The code provided represents a computational model of neuronal activity within cortical networks, specifically aiming to investigate the influence of specific connectivity on feature selectivity, as explored in the referenced study by Sadeh, Clopath, and Rotter (2015). Below, I outline the biological basis of this model as reflected in the code: ### Biological Basis 1. **Neuronal Composition**: - **Total Neurons**: The network consists of 1000 neurons. - **Excitatory and Inhibitory Neurons**: The network is divided into excitatory (80%) and inhibitory (20%) neurons. This reflects biological cortical networks where excitatory neurons typically outnumber inhibitory neurons. 2. **Connectivity**: - **Probabilistic Connections**: Connections between neurons are set probabilistically, with excitatory-to-excitatory/inhibitory and inhibitory-to-excitatory/inhibitory connections having different probabilities. This mirrors synaptic connectivity in the brain, which is both structured and probabilistic. - **Feature-Specific Modulation (fs_mod_rng)**: Different degrees of synaptic specificity are modeled, from non-specific to highly specific, to study their effects on network activity and feature selectivity. 3. **Synaptic Strength and Plasticity**: - **Synaptic Weights**: The model uses specific synaptic weight values for excitatory and inhibitory synapses, represented as Excitatory Postsynaptic Potential (EPSP) and Inhibitory Postsynaptic Potential (IPSP). The inhibition dominance (g = 10) signifies that inhibitory inputs have a stronger effect, which is common in cortical circuits to maintain network stability. - **External and Feedforward Inputs**: The model incorporates external inputs akin to sensory or thalamic inputs, reflecting the interaction between local cortical circuits and broader sensory systems. 4. **Neuronal Dynamics**: - **Neuron Model**: Neurons are modeled using current-based integrate-and-fire dynamics, with parameters such as membrane time constant, resting potential, and firing threshold. - **Membrane Potential and Spiking**: Neuronal properties are defined to simulate realistic spiking behavior, including the use of membrane potential and threshold values relevant to biological neurons. 5. **Input and Orientation Selectivity**: - **Preferred Orientation (PO)**: Each neuron is assigned a preferred orientation, which aligns with how neurons in the visual cortex react preferentially to specific edge orientations within the visual field. - **Stimulus Presentation**: A set of orientations is presented to the network to evaluate how feature selectivity emerges from different connectivity configurations. 6. **Feedback and Delay**: - **Synaptic Delay**: A delay parameter captures the time lag in synaptic transmission, an essential feature in realistic neural circuits where even minimal transmission delays can impact network dynamics. 7. **Recording and Analysis**: - **Sample Neurons**: A subset of neurons is designated for detailed recording to analyze membrane potentials absent of spiking activity, enabling the study of input integration without output interference. This model framework is designed to replicate key aspects of cortical circuit function and structure, allowing for the exploration of how various degrees of connectivity specificity influence neuronal feature selectivity.