Feed-forward artificial neural network provides data-driven inference of functional connectivity.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Biology

Background:

• Functional connectivity analysis is crucial for understanding complex biological systems.
• Existing methods often require significant computational resources and extensive data.
• Accurate detection of dynamic functional networks is essential for diagnosing neurological disorders.

Purpose of the Study:

• To introduce a novel, model-free method for detecting functional connectivity using artificial neuronal networks.
• To develop a computationally efficient approach applicable to short experimental data trials.
• To validate the method on a known chaotic system and apply it to analyze epileptic brain networks.

Main Methods:

• A feed-forward artificial neuronal network model was developed for functional connectivity detection.
• The method was tested on the chaotic Rössler system for validation.
• The approach was applied to electrocorticography (ECoG) data from WAG/Rij rats with absence epilepsy.

Main Results:

• The artificial neuronal network method demonstrated accurate functional connectivity detection.
• The model showed good agreement with established results on the Rössler system.
• Functional interdependence between cortical layers and thalamic nuclei was observed following epileptic discharge onset in rats.

Conclusions:

• The proposed model-free artificial neuronal network method offers an efficient approach for functional connectivity analysis.
• This method is suitable for analyzing diverse experimental multichannel data, including limited datasets.
• The findings highlight the emergence of functional interdependence in epileptic brain networks, offering insights into absence epilepsy.