Week in Review: Number 34

Tripartite Neuron Circuit Governs Representation Consistency in the Mouse Visual Cortex
Visual processing in the brain is accompanied by a lot of noise as different groups of neurons in the visual cortex (V1) are stimulated by the same scene. Neuroscientists at MIT explored how the brain compensates for this neuronal noise to produce fidelity in visual representation. The researchers used optogenetics and two-photon microscopy to observe the brain activity of mice watching movies. They discovered that in addition to groups of excitatory pyramidal neurons that respond when images appear, two groups of inhibitory neurons work in a circuit to enforce reliability behind the scenes. In particular, low representation reliability corresponded with high activity among parvalbumin-expressing (PV) inhibitory neurons and low activity among somatostatin-expressing (SST) neurons. High representation reliability corresponded with low PV activity and high SST activity. When excitatory activity became unreliable, SST activity followed PV activity.

The PV neurons inhibit excitatory activity to control their gain, that is, to prevent them from becoming saturated amid a flood of incoming images. However, gain suppression comes at the cost of making the representation of images less reliable. Meanwhile, the SST neurons can inhibit the activity of PV neurons, the SST neurons kicking in when excitatory activity has become unreliable. This tripartite neuron circuit was then modeled with computer software. Once they learned how the circuit worked, the researchers could control the inhibitory cells to influence how consistently the excitatory neurons represented images. For example, increasing PV activity made any existing reliability less reliable, and increasing SST activity made unreliable activity more reliable. Importantly, because SST inhibitory neurons work by inhibiting PV neurons—in other words, SST neurons increase reliability by inhibiting inhibition—they would not be able to enforce reliability without PV neurons. It is an inherent tripartite circuit, due in part to differences in how these two cells form synapses. (SST neurons form inhibitory synapses at neuron dendrites while PV neurons form inhibitory synapses at neuron somas.) The authors remark, "[I]t is the co-operative dynamics between SST and PV [neurons] which is important for controlling the temporal fidelity of sensory processing." They add that the circuit also interacts and receives input from other higher order brain areas, for example, to concentrate on a particular scene with volitional attention.

Binocular Connectivity with Selective Monocular Function in the Retinogeniculate Pathway
The visual thalamus—in humans, comprised of the lateral geniculate nucleus (LGN), the thalamic reticular nucleus (TRN), and the pulvinar—relays visual stimuli from the retina to the primary visual cortex. The retinogeniculate pathway, involving the LGN, is of greatest interest to visual perception. According to a long-standing theory, signals from the two eyes are separated in the visual thalamus before being transmitted to the visual cortex. However, investigations into the visual thalamus of mice showed that a large number of neurons form synapses with retinal ganglion cells (RGCs) of both eyes. Using optogenetics, researchers confirmed this finding. However, they also discovered that very few of the neurons actually receive equal signal strength from both eyes. Rather, in most neurons, signal strength from one eye dominates over the other, with the weaker signals from the non-dominant eye not reaching threshold to trigger an action potential. The researchers hypothesize that they do not play a major role in processing visual information. In other words, although demonstrating structural binocular connectivity, these cells in the mouse visual thalamus were functionally monocular, resolving the contradictory findings in the earlier study. From a basic science standpoint, the findings showed that anatomy does not necessarily confer function; in this case, neurons in contact with one another do not necessarily communicate extensively. From the perspective of function, the study showed that even with equal access to both eyes, thalamic cells selectively establish functional connections with only one eye. Monocular connections are explained by synaptic selection rather than by anatomical overlap between RGC axons and dLGN neuron dendrites alone. Furthermore, synapses with the dominant eye strengthen while those of the nondominant eye remain immature. The scientists next plan to study how eye dominance is determined, whether there are variations in different RGC types, and whether the immature contact sites could be activated when needed, for example to treat eye conditions such as amblyopia.

Automated Pupillometry Improves Care in Neuro ICU
Assessing pupil size and reactivity provides vital information about the care of patients in the neurology ICU. Although manual methods based on visual estimates via penlight remain established practice, in large part due to convenience and equipment availability, they also present obstacles in terms of reliability, varying among operators and patient eye colors. Intubation and intravenous lines blocking a clear line of sight, as well as cases of sedation and pharmacologic paralysis, can also present challenges. The Neurology Department at Vanderbilt University Medical Center is assessing the use of an automated handheld pupillometer by NeurOptics for more accurate and objective measurements of pupillary size, shape, and light reactivity. These measurements are quantified as a Neurological Pupil index (NPi), which provides valuable information regarding subtle changes before visible changes in pupil size are seen. As an assistant professor of the program remarks, “The impact of accurate versus slightly inaccurate serial measurement can be profound...Pupillometry may allow for more urgent medical or surgical intervention in some of our sickest patients, particularly those with rising intracranial pressures that can result in permanent damage, like brain herniation.” Similarly, they report that quantitative pupillometry outperformed manual measurements in comatose cardiac arrest patients, especially in predicting unfavorable outcomes, as well as successfully identified early changes in neurologic function, intracranial pressure, and treatment response to osmotherapy, among other neurologic conditions. Automated pupillometry, even at VUMC, is not indicated for all patients, but rather for patients with the highest risk, such as patients with intracranial injury or patients who experience neurologic decline. The team anticipates automated pupillometers becoming standard of care for units that treat trauma and critical illness.

Representational Drift in Neurons of the Visual Cortex
It was previously thought that neurons of the visual cortex reliably encode information from sensory stimuli. However, experiments in mice found evidence of "representational drift" of neuronal activity, known to occur in other areas of the brain, to also happen in the visual cortex (V1). Furthermore, while flexibility in neuronal activity is expected in response to changes in learning or experience, what was surprising was that neurons in the visual cortex exhibited drift even when presented with the same stimulus across time. The experiments involved showing mice a short 30-second movie clip on a loop, while the activity of hundreds of neurons were recorded in their primary visual cortex using two-photon calcium imaging. These viewing sessions were repeated weekly for up to seven weeks. Analysis of the neuronal population activity showed that individual neurons did not respond the same way to the visual stimulus, i.e., the exact scene at specific moments in the film, when the mouse watched the film one week as compared with another week. Specifically, the single-neuron activities consisted of individual spiking episodes, and it is these spiking episodes that showed distinct variations across weeks. In other words, the long-term stability of single-neuron responses varied across the weeks. As the first author of the study argues, “What is somewhat unexpected is that even when there is no learning, or no experience changes, neural activity still changes across days in different brain areas.” Although an implicit variable that is difficult to dissociate is the learning and experience changes that occur across the weeks beyond the time periods when activity is recorded in the lab, underlying the concept of representational drift is the thought that representations, or identities, of stimuli are expected to remain stable across the life of the organism. Discovering that neuronal encoding of stimuli varies across time could point to compensatory mechanisms in downstream brain areas to maintain stable representations.

Framing Strategies Stabilize Visual Perception
When we move our eyes to look around a scene, the image of the world projected onto our retina also moves, which in turn changes the image that is sent to the brain. Yet, our brains perceive a stabilized view of the world. Psychologists studying this process, known as "visual stabilization," tested the brain's inherent "steady cam" through two experiments, conducted both in person and online. The first experiment is an example of “paradoxical stabilization,” in this case demonstrating how a stabilized view is produced by a moving frame replicated on a computer monitor. A second experiment demonstrated that participants perceived the location of objects in relation to (moving) frames, even when the objects themselves were stationary. This effect held true across a wide range of frame speeds, sizes, and path lengths tested. In other words, even though the image of what we see "moves" on our retina when our eyes move, (1) the presence of a frame in the scene stabilizes our judgments of location, and (2) our brain has a tendency to perceive objects in relation to available frames. As the senior author of the study states, “Our results show that a framing strategy is at work behind the scenes all the time, which helps stabilize our visual experience.”

Social Interaction as Seen through Pupils
As part of the autonomic sympathetic nervous system, the pupils of our eyes change size in response to arousing emotional stimuli. Psychologists in China were interested in exploring whether pupil size could also serve as a window to social motivations, particularly social interaction. In a set of three experiments, the scientists tested pupillary response to second-person social interaction, observation of third-person social interaction, and communicative social interaction. They found that pupil size enlarged when the participants perceived a single agent sending interactive intention toward them as compared to toward others, and when they viewed two agents interacting with each other as compared to facing away from each other. These pupil dilation effects, however, relied on correct understanding of the communicative intentions of the interacting agents. The experiments show that, perhaps related to well-documented relationships between pupils and emotional response, pupil size could also indicate perception of social interaction. In brief, humans are social beings possessing intrinsic motivations toward interacting with others and maintaining social relationships. The study adds insight to the visual system's sensitivity to social interaction, with pupils being a potential biomarker for early detection of social cognitive disorders.

In Other News
(1) NJIT secures $3.7 million to study concussion-induced eye disorder (Related)
(2) Hippocampus could help guide visuospatial attention and short-term memory
(3) Diversity in pupil shapes of frogs and toads