Dr. David K. Merwine

Research Assistant Professor

Linear and nonlinear spatiotemporal processing in the retina; Activity- and sensory-dependent mechanisms of retinal development.

Curriculum Vitae (Word)
Phone: (213) 821-2072
Fax: (213) 821-2070
E-mail: <a href="mailto:dmerwine@bmsr.usc.edu" dmerwine@bmsr.usc.edu



  • 6/87 – 9/97 University of Alabama at Birmingham, Ph.D., Vision Science. Dissertation Thesis: “Analysis of the Center-Surround Interaction of Rabbit Retinal Ganglion Cells: Physiology and Pharmacology”.
  • 8/83 – 12/83 Kansai Gaidai University, Exchange Program.
  • 9/80 – 12/84 Juniata College, B.S., Pre-Medicine.


  • 9/01 – Present Research Assistant Professor, University of Southern California.
  • 9/96 – 9/01 Research Associate, Smith-Kettlewell Eye Research Institute.
  • 9/95 – 3/96 Consultant, Smith-Kettlewell Eye Research Institute.
  • 6/87 – 9/97 Graduate Fellow, University of Alabama at Birmingham.


My research has focused on retinal information processing. For my thesis, I examined the center-surround spatial interaction of retinal ganglion cells. These cells, the output neurons of the eyes, transmit a highly encoded signal, with different types encoding differing aspects of the visual world. The center-surround interaction provides each type with a mechanism for edge-enhancement, as well as gain control. I found a primarily non-linear center-surround interaction for all ganglion cell types in the rabbit. I have also collaborated on numerous studies of directionally selective (DS) retinal cells. DS cells provide critical information for the control of eye movements, attention, locating objects in space, etc. The retinal computation of directional selectivity is fascinating because it breaks several neural dogmas.

There are several practical applications for studying retinal processing. First, it is well known that there are pathological conditions that damage portions of the retina (retinitis pigmentosa and macular degeneration, for example), while leaving the central visual pathways intact. Therefore, several labs (including the newly established Retinal Institute at USC’s Keck School of Medicine) are developing prosthetic devices that will perform the lost retinal functions. Second, there have been tremendous gains in the development of artificially intelligent machines. Although there may be many ways to create a machine that can ‘see’, an obvious way would be to imitate a system that already exists. Thus, a detailed description of retinal functioning could have multiple applications.

Recently, I have expanded my retinal studies into the field of development, in part due to recent successes in retinal transplantation. During normal retinal development, ganglion cells ‘learn’ how large they should be, and how they should retinotopically connect to other brain structures. Imitating the normal developmental process could result in more successful transplants. Additionally, it may be a better strategy to allow a robotic eye to ‘learn’ visual functions rather than encode them. Thus, a detailed understanding of retinal development could prove invaluable in both clinical and AI applications.

Research Interests