Neural waves and computation in a neural net model I: Convolutional hierarchies.

Area of Science:

• Computational neuroscience
• Neurobiology
• Artificial intelligence

Background:

• Neuromorphic network models offer insights into brain computation.
• Hierarchical systems like the mammalian visual cortex are complex.
• Spontaneous neural activity plays a role in neural processing.

Purpose of the Study:

• Investigate the computational resources of a neuromorphic network model.
• Examine the formation of hierarchical feed-forward structures.
• Analyze the emergent logic within neural networks.

Main Methods:

• Utilized a previously introduced neuromorphic network model.
• Simulated spontaneous wave-like neural activity.
• Investigated local Hebbian modulation and extra-synaptic effects.

Main Results:

• Demonstrated spontaneous local convolution driven by wave-like activity.
• Showed the formation of logical gate-like neural motifs into Hubel-Wiesel type structures.
• Identified the significant role of extra-synaptic effects.
• Confirmed the emergence of non-Boolean logic.

Conclusions:

• Spontaneous activity and Hebbian modulation are key to forming hierarchical neural structures.
• Extra-synaptic effects are crucial for these computational processes.
• The emergent non-Boolean logic has implications for understanding brain function and conditions like schizophrenia.