Given the simple firing rate model, τdv/dt=-v+w · u, in which v is the rate of the postsynaptic neuron, u and w are vectors representing the rates of the presynaptic neurons and the corresponding synaptic weights, respectively, which of the following equations describes the basic Hebb learning rule? *77. (part 24, easy) 105-Final
A. τwdw/dt=vu
B. w=vu
C. dw/dt=u
D. u=vw
查看答案
The long-term synaptic potentiation is specific because 75. (part 23, medium) 105-Final
A. when a neuron is activated by a high frequency input at a specific synapse, all other inactivated synapses also become facilitated.
B. induction of the synaptic potentiation only occurs at the synapses that undergo high frequency stimulations, not at the other inactivated synapses.
C. when a neuron is activated by a high frequency input at a specific synapse, other co-activated synapses also facilitated.
D. only one specific synapse in each neuron can be facilitated.
In the simple neural oscillator described in the lecture part 22, what do you expect if we change the time constant of the inhibitory neuron from small to large? 72. (part 22, medium) 105-Final
A. The system does not change its oscillation amplitude.
B. The oscillation stops
C. The system changes from a damped oscillation to an oscillation with a constant amplitude.
D. The oscillation accelerates.
Which of the following statements is not correct for a memory network described in the lecture part 21? 69. (part 21, difficult)105-Final
A. When an input is similar to a memorized item, the network has a high probability to converge to the state corresponding to the item.
B. The activity vectors that correspond to each memorized item have to be parallel to each other.
C. When the input is not similar to any memorized item, the network is likely to return to the baseline state.
D. The activity corresponding to a memory persists after the stimulus offset.
Which of the following conditions can turn a recurrent network into an input integrator? 66. (part 20, medium) 105-Final
A symmetric network with an eigenvalue smaller than zero
B. The eigenvalues of the network are all zero.
C. A symmetric network with an eigenvalue equal to 1.
D. A recurrent network can never be an input integrator.
微信支付 
支付宝支付