Synaptic impairment in neuronal circuits hinders memory recall, memory rescue, and learning

Connectivity-reduction

C++ code. The program is operated under windows with a gnu complier, and RandNum.h (from http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/emt.html) and run.h provided whitin this repository to run the program.

The model circuit is composed of 2000 excitatory neurons and 400 inhibitory neurons, and probability 0.2, the neurons were randomly connected. In inhibitory population, 200 neurons are picked to receive inhibitory pulse to mimic CA1 neurons rececving input from MSPV neuron. The neurons are labelled as:

0-1999 excitatory neurons; 2000-2199 inhibitory neurons; 2200-2399 inhibitory neurons that may receive inhibitory pulse;

The model used in this work is composed of conductance-based integrate-and-fire neurons, short-term plasticity and long-term plasticity, and the parameters used are biologically plausible, justified in previous works (Brunel and Wang, 2003; Dayan and Abbott, 2005; Hempel et al., 2000; Markram et al., 1998; Mongillo et al., 2008; Pfister and Gerstner, 2006). In particular, each layer has 2000 excitatory neurons, and 400 inhibitory neurons, n_I. Each neuron in L1 and L2 received background input from 400 independent Poisson trains.

In main.cpp, you can decide

The recoding number, count. With or without rescue input (inhibitory pulse to inhibitory neurons). The recsue input frequency if resuce input is applid. Network connectivity (from 0 to 1) connectivity. With or without learning.

the numbering of the simulation, count To run the simulation, you can run cmd in window, go to the directory folder and type mingw32-make. It will generate an exe, named "main.exe".

By running "main.exe", 7 txt files will be generated, which are

E=3.00I=8.00P={connectivity}spike{count}0{connectivity}.txt The txt has 2 columns, which are the spike time and neuron label (Each neuron has its unique label). i.e. 0-1999 are L1 excitatory neurons; 2000-2399 are L1 inhibitory neurons; 2800-4799 are L2 excitatory neurons; 4800-4999 are L2 Re inh neurons; 5000-5199 are FF inh neurons.

text_{count}_AMPA.txt The txt has 13 colums. The 1st column is running time of simulation in unit ms. The 2-11 columns are the total recurrent AMPA current recived by neurons in 10 different excitatory engrams. The 12th and 13th column are for inhibitory neurons that do not receive inhibitory pulse input and receive inhibitory pulse respectively.

text_{count}_NMDA.txt Same as text_{count}_AMPA.txt, but the 2-13 columns are total recurrent NMDA current.

text_{count}_GABA_L1.txt Same as text_{count}_AMPA.txt, but the 2-13 columns are total recurrent GABA current from inhibitory neurons that do not receive inhibitory pulse input.

text_{count}_GABA_L2.txt Same as text_{count}_AMPA.txt, but the 2-13 columns are total recurrent GABA current from inhibitory neurons that receive inhibitory pulse input.

text_{count}_V.txt Same as text_{count}_AMPA.txt, but the 2-13 columns are total membrane potential of neurons.

E=3.00I=8.00P={connectivity}_W_{count}_0_{connectivity}.txt A matrix representing the final synaptic weights after learning. Each number of the matrix is the synaptic weight from the neuron with label (column number -1) to neuron with label (row number -1).