Algorithm 3 Patrolling.

Parameters: gain ${\displaystyle \gamma }$, noise ${\displaystyle ϵ}$, habituation ${\displaystyle \beta }$, recovery time ${\displaystyle \tau }$ Input: map synapses ${\displaystyle \mathbf{M}}$   ${\displaystyle h_i\leftarrow 1}$, for all point cells ${\displaystyle i}$                 ${\displaystyle ▹}$ starting sensitivity of point cell at node ${\displaystyle i}$   ${\displaystyle s\leftarrow x}$                                                 ${\displaystyle ▹}$ begin patrolling at node ${\displaystyle x}$   while patrolling do     ${\displaystyle h_s\leftarrow h_s{\mathrm{e}}^{-\beta }}$                                          ${\displaystyle ▹}$ habituation of point cell ${\displaystyle s}$     ${\displaystyle h_i\leftarrow 1-(1-h_i){\mathrm{e}}^{-1/\tau }}$, for all ${\displaystyle i}$                   ${\displaystyle ▹}$ resensitization of all point cells     for all nodes ${\displaystyle j}$ that neighbor ${\displaystyle s}$ do                   ${\displaystyle ▹}$ agent tests available options       ${\displaystyle u_i(j)\leftarrow {\delta }_{i,j}{h}_j}$ for all ${\displaystyle i}$                     ${\displaystyle ▹}$ point cell output with agent at node ${\displaystyle j}$       ${\displaystyle \mathbf{v}(j)\leftarrow {(\frac{1}{\gamma }\mathbf{1}-\mathbf{M})}^{-1}\mathbf{u}(j)}$                                         ${\displaystyle ▹}$ map output       ${\displaystyle p(j)\leftarrow \frac{1}{Z}\sum _i{v}_i(j)+\eta }$     ${\displaystyle ▹}$ sum of map output with noise, normalized so max = 1     end for     ${\displaystyle s\leftarrow \underset{j}{\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{m}\mathrm{a}\mathrm{x}}p(j)}$            ${\displaystyle ▹}$ move to neighbor node with the highest patrol signal   end while