Learning and executing goal-directed choices by internally generated sequences in spiking neural circuits

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Neural ensemble recordings link hippocampal neural sequences to goal-directed spatial decision-making in rats.
• Understanding the synaptic mechanisms driving these internal sequences is crucial for deciphering decision-making processes.

Purpose of the Study:

• To investigate the synaptic mechanisms underlying internally generated neural sequences involved in spatial decision-making.
• To elucidate how combined synaptic plasticity mechanisms contribute to adaptive navigation.

Main Methods:

• Development and investigation of a spiking neural circuit model.
• Incorporation of two synaptic plasticity mechanisms: spike-timing dependent plasticity (STDP) and synaptic scaling.
• Quantitative comparison of model dynamics and decision-making accuracy with experimental data.

Main Results:

• The model, with combined STDP and synaptic scaling, generated forward-sweeping neural sequences consistent with experimental observations.
• Inclusion of both plasticity mechanisms was necessary for accurate sequence generation and decision-making; single mechanisms failed.
• The model demonstrated adaptive responses to changing cue-goal associations, performing comparably to a Kalman filter.

Conclusions:

• The combination of synaptic plasticity mechanisms on different timescales is a plausible mechanism for generating internal neural sequences.
• This dual plasticity approach supports adaptive spatial decision-making and navigation in complex environments.
• The findings provide a computational framework for understanding neural sequence formation and its role in behavior.