# Biological Basis of the Computational Neuroscience Model ## Introduction The provided code snippet is part of a computational neuroscience model designed to simulate aspects of neural dynamics, specifically focusing on orientation selectivity in cortical networks. This model is based on a study by Sadeh and Rotter (2015) that investigates how various factors, including single neuron properties and network dynamics, contribute to orientation selectivity. ## Model Components ### Neuron Types and Numbers - **Two Neuron Populations**: The model simulates a network of 5000 neurons, divided into excitatory (marked by `NE`) and inhibitory (`NI`) populations, with an 80:20 ratio. This ratio is consistent with observations in the mammalian cortex where excitatory neurons often dominate the network. ### Recurrent Coupling - **Synaptic Strength**: The recurrent synaptic connections are defined by `J_rec`, with positive weights for excitatory-to-excitatory and excitatory-to-inhibitory synapses, and negative weights for inhibitory-to-excitatory and inhibitory-to-inhibitory synapses. The inhibitory connections are stronger by a factor of `g = 4.0`, reflecting the inhibition-dominated dynamics often observed in cortical networks. - **Connectivity Probability**: `eps_exc` and `eps_inh` represent the connection probability within excitatory and inhibitory populations, modeling the sparse yet structured connectivity in cortical networks. ### Non-local Input - **Background Activity**: The model incorporates a non-local network of 5000 neurons (`n_ext`) providing input, simulating the spontaneous background activity present in the brain (`r_ext`) that affects local circuits. - **Synaptic Input**: The synaptic strength from this non-local network is set to `j_ext = 0.2`, representing the moderate excitatory post-synaptic potentials contributed by external sources. ### Feedforward Input from LGN - **LGN Input**: The model includes a feedforward input from the lateral geniculate nucleus (LGN), represented by `n_ffw`, which simulates the visual input pathway. Each cortical neuron receives input from an average of 50 LGN neurons, with a firing rate (`r_ffw`) reflecting the typical activity level of thalamic relay cells. - **LGN to Cortex Synapses**: The impact of this input is determined by `j_ffw`, a higher synaptic weight, indicating significant excitatory postsynaptic potentials due to these afferent connections. ### Neuronal Orientation - **Orientation Preferences**: Initial preferred orientation (PO) angles (`po_init`) are assigned randomly across the neuron population, with sample neurons having defined orientations. This simulates the columnar organization of orientation selectivity in the visual cortex, where neurons respond preferentially to specific orientations of visual stimuli. ## Summary The code defines parameters for a network model that captures key elements of cortical microcircuitry relevant to orientation selectivity, a fundamental property of visual processing in the brain. It incorporates biological realism through its division into excitatory and inhibitory neurons, balances between recurrent and external inputs, and structured orientation preferences. Each of these components is grounded in empirical observations of cortical structures and dynamics, particularly in the context of primary visual cortex (V1) functioning.