Brain Computing Lab
Project 2
Read More

The role of neural feedback control for object recognition during partial occlusion


One prominent feature of biological visual systems, including our own, is the ability to recognize objects independent of various transformations, such as location and orientation. One of the most dramatic of these transformations is occlusion: non-transparent objects in the line of sight closer to the observer will block objects further away from forming a complete image on the retina. While most previous studies of the neural mechanisms of object vision mainly focus on the representation of unoccluded objects, only a few studies aimed to address how the challenge of partial occlusion is resolved in our brain. Previous studies have suggested that visual neurons respond to partially occluded objects in a compromised way; but the whole picture is still unclear. To fully understand how the brain resolves partial occlusion, two main questions need to be answered: 1) How does the brain represent the occluded item, the occluder and their relationship? 2) What is the neural mechanism underlying the recognition of partially occluded items?


In this project we propose to study how primate brain performs object recognition under partial occlusion and the specific roles of feedback from PFC to IT in this process. We will record simultaneously from both areas while the monkey matches partially occluded objects to unoccluded objects. Studying how feedfoward and feedback signals flow between IT and PFC cooperate in recognizing partially occluded objects will help us construct new object recognition systems.



Mohammad-Reza Abolghasemi Dehaqani

Jalaledin Noroozi


Ehsan Rezayat

rezayat [at] ipm [dot] ir

Atlas Shahamati




Le Chang

University of Chinese Academy of Sciences