Merging information in the entorhinal cortex: what can we learn from robotics experiments and modeling?

Area of Science:

• Neuroscience
• Computational Neuroscience
• Robotics

Background:

• Place recognition relies on integrating sensory inputs.
• The entorhinal cortex (EC) is crucial for spatial cognition.
• Mammalian brains process 'what' and 'where' visual information.

Purpose of the Study:

• To explain how the entorhinal cortex (EC) forms compact place codes.
• To investigate the role of conjunctive and grid cells in spatial memory.
• To propose a model for grid cell formation based on information compression.

Main Methods:

• Review of existing evidence on visual processing and place cells.
• Analysis of computational models and robotics experiments.
• Hypothesizing mechanisms for information convergence and compression in the EC.

Main Results:

• Visual 'what' and 'where' information converges in the EC.
• Conjunctive cells merge multimodal information for place cells.
• Modulo projection of cortical activities may explain grid cell formation and visual exploration.

Conclusions:

• The EC constructs robust, compressed neural codes for places.
• The hippocampus uses these codes for recognition and predicting state transitions.
• This model unifies findings on place and grid cells.