$te$	transfer entropy value
g	the binning threshold
$t{e}_{j,i}^l$	obtained $te$ calculated using the time series produced by binning the outputs of neurons’ index i from layer $l-1$ and index j from layer l
$t{e}_{j,i}^{r,n}$	obtained $te$ using the time series from outputs of neurons’ index i from layer $l-1$ and index j from layer l, at epoch r and sample n
$o_i^{r,n}$	output of neuron having index i—that is layer $l-1$, at epoch r using sample n
$o_j^{r,n}$	output of neuron having index j—that is layer l, at epoch r using sample n
$s_i^{r,n}$	binning the output of neuron index i—layer $l-1$, at epoch r using sample n, with threshold g
$s_j^{r,n}$	binning the output of neuron index j—layer l, at epoch r using sample n, with threshold g
CNN	Convolutional Neural Network
FF+FB	Feedback Transfer Entropy—our proposed method
MLP	Multi-layer perceptron
FF	Non Feedback network, regular MLP architecture and algorithm
TE	Transfer Entropy