Exercise 4 - Modeling and Simulating Biological Neural Networks

Question 1

Goal of neuroscience

Answer 1

the ultimate goal of neural science is to understand how the flow of electrical signals through neural circuits gives rise to mind - to how we perceive, act, think, learn, and remember

Question 2

computational neuroscience

Answer 2

mathematical tools and theories are used to investigate brain function.

Question 3

Main components of a neuron

Answer 3

• Soma: cell body containing the nucleus
• Dendrites: receive electrical impulses from other neurons through synapses
• Axon: emerges from the soma at the axon hillock and conducts electrical impulses to other neurons.

Question 4

polarity of a neuron meaning

Answer 4

number of axons and dendrites

Question 5

Into what can neurons be classified by their functionality?

Answer 5

sensory neurons, interneurons, and motor neurons

Question 6

resting membrane potential meaning

Answer 6

• in the idle state of the neuron, the sum of the different ion flows between the inside and the outside of the neuron results in an equilibrium state

Question 7

steps in action potential

Answer 7

• depolarization: positive charge flows into the neuron and membrane potential increases
• action potential: membrane voltage quickly increases to a peak, the action potential
• hyperpolarization: membrane voltage falls back below the resting state

Question 8

two phases of refactory period

Answer 8

• absolute refactory period: neuron cannot emit any spike

- relative refactory period: eliciting an action potential requires strong stimuli

Question 9

Two different models propagating action potentials

Answer 9

• Electrotonic conductance (short distance, fast)

- active regeneration of action potentials (long distance, slow)

Question 10

how to enhance neural signal transmission?

Answer 10

saltatory conduction of myelinated segments

Question 11

4 properties of an action potential

Answer 11

threshold
all-or-none event
propagation (propagation over long distances through self-regeneration)
refractory period

Question 12

types of synapses

Answer 12

electrical

chemical with synaptic cleft

Question 13

How are the cells in the brain called?

Answer 13

glial cells have a supporting function

Question 14

Two types of glial cells

Answer 14

Microglia: part of immune system
Macroglia: formation of myelin sheath around axons

Question 15

detailed compartmental models

Answer 15

capture the detailed neuron morphology based on anatomical reconstructions

Question 16

reduced compartmental models

Answer 16

• comprised of only a few dendritic compartments

Question 17

single-compartment model

Answer 17

• completely ignores the morphology of the neuron.

Question 18

cascade model

Answer 18

purely functional models that only represent abstract computations and no biological detail

Question 19

black-box models

Answer 19

consider the neuron as an input/output system

Question 20

McCulloch-Pitts Model

Answer 20

• first computational neuron model
• inspired by the all-or-none property of biological neurons.
• if no inhibitory synapse is active and the sum of active excitatory synapses crosses a threshold, the output of the neuron is 1

Question 21

What was developed after the McCulloch-Pitts model?

Answer 21

The general analog neuron model. The simple McCulloch-Pitts model was generalized by adding synaptic weights and an activation function
A(w1x1… + b)

Question 22

Common activation functions

Answer 22

Binary step function
logistic function
tanh
ReLu
ELU = exponential linear unit
Gaussian

Question 23

what are analog neuron models not able to reproduce?

Answer 23

temporal dynamics of biological neurons.

Question 24

How do spiking neuron models represent biological neurons?

Answer 24

• cell membrane as capacitor
• flows of ions are electrical currents
• ion channels = resistor

Question 25

different types of spinking neurons

Answer 25

regular spikes
instrinsically bursting
chattering
fast spiking
thalamo-cortical
low-threshold spiking

Question 26

Hodgkin-Huxley Model

Answer 26

• model contains a single compartment
• represents electrical properties
• computationally complex

Question 27

Difference between Hodgkin-Huxley Model and Leaky Integrate-and-Fire model?

Answer 27

Leaky Integrate-and-Fire Model is a phenomenological neuron model that captures the overall behavior of a biological neuron but does not show the underlying dynamics.

Question 28

AdEx Model

Answer 28

• Extension of the Leaky Integrate-and-Fire model and combines it with the Hodgkin-Huxley model to produce biologically realistic firing patterns.

Question 29

Types of encoding in spike trains

Answer 29

• rate encoding = averaged spike rate within a certain time window
• time-to-first spike coding = information is encoded in the time of emission of the first spike after a reference point.
• coding by synchrony or correlation: population codes are based on the activity patterns of a group of neurons: information is encoded as the averaged overall activity of the population.

Question 30

Model in spiking neurons vs. analog neurons

Answer 30

dynamical system, activation function

Question 31

Communication in spiking neurons vs. analog neurons

Answer 31

Digital spikes, analog values

Question 32

encoding of information in spiking neurons vs. analog neurons

Answer 32

rate, time-to-first spike, synchrony ; real numbers

Question 33

computational complexity in spiking neurons vs. analog neurons

Answer 33

high, low

Question 34

biological plausibility in spiking neurons vs. analog neurons

Answer 34

high, low