2 block (iii.)Allows for characteristic rapid spike in Ca +2 levels that (iv.) cause LTP mechanisms to be activated.

Materials Data GRIN1 glutamate receptor, ionotropic, N-methyl D-aspartate 1 [ Homo sapiens ] Gene location: 9q34.3 Genomic Region: 30.366 kb (3 isoforms) mRNA transcript: 4265 bp Length: 901 aa long; weight: 99.17 kDa MSTMRLLTLALLFSCSVARAACDPKIVNIGAVLSTRKHEQMFRE AVNQANKRHGSWKIQLNATSVTHKPNAIQMALSVCEDLISSQVYAILVSHPPTPNDHF TPTPVSYTAGFYRIPVLGLTTRMSIYSDKSIHLSFLRTVPPYSHQSSVWFEMMRVYSW NHIILLVSDDHEGRAAQKRLETLLEERESKAEKVLQFDPGTKNVTALLMEAKELEARV IILSASEDDAATVYRAAAMLNMTGSGYVWLVGEREISGNALRYAPDGILGLQLINGKN ESAHISDAVGVVAQAVHELLEKENITDPPRGCVGNTNIWKTGPLFKRVLMSSKYADGV TGRVEFNEDGDRKFANYSIMNLQNRKLVQVGIYNGTHVIPNDRKIIWPGGETEKPRGY QMSTRLKIVTIHQEPFVYVKPTLSDGTCKEEFTVNGDPVKKVICTGPNDTSPGSPRHT VPQCCYGFCIDLLIKLARTMNFTYEVHLVADGKFGTQERVNNSNKKEWNGMMGELLSG QADMIVAPLTINNERAQYIEFSKPFKYQGLTILVKKEIPRSTLDSFMQPFQSTLWLLV GLSVHVVAVMLYLLDRFSPFGRFKVNSEEEEEDALTLSSAMWFSWGVLLNSGIGEGAP RSFSARILGMVWAGFAMIIVASYTANLAAFLVLDRPEERITGINDPRLRNPSDKFIYA TVKQSSVDIYFRRQVELSTMYRHMEKHNYESAAEAIQAVRDNKLHAFIWDSAVLEFEA SQKCDLVTTGELFFRSGFGIGMRKDSPWKQNVSLSILKSHENGFMEDLDKTWVRYQEC DSRSNAPATLTFENMAGVFMLVAGGIVAGIFLIFIEIAYKRHKDARRKQMQLAFAAVN VWRKNLQSTGGGRGALQNQKDTVLPRRAIEREEGQLQLCSRHRES

GRIN1 glutamate receptor, ionotropic, N-methyl D-aspartate 1 [ Homo sapiens ]

Gene location: 9q34.3 Genomic Region: 30.366 kb (3 isoforms)

mRNA transcript: 4265 bp

Length: 901 aa long; weight: 99.17 kDa

MSTMRLLTLALLFSCSVARAACDPKIVNIGAVLSTRKHEQMFRE AVNQANKRHGSWKIQLNATSVTHKPNAIQMALSVCEDLISSQVYAILVSHPPTPNDHF TPTPVSYTAGFYRIPVLGLTTRMSIYSDKSIHLSFLRTVPPYSHQSSVWFEMMRVYSW NHIILLVSDDHEGRAAQKRLETLLEERESKAEKVLQFDPGTKNVTALLMEAKELEARV IILSASEDDAATVYRAAAMLNMTGSGYVWLVGEREISGNALRYAPDGILGLQLINGKN ESAHISDAVGVVAQAVHELLEKENITDPPRGCVGNTNIWKTGPLFKRVLMSSKYADGV TGRVEFNEDGDRKFANYSIMNLQNRKLVQVGIYNGTHVIPNDRKIIWPGGETEKPRGY QMSTRLKIVTIHQEPFVYVKPTLSDGTCKEEFTVNGDPVKKVICTGPNDTSPGSPRHT VPQCCYGFCIDLLIKLARTMNFTYEVHLVADGKFGTQERVNNSNKKEWNGMMGELLSG QADMIVAPLTINNERAQYIEFSKPFKYQGLTILVKKEIPRSTLDSFMQPFQSTLWLLV GLSVHVVAVMLYLLDRFSPFGRFKVNSEEEEEDALTLSSAMWFSWGVLLNSGIGEGAP RSFSARILGMVWAGFAMIIVASYTANLAAFLVLDRPEERITGINDPRLRNPSDKFIYA TVKQSSVDIYFRRQVELSTMYRHMEKHNYESAAEAIQAVRDNKLHAFIWDSAVLEFEA SQKCDLVTTGELFFRSGFGIGMRKDSPWKQNVSLSILKSHENGFMEDLDKTWVRYQEC DSRSNAPATLTFENMAGVFMLVAGGIVAGIFLIFIEIAYKRHKDARRKQMQLAFAAVN VWRKNLQSTGGGRGALQNQKDTVLPRRAIEREEGQLQLCSRHRES

Detailed Specifications and Design

Detailed specs/design Structure: Alpha helix and Beta sheet motifs create a binding pocket of the NMDA NR1 ligand-binding core (center)

Structure:

Alpha helix and

Beta sheet motifs create a binding pocket of the NMDA NR1 ligand-binding core (center)

Details specs/design: Ligand Binding Core Pocket Hydrogen Bonding (dotted black lines) is responsible for binding glutamate in S1S2 ligand binding core at Residues: 405, 516, 518, 523, 732, 731. Water molecules (green) also make an important contribution to binding.

Hydrogen Bonding (dotted black lines) is responsible for binding glutamate in S1S2 ligand binding core at Residues: 405, 516, 518, 523, 732, 731. Water molecules (green) also make an important contribution to binding.

Details Specs/design: Binding affinity to different ligands Data on binding affinities and the ligand binding core can help in the understanding of the kinetics.

Data on binding affinities and the ligand binding core can help in the understanding of the kinetics.

Also use AFM to give structural and functional information J. Struct. Biol., 142 , 369 (2003). Curr. Op. Struct. Biol. , 16 , 489 (2006). Neurobiol. Aging, 27 , 546 (2006). J. Cell. Biochem., 97 , 1191 (2006).

Also use AFM to give structural and functional information

*What* GRIN1 do? What does a neuroreceptor need need? Neurochemical binding pocket Transmembrane domains Ion channel or catalytic cytosolic domain GRIN1 has S1S2 ligand binding core composed of two domains (red and blue) GRIN1 has transmembrane domains (loops 1-3) that anchor it to the postynaptic cleft Selectivity Filter: The transmembrane domains also encase a channel that discriminates based on size allowing Ca +2 through but not Mg +2

What does a neuroreceptor need need?

Neurochemical binding pocket

Transmembrane domains

Ion channel or catalytic cytosolic domain

GRIN1 has S1S2 ligand binding core composed of two domains (red and blue)

GRIN1 has transmembrane domains

(loops 1-3) that anchor it to the postynaptic cleft

Selectivity Filter: The transmembrane

domains also encase a channel that

discriminates based on size allowing Ca +2

through but not Mg +2

Prototypes through time Sodium/Potassium Pumps that are involved in action potential propogation Voltage-Gated: Do not depend on binding ligands. -Transmembrane domain -Selectivity filter J.Mol.Biol. , 284 , 401 (1998).

Sodium/Potassium Pumps that are involved in action potential propogation

Voltage-Gated: Do not depend on binding ligands.

-Transmembrane domain

-Selectivity filter

Evolutionary tree of GRIN1 http://www.ncbi.nlm.nih.gov/blast/treeview/blast_tree_view.cgi?request=page&rid=SSRCRD3F013&dbname=nr&queryID=lcl|30291

That reminds me of…

“ How Did Nature Solve the information processing problems through the development of neural networks as well as the subsequent training and coordination with other networks?” Daniel Carlin, Daniel Cook, Joshua Mendoza-Elias Macro Micro Nano Macro Micro The Physical and Chemical Limitations that Nature Overcame were: solubility, dispersibility, surface area , conduction, signal modulation The Problem Nature HAS in Disease is: the severing of a neural network: paralysis, loss of sensation. Nano LTP-KO: NMDA Receptor Agonist CPP No change synaptic efficacy No Learning Loss of memory

Remember the role of NMDAr in LTP?

Loss of function: NMDA Receptor Antagonist Knocking out LTP Agonist - def - An antagonist is a chemical agent that is able to prevent the normal functioning and of a neuroreceptor. CPP ( 3-(2-Carboxypiperazin-4-yl)propyl-1-phosphonic acid ) blocks the normal functioning of the NMDA receptor and by causing loss of function of the NMDA receptor. The NMDAr is an important part component in the mechanisms of LTP and LTD. (C.) Loss of LTP causes the inability of neurons to create synaptic pathways (attractors) that will store information. Meanwhile LTD will cause information degradation by “detraining” established synaptic pathways. The result is a loss of memory. (A.) Loss of LTD causes the inability to degrade information. (B.) Normal functioning of LTP and LTD in concert Nature Neuroscience . 5 (1): 48-52 (2002) Nature Neuroscience . 5 (1): 6-8. (2002)

Recovery of CPP Antagonist KO effect The antagonist effect of CPP can be compensated by increasing the amount of NMDA receptors, increasing the amount of AMPA receptors, or a combination of both. In addition, overtime receptors will be replaced and new ones that have not bound CPP will begin to compensate. The process of replacement can also be accelerated by using monoclonal antibodies against CPP to aid in elimination of CPP bound NMDA receptors from the body. Increase NMDAr expression Increase AMPAr expression Antibodies against CPP Time

CPP binds to the S1S2 ligand binding core by Does not allow conformational changes that wil allow Ca+2 to enter the postynaptic cleft. NMDAr looses ability to generate rapid Ca +2 spikes. AMPAr do not activate cellular mechanisms Long-term effects in gene expression Synaptic efficacy does not increase. Synapses shrink or stay the same. Micro: Changes in Cell Morphology are no longer possible

CPP binds to the S1S2 ligand binding core by

Does not allow conformational changes that wil allow Ca+2 to enter the postynaptic cleft.

NMDAr looses ability to generate rapid Ca +2 spikes.

AMPAr do not activate

cellular mechanisms

Long-term effects in gene expression

Synaptic efficacy does not

increase. Synapses shrink

or stay the same.

Macro: Insulin Resistance Syndrome No synaptic pathways (attractors) can be made. No memories. No new information or skills can be formed or gained that are not instinctual. But its great if you finsihed studying for a test, and don’t want LTD to make you forget what you studies for.

No synaptic pathways (attractors) can be made.

No memories.

No new information or skills can be formed or gained that are not instinctual.

But its great if you finsihed studying for a test, and don’t want LTD to make you forget what you studies for.

Foreward Engineering: Building a “Wet Drive” Daniel Calrin B.S.E. Daniel Cook Joshua Mendoza-Elias

Background: Neurons We already know the LTP and LTD are the basis of forming synaptic pathways and memories. A complete molecular description of cellular signalling mechanisms are not fully characterized, BUT we still know basically what happens and can abstract this and represent this in a model. So can we train neurons an in silico model that approximates neural networks i n vivo? Can we approve upon this model Can we use this biologically inspired devices?

We already know the LTP and LTD are the basis of forming synaptic pathways and memories.

A complete molecular description of cellular signalling mechanisms are not fully characterized,

BUT we still know basically what happens and can abstract this and represent this in a model.

So can we train neurons an in silico model that approximates neural networks i n vivo?

Can we approve upon this model

Can we use this biologically inspired devices?

Short-term and Long-term Effects

Background continued: The mechanisms of Long-term Depression

The mechanisms of Long-term Depression

The Basis of Hebbian Learning

Foundation for our Computer Model

Types of Neuron Inhibitory Neurons Input Neuron Output Neuron Neuron is voltage-clamped, presynaptic to all neurons in model Inhibitory neurons depress post-synaptic neurons Excitatory Neurons Average firing rates solved at each time step Learning rule determines change in synaptic strength inhibitory synapse excitatory synapse Key: 1 2 3 4 α = - 1 4

Synaptic strengths 1.0 1.0 t=t 0 t=t 0+1 + α In phase Out of phase + α Two neurons are firing full-speed: Strengths increase by factor of alpha 1.0 0.0 - β One neuron v i is firing but v j is not: Strengths decrease by factor of beta - β v i v j v i v j v i v j v i v j

Inhibitory Neurons t=t 0 t=t 0+1 Excitatory Inhibitory = ...+ w i,j v i +... = ...- w i,j v i +... An inhibitory neuron v i is firing, depressing the post-synaptic neuron Weighted v i is summed negatively into v j Weighted v i is summed positively into v j An excitatory neuron v i is firing, potentiating the post-synaptic neuron v i v j v j v i v i v j v i v j

In-phase & out-of-phase components, but we could not teach the model complete phasic inversion Need further development to do this: one-way connections (i.e. some strengths are 0) Results

In-phase & out-of-phase components, but we could not teach the model complete phasic inversion

Need further development to do this: one-way connections (i.e. some strengths are 0)

Phase components Learning rule: α = | v 1 - v N | Input Neuron Output Neuron Convergence In phase component Out of phase component time firing rate

Learning rule: α = | v 1 - v N |

Output Neuron Maxima & Minima Local max near minimum Local min near maximum Maximum at maximum time firing rate Input Neuron

Further developments Short-term Sparse synapse matrix (i.e. some synapses are strength 0) Asynchronous firing Multi-dimensional training (i.e. for character recognition, sound recognition, etc.) Long-term Ca +2 Modeling Gene Expression Profile (DNA microarray data to reflect changes in synaptic efficacy)

Short-term

Sparse synapse matrix (i.e. some synapses are strength 0)

Asynchronous firing

Multi-dimensional training (i.e. for character recognition, sound recognition, etc.)

Long-term

Ca +2 Modeling

Gene Expression Profile (DNA microarray data to reflect changes in synaptic efficacy)

Biological parallel with In Silico Starvoytov et al. 2005 Light-drected stimulation of neurons on silicon wafers. J Neurophysiol 93 : 1090-1098.

LDS in concert with Computer Simulation MEAs vs. LDS More real-time data More quickly Scans: Works on variably connected neural networks

MEAs vs. LDS

More real-time data

More quickly

Scans: Works on variably connected neural networks

MEAs Neural Network controls Robot Redwood Center for Theoretical Neuroscience Inaugural Symposium UC Berkeley, October 7, 2005 Keywords: Theoretical Neuroscience; Brain; Machine Learning; Vision; Robotics Contact Information: Kilian Koepsell , Redwood Center for Theoretical Neuroscience, University of California, Helen Wills Neuroscience Institute, 132 Barker, MC #3190, Berkeley, CA 94720-3190 This research suggests that further model refinement, a software interface can be made that will allow Hybrots with larger networks greater sophistication and ability. Hybrots Video: Hybrots Info Webpage:

Redwood Center for Theoretical Neuroscience Inaugural Symposium UC Berkeley, October 7, 2005

Keywords: Theoretical Neuroscience; Brain; Machine Learning; Vision; Robotics

Contact Information: Kilian Koepsell , Redwood Center for Theoretical Neuroscience, University of California, Helen Wills Neuroscience Institute, 132 Barker, MC #3190, Berkeley, CA 94720-3190

This research suggests that further model refinement, a software interface can be made that will allow Hybrots with larger networks greater sophistication and ability.

Acknowledgements: Dr. David Needham Dr. Ragachavari Dr. Nestor

Dr. David Needham

Dr. Ragachavari

Dr. Nestor

Add a comment

Related presentations

Related pages

Bow Tie 1 Fashion Fabric Pattern - Documents

Crocheted Bow Tie Pattern Crocheted Bow Tie Materials - Deborah Norville - Size US-G/4.00mm Serenity Sock Yarn Abbreviations -st=S titch -Rnd=Round ...
Read more

Self Tie Bow Tie: A Guide For Pink Bow Ties - Documents

Self Tie Bow Tie: A Guide For Pink Bow Ties. by carmats799
Read more

Crocheted Bow Tie Pattern - Documents - docslide.us

Download Crocheted Bow Tie Pattern. Transcript ...
Read more

Bow Ties | Formal Mens Bow Ties and Bowties | The Tie Bar

With most bow ties below $20, The Tie Bar offers premium quality at a great value. ... SHOP BY BOW TIE COLOR. Black Bow Ties; Silver Bow Ties; White Bow Ties;
Read more

Bow Tie Television

NEP Bow Tie was formed in 1990. Today our business is diverse and service based. Working on behalf of Governments, ... © 2016 BOWTIE TELEVISION LIMITED |
Read more

Rose gold iced out bow tie shamballah box chain adjustable ...

Rose Gold Iced Out Bow Tie Shamballah BoxChain Adjustable Bracelet Price : ...
Read more

PowerPoint Presentation - Trailering - United States Army

... tie downs, winch and bow restraint Jack stand and ... and slowly put on trailer Attach winch cable to bow eye and ... Presentation Author:
Read more

The History of Bowtie - CGE Risk

History. It is said that the first ‘real’ Bowtie diagrams appeared in the (Imperial Chemistry Industry) course notes of a lecture on HAZAN ...
Read more