Simple synaptic modulations implement diverse novelty computations.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Synaptic Plasticity

Background:

• The brain must identify novel stimuli for survival amidst changing environments.
• Inhibitory microcircuits in the neocortex regulate activity based on experience, with distinct neuron subtypes showing varied novelty responses.
• Understanding experience-dependent neural circuit function requires a flexible and biologically plausible novelty detection mechanism.

Approach:

• Introduced familiarity modulated synapses (FMSs), a novel unsupervised learning mechanism based on multiplicative synaptic modifications.
• FMSs operate under continual learning, lack specialized architecture, and enable rapid novelty detection without feedback.
• Implemented FMSs in an experimentally-constrained model of a visual cortical circuit.

Key Points:

• FMSs successfully reproduced three experimental novelty effects: absolute, contextual (oddball), and omission novelty.
• The model demonstrated functional diversity within cell subpopulations, predicting heterogeneous individual neuron signals.
• The mechanism relies solely on presynaptic or both pre- and postsynaptic activity for synaptic modification.

Conclusions:

• Simple plasticity mechanisms within cortical circuits can generate distinct, experience-dependent novelty responses.
• FMSs offer a computationally tractable and theoretically flexible framework for studying novelty detection.
• Findings provide experimentally testable predictions regarding connectivity and synaptic dynamics in novelty processing.