[Is understanding the mechanism of perception possible?]

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Perception involves the brain's interpretation of sensory information, primarily visual in humans.
• Visual information is processed in the occipital cortex and further deciphered in the temporal cortex.
• Magnetic Resonance Imaging (MRI) techniques provide insights into brain activity during perception.

Purpose of the Study:

• To discuss the neural codes used by neurons in the temporal cortex for visual recognition.
• To examine two specific instances: recognition of written words and recognition of human faces.
• To present hypotheses and findings related to invariant visual form recognition.

Main Methods:

• Review of existing hypotheses on neural coding for visual recognition.
• Application of a hierarchical model of local combination detectors for word recognition.
• Analysis of neural activity from neuron patches in the inferior temporal cortex for face recognition in macaques.

Main Results:

• An effective hypothesis for invariant visual form recognition in reading was developed based on neuronal properties.
• Face recognition in macaques was localized to specific neuron patches in the inferior temporal cortex.
• A system for view-invariant facial identification was identified, with the neural code for 200 neurons cracked.

Conclusions:

• Neural codes in the temporal cortex are crucial for deciphering complex visual information.
• Hierarchical processing and specific neuronal assemblies support invariant recognition of both written words and faces.
• Understanding these neural codes advances our knowledge of brain function in perception and identification.