Dear vision scientists,
Recent evidence shows that lateral geniculate nucleus (LGN) has more sophisticated functions than a mere relaying hub. We provide a rather unfiltered and eclectic review of the LGN functions, with some focus on its computational role, as well as its role in higher level cognitive process (specifically higher level vision). In particular, we argue that the LGN has a variety of complex nonlinear behaviors (perhaps the most interesting nonlinearity involves feedback from cortex). We suggest further development of feedback models of high-level vision considering the LGN function. Because of LGN’s contribution in temporal aspects of object processing, modeling LGN function particularly becomes more important when considering non-stationary stimuli (such as videos). Predictive coding and expectation coding models are in this direction.
Best
--
*_____________________________________________________________*
Masoud Ghodrati,
Department of Physiology,
Monash University, Melbourne, Australia.