Table 3.

Parameters of the computational model

Parameter	Description	Approach	Withdrawal
$r_E$	Excitatory (E) firing rate
$r_I$	Inhibitory (I) firing rate
$W_{EE}$	E-to-E synaptic weight	5	0.1
$W_{EI}$	I-to-E synaptic weight	−0.1	−10
$W_{IE}$	E-to-I synaptic weight	0.1	10
$W_{II}$	I-to-I synaptic weight	−1	−0.1
${\theta }_E$	E baseline input	10	10
${\theta }_I$	I baseline input	10	1
$W_E$	Input synaptic weight to E	1	0.1
$W_I$	Input synaptic weight to I	10	1
$\theta$	Neural threshold	20
$I$	Input

Numerical values of the parameters in a computational scheme in which visual inputs are combined to mimic personal space coding (Fig. 11B). Following the widely used framework of the firing rate neural models (Wilson and Cowan, 1972), the firing rates of excitatory and inhibitory populations were modeled as the following:

$$\frac{d{r}_E}{dt}=-r_E+W_{EE}f\left(r_E\right)+W_{EI}f\left(r_I\right)+{\theta }_E+W_{E}I$$

$$\frac{d{r}_I}{dt}=-r_I+W_{IE}f\left(r_E\right)+W_{II}f\left(r_I\right)+{\theta }_I+W_{I}I,$$

where $r_E$ and $r_I$ are the firing rate of excitatory and inhibitory populations, respectively. $W_{EE}$ is the synaptic strength of excitatory-to-excitatory connections, $W_{EI}$ is the strength of inhibitory-to-excitatory connections, $W_{IE}$ is the strength of excitatory-to-inhibitory connections, and $W_{II}$ is the strength of inhibitory-to-inhibitory connections. Each population has a baseline activity represented by ${\theta }_E$ and ${\theta }_I$, respectively. Finally, each population receives the external input $I$ weighted by synaptic strengths $W_E$ and $W_I$, respectively. Each population receives the firing rate of the other population as input through a sigmoid nonlinear function $f\left(x\right)=1/\left(1+\text{exp}(-(x+\theta ))\right)$, where $\theta$ represents the intrinsic soft threshold of the population activity. The variable of interest in these simulations was the stable steady-state firing rate of the excitatory population as a function of the strength of the external input $I$.