Visiting Research Fellow, Homi Bhabha Centre for Science Education (HBCSE), Mumbai · January 2020 – July 2021
Supervisor: Prof. Sanjay Chandrasekharan
Poster: Sheetal Jadhav and Sanjay Chandrasekharan, “She saw me and made a face: functional network analysis reveals role of the limbic system in face recognition”. 7th Annual Conference of Cognitive Science (ACCS7), Indian Institute of Science, Bangalore (online), 23 January 2021. Abstract book, pp. 44–45.
About
Faces carry some of the most important social information we receive: identity, emotional state and where someone is looking. Face processing is therefore tied to emotion, action and social cognition, which suggests that it recruits brain regions well beyond the core face-selective areas. This project asked which wider regions take part in face processing, and what they might contribute.
Methods
- Data: the Haxby fMRI dataset, an open dataset from a study of face and object representation in the human brain. Six subjects each passively viewed greyscale images of eight object categories, including faces, in 12 runs.
- Preprocessing: brain extraction and motion correction in FSL.
- Seed-based functional connectivity: with Nilearn in Python, the BOLD signal of a seed region, the occipital fusiform gyrus (part of the core face-processing network), was correlated with every other voxel in the brain to build a functional connectivity map, followed by a group analysis across all six subjects.
Findings
- The face-correlated network was bilateral and included the superior and middle frontal gyri, the cingulate gyrus, the insula and the medial occipitotemporal gyrus.
- The cingulate gyrus points to a role for the limbic system: making sense of, and acting on, facial expressions and emotions, which a purely geometric recognition of a face does not provide.
- The posterior cingulate is part of the default mode network, which suggests that internally directed processes such as memory retrieval and planning are also part of face recognition.
