Gruber Lecture: Brain Circuits for Active Vision. Robert Wurtz, PhD, National Eye Institute/National Institutes of Health After an brief introduction to the Gruber International Prize Program by Sarah Hreha, Executive Vice President of the Peter and Patricia Gruber Foundation, Michael Goldberg (and his excellent bow tie) present the 2010 International Research Award for Neuroscience to Laura Colgin and Jason Shepherd. Congratulations to them both.
This is followed by the presentation of the Gruber Prize to Robert Wurtz by Sten Grillner (chair of the Neuroscience selection advisory board for the Gruber Foundation). Wurtz long, auspicious career is described, highlighting his groundbreaking work on the visual system of awake-behaving primates. The official text of the award: “Honored for his pioneering work in the neurophysiology of visual cognition, which has led scientists to a deeper understanding of how the brain is organized to produce behavior.”
And after a few acknowledgements, Robert Wurtz begins his talk.
Wurtz notes the usefulness of having post-docs in the lab… for when a monkey gets loose.
When presented with a visual scene, we have the perception that we see every part of that scene with equal clarity at all times. However, neuroscience has shown that this is emphatically not true, we constantly make saccadic eye movements that focus our high acuity vision to examine different parts of the visual field at any time. These eye movements shift the fovea, and visual attention, displacing the retinal image of the visual scene and producing a blurred sweep of visual information during the saccade movement itself. And yet with all this movement and displacement, our brain computes the visual information to produce the illusion of a unified, clear visual field.
The talk will address four points: 1) an overview of the brain areas involved in active vision, 2) attention (enhancement of vision with saccades), 3) suppression (decreases of the inter-saccade blurring) and 4) the mechanism that generates saccades, which underlies both attention and suppression.
First, Wurtz describes the methodologies used to study active vision in awake monkeys. The basic setup is to record neuronal activity from head-fixed monkey who are fixating on a visual location.
Next, he outlines the basic system underlying the generation of saccadic eye movements. From V1, the pathway extends to the posterior parietal cortex and the frontal cortex, and thereon to the superior colliculus, and to the midbrain and pons for generation of motor output. To highlight the role of the superior colliculus, which is a critical component of the saccade generating circuit. For the purposes of the talk, the SC can be divided into two layers, a superficial, retinorecipient layer, and a deeper layer that encodes information regarding planned saccade movements. These visual-motor neurons display a burst of activity immediately prior to the eye movement.
With this type of activity introduced, Wurtz presents a two-fold experimental strategy: 1) correlation of neuron activity to behavior and 2) perturbation of neuron activity via activation or inactivation of neurons. He notes that in the second part of this strategy, classic techniques have been chemical or electrical means, but that recently they have turned to optogenetic inactivation of neurons located in the deep layers of the superior colliculus (this in collaboration with Ed Boyden’s lab).
Correlating and perturbing neuronal activity leads to classical view of superior colliculus, namely that it is largely involved in motor generation of the saccade. Newer research has suggested that the superior colliculus is also a major source of ascending pathways to inform the cerebral cortex about the saccade. These ascending pathways leave the superior colliculus and go directly to the thalamus before going to cortical areas (for example, SC to MD to FEF; SC to Pulvinar to LIP/MT or SC to Reticular Nucleus to thalamocortical relay neurons (LGN)). Wurtz will focus on the SC to Pulvinar to LIP/MT pathway.
We now turn to the topic of spatial attention, specifically looking at the phenomenon of change blindness. We are showed a demonstration of change blindness, which demonstrates the power of spatial attention to direct
Wurtz describes two types of attention. The first he calls onset attention, which generates saccades to new objection (also known as bottom up, or involuntary attention) – this is eliminated during the change blindness test. The second type is goal directed attention (aka top-down attention) which remains during the change blindness test. Wurtz will discuss this second form of attention, and the role of the superior colliculus. Results from experiments demonstrated a motor theory of attention which postulates that the mechanisms generating saccades to a target contribute to a shift of attention to that target – this theory was demonstrated particularly well by Tirin Moore’s 2002 experiments in stimulating FEF to shift attention and gaze.
The question for Wurtz examined was if the superior colliculus saccade generation activity contributes to attention modulation in cortex. They used a change blindness task, modulating SC activity to attempt to alter cortical processing of visual information. He first describes the paradigm of motion detection change blindness. Briefly, the subject fixates, is presented with 3 groups of dots that are moving in a direction, with out of those groups switching directions after a blank screen is shown (to induce change blindness). Cueing the participant to which group is likely to switch will shift visual attention, increasing the ability of the participant to detect the change in direction. If the visual cue is replaced by stimulation of the superior colliculus, they see a significant improvement in the ability of monkeys to detect change in the stimulus. So stimulating a neuronal component involved in saccade preparation produces a behavioral effect similar to attention. From this research, Wurtz notes with great satisfaction a cognitive function like spatial attention can be understood at the mechanistic level of neuronal activity.
Wurtz now moves on to the suppression of visual activity that occurs during the saccadic eye movement itself. The classic explanation involves corollary discharges, where information generated in sensorimotor processing is sent as a corollary discharge to other brain areas, where it informs those areas about the movement that is about to be made. Where are the saccade corollary discharges produced? Wurtz suggest it is the superior colliculus – he notes that they have identified two corollary discharge pathways: SC to MD to frontal cortex as well as SC to inferior pulvinar to occipital/parietal cortex. To summarize the research on this first pathway: the SC to FEF path provides a corollary discharge that acts of compensate for the displacement of the image on the retina. The second pathway may contribute to elimination of the blur from saccades (aka saccadic suppression).
Why look at superior colliculus as the origin of corollary discharge for saccadic suppression? Recordings from SC neurons show reduced responses to visual stimuli during movement. Furthermore, saccadic suppression starts before the saccade, and SC neurons demonstrate a decrease in responsiveness to visual stimuli immediately prior to saccadic movement. This resulted in a hypothesis stating that intermediate SC layers receive inhibitory inputs – slice recordings support this as occurring. But how does SC-mediated saccadic suppression get into the cortex? Recordings from MT neurons demonstrate saccadic suppression similar to the suppression seen in SC neurons. But are is this cortical suppression the result of the SC suppression? One problem is that it is not clear what the actual pathway is. Wurtz describes experiments that have been done to locate the pulvinar relate in the circuit between the SC and MT. He notes that they have found some relay neurons in a subregion of the inferior pulvinar that are connected to both SC and MT (this research done by Rebecca Berman and involved recording from pulvinar neurons while stimulating both SC and MT).
So does this circuit carry saccadic suppression? To test this, Berman has been recording MT neurons that show saccadic suppression while inactivating SC – her results show that in the absence of SC activity, there is an increase in MT neuron activity – suggesting that suppression in SC contributes of saccadic suppression in cortex. Wurtz notes that MT responses are a combination of SC and V1 inputs – inactivation of the SC removes one (inhibitory) input, leaving only the excitatory V1 inputs.
[Bloggers commentary: the role of the SC-pulvinar-MT connection in suppression of MT activity should perhaps be taken with a grain of salt. The exact components underlying the suppression of MT are not as clear as Wurtz suggested. The pulvinar is not an inhibitory area, so how it could be directly suppressing MT activity is not clear. One possibility is that inhibitory circuitry within in the SC is responsible but again, the exact players involved have not been identified. And intra-SC inhibition would not support their model of MT activity being the sum of SC and V1 inputs, with knocking out SC resulting in increased activity because of removal of inhibitory drive. Indeed, that model would require pulvinar to be responsible for inhibitory drive onto MT, something that is not supported by any currently described circuitry. If intra-SC inhibition is responsible, then you would expect that inactivation of SC would yield MT activation equal to drive from V1 (aka removal of additional excitatory drive), with no additional enhancement of MT (as might be expected if SC inactivation removed inhibitory drive, which Wurtz suggested is the case). In conclusion, the exact connections and role of the superior colliculus in attention and saccadic suppression have not yet been fully defined, End of commentary.]
To sum up his talk, Wurtz restates his claim that our perception of visual stability can be understood at the level of simple neuronal circuits. In conclusion, he notes that the brain circuits he has discussed all have the same basic structure, with their different functions dependent on what signals are conveyed and where those signals are directed. He notes that this structure suggests that corollary discharges and spatial attention can be viewed as variant of a general scheme. Finally, he point out the progress being made in elucidating the brain circuits that underlie cognitive processes for active vision. Still to be fully described are the mechanisms underlying decision for action, working memory, and rewards/values. He notes that understanding these cognitive functions depends on studying circuits at the highest levels of the cerebral cortex, and that understanding of these functions is necessary for comprehending the many diseases of the brain. He lends his support for research utilizing the monkey brain as an experimental model, stating the monkey brain provides unprecedented opportunities for understanding a wide range of cognitive processes and diseases.