Applied Neuroscience

ing 2017Introduction to Applied Neuroscience

Logistics1. Maximum of 4 Absences2. For an excused absence, e-mailshpattendance@columbia.edu and ng2410@columbia.edu3. Two 10 minute breaks at 11:00 AM and 12:00 PM4. Joe’s Coffee in Northwest Corner Building5. Do not hesitate to ask questions6. Do request topics of interest for course7. No exams

Introduction to Applied NeuroscienceObjective: Fundamentals of NeuroscienceAgenda:1. Logistics2. Computational Neuroscience3. Neurobiology NameGradeLocationWhat is something you wouldlike to learn in this class? Share one interesting thingabout yourself.

Philosophical Questions on Computers1. What is intelligence? What is thought?2. Are these functions that a machine can have?3. If machines can display thought or intelligence, does thisimply that human cognition is a type of computationalability?4. If human cognition is a computation, does this imply thatthe human mind is a W. eConceptualNotation1815-1864GeorgeBooleThe Laws ofThought1912-1954Alan TuringTuringMachine1906-1978Kurt GödelFormallyUndecidablePropositions

Introduction to Applied Neuroscience Descriptive Models of the Brain How is information about the external world encodedin neurons and networks? How can we decode neural information? Mechanistic Models of Brain Cells and Circuits How can we reproduce the behavior of a singleneuron in a computer simulation? How do we model a network of neurons? Interpretive Models of the Brain Why do brain circuits operate the way they do? What are the computational principles underlying theiroperation?

Course Objectives1. To be able to quantitatively describe what a givencomponent of a neural system is doing based onexperimental data2. To be able to simulate on a computer the behavior ofneurons and networks in a neural system3. To be able to formulate specific computational principlesunderlying the operation of neural systems

Introduction to Applied NeuroscienceHow can we reproduce the behavior of a single neuron in acomputer SimulationSoftware AppliedNeuroscienceComputational neuroscienceprovides tools and methods for“characterizing what nervoussystems do, determining howthey function, and understandingwhy they operate in particularways” (Dayan and Abbott)1. Description Models (What)2. Mechanistic Models (How)3. Interpretive Models (Why)Professor Larry AbbottCenter for Theoretical NeuroscienceColumbia University

An Example: Cortical Receptive FieldsWhat is the receptive field of a brain cell?Classical Definition: the region of sensory space thatactivates a neuron (Hartline, 1938)Example: Region of the retina where a spot of lightactivates a retinal cellLet’s look at:I. A Descriptive Model ofReceptive FieldsII. A Mechanistic Model ofReceptive FieldsIII. An Interpretive Model ofReceptive Fields

I. Descriptive Model of Receptive FieldsOutputResponses (SpikeTrains) from aRetinal GanglionCell

II. Mechanistic Model of Receptive FieldsThe Question: How are receptive fields constructed using theneural circuitry of the visual cortex?Model suggested by Hubel andWiesel in 1960s: V1 ReceptiveFields are created fromconverging LGN inputs

III. Interpretive Model of Receptive FieldsThe Question: Why are receptive fields in V1 shaped in thisway?What are the computational advantages of such receptivefields?Computational Hypothesis: How can theimage I be represented as faithfully andefficiently as possible using neurons withreceptive fields RF1, RF2, RFn

The Human BrainEmily Dickinson

Major Brain Regions: Brain Stem and CerebellumCerebellum:Coordination ofvoluntary movementsand sense of equilibriumPons:Connects brainstem withcerebellum and involvedin sleep and arousalMedulla:Breathing, muscle tone,and blood pressure

Major Brain Regions: Midbrain and ReticularFormationMidbrain:Eye movements, visual,and auditory reflexesReticular Formation:Modulates musclereflexes, breathing, andpain perception.Regulates sleep,wakefulness, andarousal.Not anatomically wellMidbrain defined (set of nuclei inbrainstem).

Major Brain Regions: Thalamus and HypothalamusThalamus:“Relay station” for allsensory information(except smell) to thecortexHypothalamus:Regulates basic needsincluding fighting,fleeing, feeding, andmatingHypothalamus

Major Brain Regions: Cerebral HemispheresConsists of: cerebralcortex, basal ganglia,hippocampus, andamygdalaInvolved in:Perception and motorcontrol, cognitivefunctions, emotion,memory and learningCorpus collosum

The NeuronNeo-cortex:Part of cerebralcortex concernedwith sight andhearing in mammals,regarding as the siteof higher intelligenceThe neo-cortex hassix layers of tissue.Pyramidal neuron:Primary componentof cortical tissue andnamed for triangularcell body (soma)

The Neuron Doctrine“The neuron is the appropriate basis forunderstanding the computational andfunctional properties of the brain” (1891)

The Neuron

Properties of a Neuron Contents of the neuronenclosed within a cellmembrane, which is a lipidbilayer The bilayer is impermeableto charged ions Each neuron maintains apotential difference acrossits membrane Inside is -70 to -80 mVrelative to outside Ionic pump maintains 70 mV difference byexpelling Na out andallowing K ions in

Electrophysiology of a NeuronNernst EquationE Membrane Potential atwhich current flow due todiffusion of ions is balanced byelectric forcesRT æ [outside] öE ln ç zF è [inside] øCell Membrane[Na ], [Cl-], [Ca2 ]OutsideInside[K ][K ], [A-][Na ], [Cl-], [Ca2 ]

Membrane Proteins: The Gatekeepers Properties in membranesact as pores or channelsthat are ion-specific Ionic channels are gated Voltage-gated:Probability of openingdepends on membranevoltage Chemically-gated:Binding to a chemicalcauses channel to open(neurotransmitters) Mechanically-gated:Sensitive to pressure orstretch (sensoryneurons)

Neuronal Signaling Different types of gatedchannels are involved inneuronal signaling Graded Potentials: travelover short distances andare activated by theopening of mechanicallyor chemically gatedchannels Action Potentials: travelover long distances andare generated by theopening of voltage-gatedchannels

Action Potential Depolarization: adecrease in thepotential differencebetween the insideand outside of thecell Hyperpolarization: anincrease in thepotential differencebetween the insideand outside of thecell Repolarization:returning to theresting membranepotential from eitherdirection

Action Potential A “threshold amount” ofdepolarization triggers an actionpotential.All-or-noneMore voltage-gated channelsopen and more Na ions enter.This continues until themembrane potential reaches theNa equilibrium potential of 50mV.As the inside of the cell becomesmore positive, voltage-gated K channels open.K moves out and the restingpotential is restored.

Action Potential PropagationThe action potential is propagated along the axon of theneuronVoltage-gatedsodium channelsVoltage-gatedpotassium channels

Neural Circuits in Drosophila Larval BrainFigure (Top)Z- Projections of Confocal Stacks of Larval Nervous Systems Illustrating Selected Expression Patterns in the InSITE CollectionA and B. Extremely sparse lines with expression in (A) the brain and (B) the abdominal neurons. C and D. Lines with expression in clusters of postembryonic neurons (pc). Arrows show sparse expression pattern. E. Strong expression in mushroom bodies (MB) and abdominal interneurons (in). F.Expression in the optic lobes (OL) and post-embryonic neurons. Abd, abdomen; Br, brain; Tx, thorax.Figure (Bottom)Examples of Glial Expression in the InSITE CollectionA. Perineurial and subperineurial glia B. Ensheathing glia. C. Astrocyte-like glia D. Midline Glia. Abd, abdomen; Br, brain; Tx, thorax.

Lineage Lines in Drosophila Larval BrainFigureExamples of Cell-Type Expression in Lineage Groups in the InSITE CollectionA, D, and E. All of the members of the post-embryonic lineages express GFP from the NBs to the oldest neurons.B. Expression in NBs and GMCs. C. Lineage expression exclusive to brain and thorax. NB, neuroblast; GMC;ganglion mother cell.

Introduction to Image RegistrationWhat is registration?Registration is the determination of geometrical transformation thataligns points in one view of an object with corresponding points inanother view of that object or another object“View” refers to the physical arrangement of an object in space (a3-D or 2-D Image)Inputs of Image Registration: 2 Views to be Registered Target/Source Fixed/MovingOutput of Image Registration: Geometrical Transformation(a mathematical mapping of points in one view to points in thesecond view)

Example of Image RegistrationNote: Image Fusion is not the same as Image Registration

Image Registration to Standardized TemplateFigure(A, A’) and (D,D’) Correction for Bending Larval Body.(B,B’) and (C,C’) Correction for Mushroom Body Size. Arrows indicate regions for correction.

Identified Issues in Image RegistrationFigureA. Expression pattern larger than template (Scaling Issue)B. Expression pattern smaller than template (Scaling Issue)C. Expression pattern twisted (Motion Correction Issue)Arrows indicate regions of misalignment.

Scaling Correction Method in Image RegistrationFigureScaling Correction Method by Manual B-Spline Deformation

Motion Correction Method in Image RegistrationFigureMotion Correction Method by Manual B-Spline Deformation

Dendrite Morphology AnalysisGraph Theory in NeuroscienceA dendritic tree is represented by a set ofnodes connected by edges. Branch orderbegins at the root (node with an index of 1).All edges lead away from the root. Thisdefines their directionality uniquely.

Dendrite Morphology AnalysisDendritic Tree Area: area circumscribed by convex hullTotal Dendritic Length: sum of all dendritic segments identified in askeletonization of the arborTotal Branch Point Number: sum of branch points identified in askeletonized rendition of the arbor

Sholl AnalysisSholl Analysis:Technique to describeneuronal arbors and quantifymorphological complexityHow does it work?Sholl Analysis creates aseries of concentric shellsaround a neuronal arbor,and counts how many timesconnected voxels definingthe arbor intersect thesampling shells.

Motivation to Modify Sholl AnalysisA. Binary Image of Drosophila Class IV neuronB. A Sholl plot visualization showing the number of intersections of thedendritic tree with circles of increasing radius from the center of thedendritic arborC. A Modified Sholl plot visualization using concentric irregularpolygons

Results of Modified Sholl AnalysisAnalysis of Paclitaxel TreatedClass IV Drosophila sensoryneurons

Results of Modified Sholl AnalysisAnalysis of Class IV Drosophila sensory neurons

Next Time:Single-Neuron Models

Introduction to Applied Neuroscience Objective: Fundamentals of Neuroscience Agenda: 1. Logistics 2. Computational Neuroscience 3. Neurobiology Name Grade Location What is something you would like to learn in