Synaptic weights that correlate with presynaptic selectivity increase decoding performance.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Neuronal activity in the visual cortex is often described by tuning curves shaped by Hebbian plasticity.
• This suggests neural circuits connect neurons with similar functional preferences more strongly.
• Recent findings indicate postsynaptic preference depends on activated spine count, independent of individual spine strength.

Purpose of the Study:

• To investigate how synaptic plasticity influences postsynaptic functional preference in the visual cortex.
• To develop a computational model that reconciles previous findings with new experimental data.
• To explore the role of input number versus input strength in defining postsynaptic selectivity.

Main Methods:

• Developed a computational model of visual cortex neural circuits.
• Implemented a plasticity rule where synaptic weights correlate with presynaptic selectivity, independent of postsynaptic similarity.
• Simulated stimulus presentation and analyzed the model's ability to decode stimuli.

Main Results:

• The model demonstrates that postsynaptic functional preference can be defined by the number of activated inputs.
• Synaptic weights correlating with presynaptic selectivity, irrespective of functional similarity, can shape circuit organization.
• The model achieved stimulus decoding performance comparable to maximum likelihood inference.

Conclusions:

• Computational models incorporating specific plasticity rules can explain observed neural circuit organization in the visual cortex.
• The number of activated inputs, rather than their individual strength or functional similarity, can be a key determinant of postsynaptic selectivity.
• This work provides a framework for understanding how synaptic plasticity contributes to functional subnetworks and stimulus processing in the visual cortex.