Week 1:

• Understanding brain using computational models
  • Descriptive models of brain
    • Encoding (Study how neurons respond to stimuli) and Decoding (Important in field of brain computer interface) model of brain
  • Mechanistic models of brain
    • Simulate behaviour of single neuron
    • Simulate network of neurons
  • Interpretive or Normative model of brain
    • Understanding why the brain works the way it works
    • Study computational principles underlying a neuron
• Computational Neuroscience is the study of how brain generates behaviours
  • Characterising what nervous systems do ie descriptive models
  • Determining how they function ie mechanistic models
  • Understanding why they operate in particular ways ie interpretive models
• Receptive Fields
  • Hubel and Wiesel Experiment with cats
  • Specific properties of a sensory stimulus that generate a strong response from a cell
• Descriptive models
  • Retinal Ganglion cells
  • Lateral Geniculate Nucleus
  • Center Surround receptive fields in retina
    • On center off surround
    • Off center on surround
  • Primary Visual Cortex
  • Cortical Receptive Fields ie oriented receptive fields
• Mechanistic models
  • Oriented receptive fields from center surround receptive fields
  • Multiple converging LGN cells info is captured by V1 cell
  • Above Hubel and Wiesel model does not take into account recurrent connections
• Interpretive models
  • Computational Advantages of receptive fields
  • Effiecient Coding Hypothesis
  • Image as linear combination of receptive fields
  • Optimization problem of minimizing squared pixel wise errors in reconctructed and actual images
  • Random RF and run efficient coding methods on images
  • Efficient Coding algos
    • Sparse Coding
    • ICA
    • Predictive Coding
• Neurons Doctrine
  • Neuron is fundamental structural and functional unit of brain
  • Majority are discrete cells
  • Info flows from dendrites to axons via cell body
  • EPSP (Excitatory Post Synaptic Potential)
  • When Summation of EPSP’s surpases threshold then we have action potential
  • Node of Ravier
  • Leaky bag of charged liquid with a cell memberane made of lipid bilayer
  • Ionic Channels are proteins which are selective and embedded in memberanes to allow ions to flow in or out
  • Resting Potential of -70 mV because Na and Cl conc. is higher outside and K and organic Anions conc. higher inside
  • Ionic Pump maintains the potential by explelling K and allowing Na inside
  • Ionic channels are gated
    • Voltage gated
    • Chemically gated eg synapses
    • Mechanically gated
  • Gated channels allow neuronal signaling
    • Chemically gated channels change local membrane potential then voltage gated channels open/close
    • Depolarization(+ve change in voltage)
    • Hyperpolarization(-ve change in voltage)
  • Action Potential
    • Strong depolarization opens Na channels causing Na influx, more channels open and then they inactivate
    • Outflux of the K ions and then K channels also close
  • Myelination of Axons
    • Ensure flat long range spike communication
    • Saltatory Conduction is when action potential hops from one non-myelinated region (Node of Ravier) to another and ensures lossless signal propagation
    • In case of MS, the cell loses the ability to send action potentials