CogSci Colloquium Talk:
Reverse engineering human visual intelligence: Developmental origins and encoding models

Dr. N. Apurva Ratan Murty

Postdoctoral fellow at The Center for Brains, Minds and Machines, Massachusetts Institute of Technology

Title: Reverse engineering human visual intelligence: Developmental origins and encoding models
Date: Thursday, 24th of February 2022
Time: 05:00 – 06:00 PM IST (6:30 – 7:30 AM EST)
Calendar: Add to gCal


The last quarter century has provided extensive evidence that some regions of the human cortex are selectively engaged in processing a single specific domain of information, from faces, places, and bodies to music, language, and other people’s thoughts. This work dovetails with earlier theories in cognitive science highlighting domain specificity in human cognition, development, and evolution. But many questions remain unanswered about even the clearest cases of domain specificity in the brain. In my talk I will describe a series of experiments that investigate two such outstanding questions about the cortical organization of the human ventral visual cortex (VVC).

The first part will focus on the developmental origins of the fusiform face area (FFA) a canonical node in the VVC that responds selectively to faces. How does the FFA arise in development, and why does it develop so systematically in the same location across individuals? Preferential fMRI responses to faces arises early by around 6 months in humans (Deen et al., 2017). Arcaro et al (2017) have further shown in monkeys that regions that later become face selective are correlated in resting fMRI with foveal retinotopic cortex in newborns, and that monkeys reared without ever seeing a face show no face-selective patches. These findings have been taken to argue that 1) seeing faces is necessary for the development of face selective patches, and 2) face patches arise in previously fovea-biased cortex because early experience with faces is foveally biased. I will present evidence against both these hypotheses by demonstrating robust face selective responses in congenitally blind participants during a novel experimental paradigm in which participants haptically explore 3D-printed stimuli.

In the next part of the talk, I will take a step back and critically evaluate the evidence for cortical selectivity. Even though cortical selectivity has provided important evidence for domain-specific theories of human cognition, development, and evolution, they are not computationally precise and remain vulnerable to empirical refutation. Given these critical shortcomings, is the theory even true? To address these issues, I will show some recent efforts to reverse engineer these regions by constructing deep artificial neural network (ANN)-based encoding models that predict the observed response to novel images, outperforming descriptive models and even professors! I will then further demonstrate how we can apply these models to make ever stronger inferences about theories of category selectivity which points the way for future research characterizing the functional organization of the human brain with unprecedented computational precision.

About the Speaker
Ratan obtained his PhD in Neuroscience from the Indian Institute of Science with Prof. SP Arun. He is currently a postdoctoral fellow at the Center for Brains, Minds and Machines, McGovern Institute for Brain Research at the Massachusetts Institute of Technology with Profs. Nancy Kanwisher and Jim DiCarlo. His current research interests are to elucidate the neural codes and algorithms that enable our rich visual perception of the world and how those codes mature over the course of visual development.