An open hypothesis: is epilepsy learned, and can it be unlearned?

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Neural plasticity is fundamental for learning and brain development.
• Learning processes can introduce instability within neural systems.
• Epilepsy involves spontaneous seizures, but its connection to learning remains unclear.

Purpose of the Study:

• To investigate the link between learning mechanisms and the development of spontaneous seizures.
• To explore whether Hebbian learning systems possess inherent vulnerabilities to epileptogenesis.
• To propose a novel hypothesis for epilepsy as a learned network-level response.

Main Methods:

• Analysis of constraints on spontaneously active Hebbian learning systems.
• Theoretical investigation into the relationship between learning algorithms and neural network stability.
• Formulation of a hypothesis based on computational principles.

Main Results:

• The capacity for Hebbian learning confers an intrinsic vulnerability to epileptogenesis in neural systems.
• Learning is identified as a destabilizing process at the algorithmic level.
• Epilepsy is hypothesized to be an abnormal learned response to repeated stimuli.

Conclusions:

• Epilepsy may represent a learned network-level phenomenon.
• The hypothesis suggests that epilepsy could be reversible by 'unlearning' the response.
• This perspective offers a potential new therapeutic avenue for epilepsy treatment.