Detecting and estimating signals over noisy and unreliable synapses: information-theoretic analysis

Area of Science:

• Neuroscience
• Computational Neuroscience
• Information Theory

Background:

• Neuronal response precision is key to the neural code.
• Understanding neuronal noise sources is vital for information transfer.
• Synaptic transmission is inherently unreliable and probabilistic.

Purpose of the Study:

• To investigate the impact of unreliable synaptic transmission on information transfer.
• To derive theoretical lower bounds on cortical synapse capacity.
• To characterize information transfer efficacy under signal estimation and detection paradigms.

Main Methods:

• Developed a simple model of a cortical synapse.
• Derived theoretical lower bounds on synaptic capacity.
• Analyzed information transfer under signal estimation (mean firing rate) and signal detection (binary input) paradigms.

Main Results:

• Single cortical synapses exhibit unreliable information transmission.
• Optimal strategies were derived for both signal estimation and detection.
• Parameter values from neocortex were used for analysis.

Conclusions:

• Cortical synapses alone have limited information transmission capacity.
• Redundancy through multiple synapses substantially improves information capacity.
• This highlights the importance of synaptic redundancy in neural information processing.