Help shape GoConqr. What should we add next?

Flashcards

Neural Networks

S. Satyanarayan Nayak Modelling and SImulation English 16 cards

Firing rate model mathematical representation ?

Cards in this deck (16)

  1. 1What is Hodgkin and Huxley model ?It was an experiment to determine the elector-chemical details which is responsible for the functioning of neurons (mainly spiking).
    A giant Squid was used for the experiment.
    The Action potential(spike) of the neuron was measured (Membrane-current = Ionic-current + CapactiveCurrent)
    Action potential varies in shape, a common all-or-none model was chose, where either the Action Potential exists or Not. And is irrelevant to other factors like the axon length, shape, width, the shape of the Action potential and so on.
    This proved to be such a dominant feature hence the scientific community considered only the all-or-none events into consideration in the calculations.
    This model enabled a large class of quite diverse membrane phenomenon to be analysed and modeled in terms of simple variables.
    Action potential is the spike.
  2. 2What is membrane potential ?Membrane potential is the ionic different inside and outside the neuron's plasma membrane. the Na+, K+, Cl- and organic anions (-ve) together comprise the membrane potential of the neuron.
    A neuron at Rest will have a -70 mV where the inside the membrane there is a higher concentration of K+ ions and outside Na+ is dominantly present.
    Due to excitatory and inhibitory signals when the potential reaches -50mv the Na+ gates are opened and the influx of Na+ ions inside the membrane will cause a spike (depolarisation). when the potential reach 30mv the K+ gates are opened and the NA-K pump will repolarized the nueron.
  3. 3What is Action Potential ?Action potential is the spike generated by the neuron when the membrane potential exceeds the threshold potential. The change in potential is due to the excitatory and inhibitory pulses received by the neuron.
  4. 4What were the scenarios which were not suitable to be tested on the Colossal Squid's Axon ?The factors responsible for the Action potential or spike in the neuron is generated, Does the spike generate when a Threshold voltage (Vth) is provided, or is a minimal threshold current (Ith) sufficient to generate the spike ? Or does the membrane require a minimum electrical charge (Qth) delivered to initiate the spike ?
  5. 5What is a Refractory Period?A biological neuron undergoes a brief time frame where its membrane is repolarised after it has recently spiked (Action potential). In this phases thousands on ions insides and outside the membrane are moved to restore the membrane potential back to resting potential.
    This is achieved by increasing the conducatance of potassium, this conductance remains active even subsequent to spike polarization (upto 12ms) during which the membrane is hyperpolarised, where it becomes very hard to generate an action potential than before. The membrane remains in a refractory state.
  6. 6Why is there no Action potential during Refractory Period ?In the Hodgkin and Huxley experiment, The neuron under the refractory state was supplied with a second current with increasing amplitudes. The second spike or Action potential was generated when the amplitude of the Second current was 23.7 times larger than the First current.
    This is unphysiological and it is de facto that there will be no Action potential will be generated when the neuron is under the Refractory period.
  7. 7In the Hodgkin Huxley model, what happens when a sustained current with same amplitude and increasing amplitude (ie, less than threshold, slightly more than or eq to threshold, more than threshold) is supplied to the neuron ?When the amplitude of the current supplied is less than the threshold current, a persistent sub-potential depolarization is observed, no Action potential is generated.
    When the amplitude of the current is sufficient to satisfy the Threshold current, A single action potential is generated and the neuron is hyperpolarized limiting it from generating a second spike, after the refractory period a second spike is generated. But a constant train of spikes (rheobase (0.065 nA)) is not generated.
    When the amplitude of the current is increased constantly, the offset depolarization following the action potential and the hyperpolarization increases until, when the amplitude of the current is 3 times the step of rehobase (0.175 nA). A second spike is generated and at 0.18 nA the membrane will generate a train of spikes (rehobase).
  8. 8What is excitatory type synapse ?If the synapse is of the excitatory type, the post-synaptic membrane potential increases upon arrival of the pre-synaptic spike.
  9. 9What is a inhibitor type synapse ?If the synapse is of the inhibitor type, the post-synaptic membrane potential decreases upon arrival of the pre-synaptic spike.
  10. 10How does excitatory and inhibitory type synapses affect membrane potential ?The membrane potential of a particular cell will vary over time,
    depending on the actual activities of the neurons it receives spikes from through
    excitatory and inhibitory synapses. When the threshold for spike generation is
    reached, the neuron fires itself and, thus, influences the potential and activity of
    all its post-synaptic neighbors
  11. 11What is the all-or-nothing model ?The Hodgkin and Huxley model found that the Action potential generated come in varied shapes.
    A common assumption was made termed the all-or-nothing model where the presence and absence of a spike is only taken in consideration and the shape and features of the spike is ignored.
    This model turned to be a dominant model in the scientific community which enabled for a large class of diverse membrane phenomenon to be analyzed in terms of simple variables.
  12. 12What is spike or pulse model ? Explain based on Assumptions, Neglected features, application and example of the model ?Assumptions: In this model, only the timing of the spike is relevant
    Neglect: The width, height of Action potential is neglected ( remember all-or-non and why it originated)
    In this model, the neural activity is a sequence of detal function ( Sigma(t-ti) ) at time ti. A slice of time is extracted and the activity of the neuron is determined based on the number of spikes generated in that time slice.
    Applications: Simplifying the spike models, modelling and simulation of the following,
    - dynamics of large interconnected networks
    - Inducing synaptic plasticity (Learning)
    Ex: Integrate and fire neurons (IaF)
  13. 13What is Firing Rate model ?A neural network model, where the description of a neural activity is reduced to taking into account only the mean firing rate.
    Ex: Obtained as the average number of spikes per unit time.
    The implicit assumption of this model is that most of the information in neural processing is contained within the mean activity and frequency of the spikes of the neurons
  14. 14Use of Firing rate model ?The firing rate model allows for a simple mathematical description of neural activity and synaptic interactions.
  15. 15Firing rate model mathematical representation ?Consider mean activity Si for neuron i. for a input from a set of J neurons when J =\= i ( self interactions is restricted).
    A biological neuron has a maximum threshold, we assume 0 <= Si, where upper limit 1 given in arbitrary units.
    The Activity of neuron 'i' is given as a non-linear response of incoming spikes which are however also represented only by the mean activity Sj ( Sj since the input for neuron i comes from a set of J),
    Firing rate model mathematical representation ?
  16. 16Assumptions, Neglect, application and example of Firing rate model ?Assumption: average rate (temporal and/or spatial) of spikes per unit of time
    carries (most) of the coded information/neural Activity.
    Neglect: precise timing of action potentials, temporal correlation of single spikes
    Extra: Firing rate model records neural activity as continuous, time-dependent firing rate.
    Application: Highly efficient simulation of network dynamics and learning process.
    Ex: Little-hopfield model.

Information

Description
Modelling and SImulation Flashcards on Neural Networks, created by S. Satyanarayan Nayak on 17/11/2019.
Views
23
Rating

More on Computational Neuroscience and Neuron Models

Feedback

What is it about?

What should we add next?