A neuromorphic retina model that reproduces neuronal images of the retinal circuit in real time.

In the vertebrate retina, images are continuously transduced into graded voltage responses in the photoreceptor cell array and processed with electrochemical properties of the following neuronal circuits. The images represented with spatio-temporal distribution of graded voltages are transformed into action potentials in the ganglion cell array that are transmitted to the brain in parallel. Previous physiological and anatomical experiments have revealed fundamental response properties of, and corresponding synaptic connections among, major types of neurons. Most of those physiological observations have been obtained from retinas isolated from animals, or retinas in anesthetized animals while using simplified optical stimulations, e.g. spot of light, randomly modulated checker patters and so on. The images received by the retina under natural visual environment, however, are quite different from those stimulations, and are continuously influenced by eye movements. Therefore, the functions of retinal circuitries inferred from the physiological experiments have to be re-examined, envisaging such a real visual environment. On this background, we have developed a neuromorphic retina model based on previous physiological experiments. I demonstrate “virtual in vivo experiments” that reproduce neuronal images of major types of retinal neurons in real time, referring to possible applications of the model to robot vision and visual prosthetic.

Biography

Dr. Tetsuya Yagi obtained his B.Sci. in physics at Nagoya University, Japan and received his Ph.D. degree in medical science from Nagoya University in 1985. Following his study on retinal physiology as a Postdoctoral Fellow at the National Institute of Physiological Science, Japan and the Rockefeller University, USA, he joined School of Computer Sci. and Systems Eng., Kyushu Institute of Technology for start-up as an Associate Professor in 1990. He moved to Graduate School of Engineering, Osaka University in 2001 as a professor. His research interests include neurophysiology of visual systems, and neuromorphic visual systems and their applications.

Distinct roles of early parallel visual pathways in motion-detection

In many species, parallel visual pathways extract information about specific features of the visual scene. In Drosophila, distinct first order interneurons are already sensitive to different features of the environment. We take apart the underlying circuit and cellular mechanisms and test the distinct role of the underlying pathways in motion-guided behaviors.

Biography

Dr. Marion Silies is a group leader at the European Neuroscience Institute in Göttingen, Germany. Originally trained as a biologist, she started to work on fly vision as a postdoc in Clandinin lab at Stanford. Her work is combining in vivo two photon imaging and behavioral analyses with genetic manipulations to understand how specific computations are carried out in visual microcircuits. Work in the Silies lab is supported by an ERC starting grant as well as the Emmy Noether program.

The computation of motion direction in the retina: Recalculating

Direction selective retinal ganglion cells (DSGCs) respond to motion in one ‘preferred’ direction. The asymmetric response relies on an asymmetric circuit architecture. Surprisingly, DSGCs can reverse their directional preference following a short repetitive visual stimulation. We use this phenomenon to shed new light on the mechanisms underlying retinal direction selectivity.

Biography

Dr. Michal Rivlin obtained her B.Sc. with honors in Mathematics and Computer Science at the Hebrew University of Jerusalem. She did her graduate studies in the Interdisciplinary Center for Neural Computation at the Hebrew University, working in Prof. Hagai Bergman’s lab on neuronal activity in the normal and parkinsonian brain. Supported by a Revson Award from the Israel National Postdoctoral Program for Advancing Women in Science, a Human Frontier Science Program Long-Term Fellowship, and an Edmond and Lily Safra Fellowship in Brain Science, Dr. Rivlin carried her postdoctoral research in the Department of Molecular and Cell Biology at the University of California at Berkeley, in Prof. Marla Feller’s lab. There, she used electrophysiology to study the direction selective retinal circuit. In Sep 2013, Dr. Rivlin joined the faculty of the Department of Neurobiology at the Weizmann Institute. Research in her lab focuses on dynamic computations in the retina, their underlying mechanisms and how they are transferred to the brain. She received the ERC starter grant in 2017.

Dr. Rivlin is married to Yedidya and has four children David (15), Naomi (13), Miriam (11) and Itai (4).

Synaptic transfer between rod and cone pathways in the mouse retina

Synaptic transmission from rod bipolar cells (RBCs) to AII amacrine cells comprises transient and sustained components that, during light responses, convey information about visual contrast and luminance, respectively. It is unknown whether AIIs transmit both components to their postsynaptic targets. The input-output characteristics of AIIs also may depend on the level of network activation, which influences analog signaling in the rod pathway. We have examined these issues in mouse retina, first with electron microscopy to examine synaptic connectivity between RBCs, AIIs and OFF CBCs. Consistent with recent reports, we found that AIIs make about half of their inhibitory outputs onto one OFF CBC subtype (CBC2). Simultaneous whole-cell recordings confirmed that RBCs transmit transient and sustained signals to AIIs. Paired recordings between AIIs and CBC2s indicated that AIIs are also capable of transient and sustained release. Finally, we recorded from RBCs and CBCs simultaneously to examine how synaptic signals are transformed by the AIIs in-between. We found that AIIs transmit transient and sustained signals to (ON) CBC5 but primarily transient signals to (OFF) CBC2, depending on the level of network activity or AII depolarization: in a relatively depolarized steady-state, AIIs transmit both transient and sustained signals received from individual RBCs. These findings reveal complex synaptic features in the rod pathway that may underlie distinct transformations of ON and OFF signals during night vision.

Biography

Jeff Diamond received his B.S. from Duke University in 1989 and his Ph.D. from the University of California, San Francisco in 1994, where he studied excitatory synaptic transmission in the retina with David Copenhagen. During a postdoctoral fellowship with Craig Jahr at the Vollum Institute, he investigated the effects of glutamate transporters on excitatory synaptic transmission in the hippocampus. Dr. Diamond joined NINDS as an investigator in 1999 and was promoted to Senior Investigator in 2007. His laboratory studies how synapses, neurons and small circuits perform computational tasks required for visual information processing in the mammalian retina.

Functional circuitry of the primate retina at cellular resolution

A major challenge in neuroscience is to decipher the functional connectivity of neural circuits. We approached this challenge in the primate retina by mapping the flow of signals between the input and output layers at cellular resolution, and using these maps along with closed-loop experiments and computational methods to infer interneuron connectivity. Large-scale multi-electrode recordings were used to examine the activity of complete populations of the retinal ganglion cell types which collectively mediate high-resolution vision (midget, parasol, small bistratified). Fine-grained white noise visual stimulation was used to separately identify the location and spectral type of each cone photoreceptor providing input to each ganglion cell. This provided functional connectivity maps at cellular resolution between complete populations of input and output neurons. Closed-loop targeted stimulation of individual cones and pairs of cones was then used to test the functional properties and interactions of cone signals. Single cone simulation indicated that ganglion cell responses were univariant with respect to the inputs from different cones, and that maps obtained with white noise stimulation accurately revealed the strength of individual cone inputs. Paired cone stimulation revealed both linear and nonlinear interactions between cone signals, consistent with the pooling of cone inputs by intermediate bipolar cells within the ganglion cell receptive field.

Biography

E.J. Chichilnisky is the John R. Adler Professor of Neurosurgery at Stanford University, where he has been since 2013 after 15 years at the Salk Institute for Biological Studies. He received his B.A. in Mathematics from Princeton Univertsiry, and his M.S. in mathematics and Ph.D. in neuroscience from Stanford University. His research program focuses on understanding the spatiotemporal patterns of electrical activity in the retina that convey visual information to the brain, and their origins in retinal circuitry, using large-scale multi-electrode recordings. His research also involves physiological experiments with electrical stimulation and computational methods aimed at advancing the design of visual prostheses for treating blindness. He is the recipient of an Alfred P. Sloan Research Fellowship, a McKnight Scholar Award, and a McKnight Technological Innovation in Neuroscience Award.