Linking visual behavior to neural processing in the thalamusMy laboratory is investigating the neural basis of natural visual perception in mammals. Our broad program is to connect the visual neurophysiology of the thalamus “down” to its underlying cellular mechanisms and “up” to its perceptual function in alert animals. Visual information is encoded by the retina and transmitted to cortex by way of the Lateral Geniculate Nucleus (LGN) of the thalamus. The description of the LGN as a simple relay from retina to cortex is a good first approximation, but it provides no insight into the functional consequences of the conspicuous active membrane properties of the cells or the massive feedback projection from visual cortex. We and others have suggested computational functions of these mechanisms on the basis of
in vivo neurophysiology in anesthetized animals. Mechanisms such as intrinsic bursting or cortical feedback are hypothesized to selectively regulate feed-forward transmission through the LGN in response to stimulus context, expectation, or task-relevance. To directly test these hypotheses we must bridge between levels of analysis, using genetics and pharmacology to manipulate circuits, in
vivo neurophysiology to probe cellular function, and visual behavior to assess perception and control attention. These experiments will be interpreted in the framework of quantitative theories.
More specifically, relay neurons in the thalamus encode the local luminance contrast in their receptive fields. Yet this encoding is sensitive to context, at least in anesthetized animals. The same visual stimulus can evoke either a single spike or a stereotyped burst depending on recent stimulus history; this is proposed to be accomplished by a feedforward mechanism involving a voltage-gated calcium channel in the relay cell. Moreover, visual stimuli outside of a relay cell’s classical receptive field can alter the physiological responses to stimuli within its receptive field. Cortical inactivation experiments suggest that cortical feedback plays a role in this stimulus-based contextual modulation of thalamic responses. At the level of perception, it is known that human judgments of local luminance can be powerfully influenced by contextual cues such as past stimuli, distant stimuli, or gestalt image structure; these effects are the basis of many strong visual illusions. Human performance at visual detection or discrimination can also be affected by cognitive context, such as expectation or spatially selective attention. There is some evidence from human functional imaging that these cognitive effects influence activity as early in the visual stream as the LGN. These stimulus-based and cognitive contextual cues are not thought to be available to an LGN relay cell in its feed-forward retinal inputs, but could be available from feed-back cortical signals. A thorough investigation of information processing in the thalamus requires that we now integrate perception and attention with physiology in the same animal, and directly test mechanisms with manipulations at the cellular and circuit level. To accomplish this goal we have developed a novel paradigm for visual physiology in alert, visually-behaving rats.
