The neural correlates of supervised and unsupervised visual learning

Animals learn to adapt and optimize their responses to the sensory information in their environment. This learning can be achieved either through explicit training (supervised learning) or by mere exposure (unsupervised learning). Previous visual learning studies in monkeys have found neuronal response modulations in various brain areas like early visual areas and inferior temporal cortex (IT) and higher order decision making areas during supervised learning.

In this project we endeavor to find neural correlates of both supervised as well as unsupervised visual learning. Pertaining to supervised learning, this project focuses on time-frequency analysis of local field potentials (LFP) recorded in area V4 of monkeys, while they were practising a coarse orientation discrimination task. In regard to unsupervised learning, this project focuses on two aspects: 1) trying to map brain regions involved during unsupervised learning using functional magnetic resonance imaging (fMRI); 2) studying the effect of sequence learning (unsupervised learning) on macaque IT neuron responses by recording and analyzing the multi-unit activity (MUA) of neurons in macaque IT cortex after unsupervised learning.