Adaptation without Plasticity.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Sensory adaptation describes how neural responses change based on recent input history.
• In the visual cortex, neurons exhibit shifts in preferred orientation after stimulus exposure.
• This adaptation is conventionally linked to synaptic plasticity.

Purpose of the Study:

• To investigate whether neural adaptation phenomena can arise from network dynamics alone.
• To determine if plasticity is a necessary mechanism for adaptation on millisecond timescales.
• To model tuning curve shifts in the visual cortex using a fixed recurrent network.

Main Methods:

• Developed a computational model of a recurrent neural network.
• Simulated network activity under varying input sequences.
• Analyzed changes in neuronal tuning curves within the model.
• Compared model outputs to experimental data from cat and macaque visual cortex.

Main Results:

• The recurrent network model reproduced observed tuning curve shifts without altering synaptic weights.
• Adaptation on timescales of hundreds of milliseconds was achieved through network dynamics.
• Fixed network properties were sufficient to generate context-dependent neural responses.

Conclusions:

• Neural adaptation, specifically tuning shifts, does not necessitate plasticity in recurrent networks.
• Recurrent connections provide a mechanism for integrating recent contextual information.
• Changes in neural tuning alone are insufficient evidence for plasticity on these timescales.