Trained as a visual psychophysicist, I have been mainly focusing my research on perceptual learning in the past decade. My study demonstrated that the learning specificities associated with perceptual learning can be removed via a double-training technique I invented, so that learning can transfer to untrained conditions for maximal efficiency. The learning transfers via double training are not limited to visual learning, but is achievable in auditory and motor skill learning. These and other related findings suggest that perceptual learning is a high level rule-based processes, and likely operates at a conceptual level. My other research topics include letter recognition and crowding, and contour integration.

Recently my research also shifted partially to two-photon imaging of macaque monkeys. Taking advantage of this cutting-edge new technology, we are able to simultaneously record hundreds of visual cortical neurons' responses at single-neuron precision. In one study we constructed cellular orientation maps in macaque V1, which confirm iso-orientation clusters, but argue against the pinwheel structures populated by optical imaging recordings. We found that most of the pinwheels are centered in blood vessel regions, suggesting that they are likely artifacts from data fitting.