Week in Review: Number 21

Ophthalmic Imaging with Multimodal PARS Microscopy
Engineers in Canada developed a multimodal laser imaging system to measure oxygen saturation and metabolism in living tissue as a potential biomarker for common posterior segment eye diseases. Known as photoacoustic remote sensing (PARS), the technology uses multiwavelength lasers to "almost instantly" image in vivo human tissue noninvasively, with relevant applications for imaging of retinal vasculature and tissue prior to structural changes and functional loss in diseases such as age-related macular degeneration, diabetic retinopathy, and glaucoma. The technical, optical aspects of the technology implement stimulated Raman scattering (SRS), which occurs when the frequency of the emitted light differs from the frequency of the incident light through a material. Multiple wavelengths of light are necessary for oxygen saturation (relative concentration) calculations; however, previous methods using dye lasers or optical parametric oscillators were limited by low pulse repetition rates (i.e., speed) and thus were not suitable for applications of in vivo ocular imaging. Instead, the SRS approach allowed for both high-speed and multiwavelength light sources. The researchers used two excitation wavelengths (532 nm and 558 nm) in this case to estimate the concentration of HbO₂ and Hb. PARS microscopy can be simultaneously combined with swept-source optical coherence tomography (SS-OCT) for enhanced image viewing, and the authors report this being the first instance of combining OCT with optical resolution photoacoustic microscopy (OR-PAM). In contrast to OCT (without the angiography function), photoacoustic microscopy is good at imaging vasculature in greater detail. The project lead comments, “We’re optimistic that our technology, by providing functional details of the eye such as oxygen saturation and oxygen metabolism, may be able to play a critical role in early diagnosis and management of these blinding diseases.” The project is in an early stage and has only been tested in animal models at this point; they are working with several ophthalmologists and hope to start human trials within two years.

Cortical Recycling in High-Level Visual Cortex
A longitudinal study by researchers at Stanford University investigated cortical recycling in the visual cortex during childhood development. The study involved about 30 children ages 5 to 12 at their first MRI and followed with subsequent MRIs over 1 to 5 years. Specifically, the researchers used functional MRI to study areas in the ventral temporal cortex (VTC) that are stimulated by the recognition of images, using a sampling of ten categories of images ranging from faces and body parts to objects, words, and places. They found that areas in childhood (5-9-year-olds) that responded to images of limbs later responded to words and faces during adolescence. The researchers emphasized that increases in face- and word-selectivity in the VTC were directly linked to decreases in limb selectivity, providing surprising evidence of cortical selectivity being repurposed from one category to another, in contrast to prior theories of childhood brain development. The first author of the study comments, “This challenges a theory of cortical development, which states that new representations, like emerging regions involved in word recognition, are sculpted on previously uncommitted cortex. Our study suggests that during childhood, cortical selectivity can change from one category to another.” Word recognition becomes increasingly relevant as children learn to read. Research into vision development in the brain could thus inform strategies related to learning.

Research in Camouflage Breaking
Camouflage is used extensively by the military. From desert sand to dense jungles, visual textures and patterns are used to break up outline and conceal location. At the same time, camouflage breaking or being able to detect and localize another's camouflage is equally important, with real-world combat implications. For example, a sniper's missed shot also reveals his location, a difference between life and death in warfare. Scientists funded by the Army Research Office are studying how to train individuals to break camouflage. In the published study, six adult volunteers with normal or corrected-to-normal vision were trained to break camouflage using a deep-learning method similar to how computer scientists train self-driving cars. Specifically, the participants looked at digitally synthesized camouflage scenes like foliage or fruit, with each scene having a 50-50 chance of containing no target versus a camouflaged target like a human head. The participants could look for as little as 50 milliseconds or for as long as they wanted. After cleansing the visual palate with a random field of pixels, participants were then asked to both acknowledge whether there was a camouflaged target and identify where the target was on the screen based on memory. Interestingly, accuracy did not decrease much in the 50 millisecond viewing scenario as compared to the free viewing scenario. In a second experiment with seven different volunteers, the researchers tested a more abbreviated training protocol with more overt visual images, and found results similar to the more extensive training protocol. The researchers also plan to study the importance of context in camouflage breaking as well as explore using these techniques to identify medical problems. Beyond the military, research into camouflage breaking could also benefit the civilian sector, for instance, in image-intensive professions such as radiology as well as binocular vision in general.

Neurotrophic Effects of PEDF & Peptides in the Retina
Researchers at the National Eye Institute are studying how a growth factor called pigment epithelium-derived factor (PEDF) protects neurons from cellular stresses such as oxidative stress, inflammation, neovascularization, and cell death. To study the mechanisms behind PEDF's beneficial properties, the team used a cell culture model where immature retinal cells are isolated from the eyes of newborn rats and grown in vitro with minimal nutrients, alongside other types of cells in the retina. They discovered that PEDF has functionally distinct domains, which previous research showed can work independently of the full-length protein. One domain called the 34-mer (formed by 34 amino acid building blocks) halts blood vessel growth. Two other domains called the 44-mer and the 17-mer (a shorter version of the 44-mer) provide anti-apoptotic signals to retinal neurons. The researchers found that like full-length PEDF, both the 44-mer and the 17-mer could preserve photoreceptors from cell death, even in the absence of proteins and cells in the usual retinal environment. Additionally, their research showed that PEDF plays a role in photoreceptor development, triggering the movement of light-sensing opsin into the budding outer segment of photoreceptors where light detection takes place. Another finding showed that the 44-mer and the 17-mer could stimulate amacrine cells, another type of cell in the retina that relays visual information, to grow neurites, projections that facilitate neuronal communication; furthermore, these two fragments were at least as effective, or better, at stimulating these connections as the full-length protein. Finally, the researchers discovered PEDF's function in processing omega-3 fatty acids such as DHA, which is important for eye health both during infant development and for eye health over time. The authors conclude, "Our findings support the neurotrophic PEDF peptides as neuronal guardians for the retina, highlighting their potential as promoters of retinal differentiation, and inhibitors of retina cell death and its blinding consequences."

Optimizing Stimulation of Optic Nerve Fibers
Scientists in Europe developed a personalized protocol for optimizing intraneural stimulation of optic nerve fibers for the blind that takes into account feedback from the viewer’s brain. Although current neurotechnology in optic nerve stimulation can only provide simple visual signals, the researchers envision designing these simple visual signals to be meaningful in assisting the blind with daily living. Their idea is to stimulate the optic nerve to induce phosphenes, the sensation of light in one’s visual field. The current technology is limited by image resolution due to contraints of size difference between the optic nerve fibers compared to the electrodes used for intraneural stimulation. A challenge for neuroprosthetics in general, intraneural stimulation of optic nerve fibers is a greater difficulty due to the extreme complexity of visual signals. The model has thus far been tested on convolutional neural networks (CNN), artificial neural networks based on machine learning and used in computer vision for detecting and classifying objects. Psychophysical tests involving ten healthy subjects were also performed to imitate optic nerve stimulation, with results compatible with CNN, according to the authors. The first author of the study states, "Our study shows that it is possible to elicit desired activity patterns in deep layers of a CNN that simulate cortical visual areas." The researchers acknowledge that the project is tremendous, taken one step at a time. They are considering working with collaborators in Rome, Italy for future clinical trials.

In Other News
(1) Discovery of the "focea" in mice
(2) Neurobiology: How mice see the world
(3) Eyeblink test in piglets, for research in human infants