We are able to see the world because our eyes transduce light into electrical activity of neurons. This process takes place in the retina, the layered tissue at the back of our eyes. Light is detected by millions of photoreceptors comparable to a pixel array. However, the information encoded in the light stimulus is not sent directly, pixel by pixel, to the brain but is preprocessed and compressed within the retina before being transmitted to the brain via the optic nerve. This process is highly sophisticated, as many different features are extracted from the visual stimulus. One such feature is the presence and direction of motion. If a small object moves in the visual scene, a specialized circuit within the retina computes its direction of motion and transmits this information to the brain.
In this project we want to build a computational model using Matlab which simulates the neural motion processing circuit and predicts how it responds to simulated light stimuli. We will then develop new light stimuli which can be used in experiments with real retinae in which we record responses of retinal neurons and compare them to the model predictions.
We are able to see the world because our eyes transduce light into electrical activity of neurons. This process takes place in the retina, the layered tissue at the back of our eyes. Light is detected by millions of photoreceptors comparable to a pixel array. However, the information encoded in the light stimulus is not sent directly, pixel by pixel, to the brain but is preprocessed and compressed within the retina before being transmitted to the brain via the optic nerve. This process is highly sophisticated, as many different features are extracted from the visual stimulus. One such feature is the presence and direction of motion. If a small object moves in the visual scene, a specialized circuit within the retina computes its direction of motion and transmits this information to the brain. In this project we want to build a computational model using Matlab which simulates the neural motion processing circuit and predicts how it responds to simulated light stimuli. We will then develop new light stimuli which can be used in experiments with real retinae in which we record responses of retinal neurons and compare them to the model predictions.
The student will have to familiarize himself with the literature describing motion-computation circuits in the retina. They will then implement a computational phenomenological model (i.e., not a biophysically realistic simulation) in Matlab. This model will take light stimuli in the form of movies as inputs and give the spiking activity of retinal neurons as output. The project will aim at designing new light stimuli which produce interesting neural output behavior. Those stimuli will then be used in experiments in which the real output of the neurons will be recorded and compared to the computational predictions.
The student will have to familiarize himself with the literature describing motion-computation circuits in the retina. They will then implement a computational phenomenological model (i.e., not a biophysically realistic simulation) in Matlab. This model will take light stimuli in the form of movies as inputs and give the spiking activity of retinal neurons as output. The project will aim at designing new light stimuli which produce interesting neural output behavior. Those stimuli will then be used in experiments in which the real output of the neurons will be recorded and compared to the computational predictions.