Impact of Photoreceptor Loss on Retinal Circuitry

Our sense of sight relies on photoreceptors, which transduce photons into the nervous system's electrochemical interpretation of the visual world. These precious photoreceptors can be disrupted by disease, injury, and aging. Once photoreceptors start to die, but before blindness occurs, the remaining retinal circuitry can withstand, mask, or exacerbate the photoreceptor deficit and potentially be receptive to newfound therapies for vision restoration. To maximize the retina's receptivity to therapy, one must understand the conditions that influence the state of the remaining retina. In this review, we provide an overview of the retina's structure and function in health and disease. We analyze a collection of observations on photoreceptor disruption and generate a predictive model to identify parameters that influence the retina's response. Finally, we speculate on whether the retina, with its remarkable capacity to function over light levels spanning nine orders of magnitude, uses these same adaptational mechanisms to withstand and perhaps mask photoreceptor loss.

Neuromodulatory Control of Early Visual Processing in Macaque

Visual processing is dynamically controlled by multiple neuromodulatory molecules that modify the responsiveness of neurons and the strength of the connections between them. In particular, modulatory control of processing in the lateral geniculate nucleus of the thalamus, V1, and V2 will alter the outcome of all subsequent processing of visual information, including the extent to and manner in which individual inputs contribute to perception and decision making and are stored in memory. This review addresses five small-molecule neuromodulators—acetylcholine, dopamine, serotonin, noradrenaline, and histamine—considering the structural basis for their action, and the effects of their release, in the early visual pathway of the macaque monkey. Traditionally, neuromodulators are studied in isolation and in discrete circuits; this review makes a case for considering the joint action of modulatory molecules and differences in modulatory effects across brain areas as a better means of understanding the diverse roles that these molecules serve.

Morphology, Molecular Characterization, and Connections of Ganglion Cells in Primate Retina

The eye sends information about the visual world to the brain on over 20 parallel signal pathways, each specialized to signal features such as spectral reflection (color), edges, and motion of objects in the environment. Each pathway is formed by the axons of a separate type of retinal output neuron (retinal ganglion cell). In this review, we summarize what is known about the excitatory retinal inputs, brain targets, and gene expression patterns of ganglion cells in humans and nonhuman primates. We describe how most ganglion cell types receive their input from only one or two of the 11 types of cone bipolar cell and project selectively to only one or two target regions in the brain. We also highlight how genetic methods are providing tools to characterize ganglion cells and establish cross-species homologies.

Precision Medicine Trials in Retinal Degenerations

The beginning of the twenty-first century was marked by the innovative use of pharmacochemical interventions, which have since expanded to include gene-based molecular therapies. For years, treatment has focused on tackling the pathophysiology of monogenic orphan diseases, and one of the first applications of these novel genome editing technologies was the treatment of rare inherited retinal dystrophies. In this review, we present recent, ongoing, and future gene therapy–based treatment trials for choroideremia, X-linked retinitis pigmentosa, Stargardt disease, and age-related macular degeneration. As these trials pave the way toward halting the progression of such devastating diseases, we will begin to see the exciting development of newer, cutting-edge strategies including base editing and prime editing, ushering in a new era of precision medicine.

Mobile Computational Photography: A Tour

The first mobile camera phone was sold only 20 years ago, when taking pictures with one's phone was an oddity, and sharing pictures online was unheard of. Today, the smartphone is more camera than phone. How did this happen? This transformation was enabled by advances in computational photography—the science and engineering of making great images from small-form-factor, mobile cameras. Modern algorithmic and computing advances, including machine learning, have changed the rules of photography, bringing to it new modes of capture, postprocessing, storage, and sharing. In this review, we give a brief history of mobile computational photography and describe some of the key technological components, including burst photography, noise reduction, and super-resolution. At each step, we can draw naive parallels to the human visual system.

Functional Organization of Extraocular Motoneurons and Eye Muscles

Eye movements are indispensable for visual image stabilization during self-generated and passive head and body motion and for visual orientation. Eye muscles and neuronal control elements are evolutionarily conserved, with novel behavioral repertoires emerging during the evolution of frontal eyes and foveae. The precise execution of eye movements with different dynamics is ensured by morphologically diverse yet complementary sets of extraocular muscle fibers and associated motoneurons. Singly and multiply innervated muscle fibers are controlled by motoneuronal subpopulations with largely selective premotor inputs from task-specific ocular motor control centers. The morphological duality of the neuromuscular interface is matched by complementary biochemical and molecular features that collectively assign different physiological properties to the motor entities. In contrast, the functionality represents a continuum where most motor elements contribute to any type of eye movement, although within preferential dynamic ranges, suggesting that signal transmission and muscle contractions occur within bands of frequency-selective pathways.

The Certainty of Ambiguity in Visual Neural Representations

Some images evoke bistable percepts: two different visual experiences seen in alternation while continuously viewing an unchanged stimulus. The Necker Cube and Rubin's Vase are classic examples, each of which gives alternating percepts of different shapes. Other bistable percepts are alternating colors or directions of motion. Although stimuli that result in salient bistability are rare and sometimes cleverly constructed to emphasize ambiguity, they have been influential for over 150 years, since the work of von Helmholtz, who considered them to be evidence for perceptual visual processes that interpret retinal stimuli. While bistability in natural viewing is uncommon, the main point of this review is that implicit ambiguity in visual neural representations is pervasive. Resolving ambiguity, therefore, is a fundamental and ubiquitous process of vision that routinely affects what we see, not an oddity arising from cleverly crafted images. This review focuses on the causes of widespread ambiguity, historical perspectives on it, and modern knowledge and theory about resolving it.

Lexical Color Categories

Color is a continuous variable, and humans can distinguish more than a million colors, yet world color lexicons contain no more than a dozen basic color terms. It has been understood for 160 years that the number of color terms in a lexicon varies greatly across languages, yet the lexical color categories defined by these terms are similar worldwide. Starting with the seminal study by Berlin and Kay, this review considers how and why this is so. Evidence from psychological, linguistic, and computational studies has advanced our understanding of how color categories came into being, how they contribute to our shared understanding of color, and how the resultant categories influence color perception and cognition. A key insight from the last 50 years of research is how human perception and the need for communication within a society worked together to create color lexicons that are somewhat diverse, yet show striking regularities worldwide.

Quantifying Visual Image Quality: A Bayesian View

Image quality assessment (IQA) models aim to establish a quantitative relationship between visual images and their quality as perceived by human observers. IQA modeling plays a special bridging role between vision science and engineering practice, both as a test-bed for vision theories and computational biovision models and as a powerful tool that could potentially have a profound impact on a broad range of image processing, computer vision, and computer graphics applications for design, optimization, and evaluation purposes. The growth of IQA research has accelerated over the past two decades. In this review, we present an overview of IQA methods from a Bayesian perspective, with the goals of unifying a wide spectrum of IQA approaches under a common framework and providing useful references to fundamental concepts accessible to vision scientists and image processing practitioners. We discuss the implications of the successes and limitations of modern IQA methods for biological vision and the prospect for vision science to inform the design of future artificial vision systems. (The detailed model taxonomy can be found at http://ivc.uwaterloo.ca/research/bayesianIQA/ .) Expected final online publication date for the Annual Review of Vision Science, Volume 7 is September 2021. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Face Recognition by Humans and Machines: Three Fundamental Advances from Deep Learning

Deep learning models currently achieve human levels of performance on real-world face recognition tasks. We review scientific progress in understanding human face processing using computational approaches based on deep learning. This review is organized around three fundamental advances. First, deep networks trained for face identification generate a representation that retains structured information about the face (e.g., identity, demographics, appearance, social traits, expression) and the input image (e.g., viewpoint, illumination). This forces us to rethink the universe of possible solutions to the problem of inverse optics in vision. Second, deep learning models indicate that high-level visual representations of faces cannot be understood in terms of interpretable features. This has implications for understanding neural tuning and population coding in the high-level visual cortex. Third, learning in deep networks is a multistep process that forces theoretical consideration of diverse categories of learning that can overlap, accumulate over time, and interact. Diverse learning types are needed to model the development of human face processing skills, cross-race effects, and familiarity with individual faces. Expected final online publication date for the Annual Review of Vision Science, Volume 7 is September 2021. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.