Inhibitory Feedback Enables Predictive Learning of Multiple Sequences in Neural Networks

Area of Science:

• Computational Neuroscience
• Neural Networks

Background:

• Anticipating future events is crucial for neuronal network computation.
• Temporal sequences in neural activity are linked to event association and anticipation.
• Mechanisms for differentiating and anticipating multiple spike sequences are unclear.

Purpose of the Study:

• To investigate how neuronal networks can differentiate and anticipate multiple spike sequences.
• To explore the role of predictive processing and inhibitory feedback in sequence anticipation.

Main Methods:

• Implemented a learning rule based on predictive processing.
• Incorporated inhibitory feedback into the neural network model.
• Analyzed network activity for sparse firing and sequence encoding.

Main Results:

• Neurons fired selectively for initial, unpredictable inputs, reducing postsynaptic firing.
• Inhibitory feedback induced sparse firing, enabling anticipation of different sequences.
• Optimal intermediate inhibition levels decorrelated neuronal activity for future input prediction.

Conclusions:

• The combination of self-supervised predictive learning and inhibitory feedback allows efficient sequence representation.
• This mechanism enables fast and accurate classification of diverse input sequences.
• Sparse, anticipatory firing independently encodes each sequence.