Seminar Details
2007-10-09 (14.00) : "Transforming visual signals into accurate 3D reach plans"
At Euler building
Organized by Mathematical Engineering
Speaker :
Blohm, Gunnar
Abstract :
Most of our interactions with objects surrounding us are guided by vision.
My research investigates how visual information about a desired reach target
is translated into a geometrically accurate 3D motor plan for the arm. To answer
this question, I used three different approaches. First, I built a geometrical model
which theoretically describes how extra-retinal eye and head position signals (or their absence)
affect this visuomotor transformation. Specific predictions from this model were compared to
behavioral human reaching performance, and revealed that the normally functioning brain uses
extraretinal signals to transform visual signals into a reach plan. This model is also important in the
analysis and interpretation of patient data where areas of the brain involved in these visuomotor
transformation pathways are damaged. Second, I designed a simple feed-forward neural network
to perform this visuomotor transformation. A detailed analysis of how this artificial neural network
solves the visuomotor conversion shows how the brain could potentially perform this transformation
and predicts which neural properties to look for in the involved brain regions. Finally, I identified neural
structures as well as the timing underlying the visuomotor transformation for reaching in humans using
functional brain imaging (magneto-encephalography). Taken together, these studies show why we need
an explicit visuomotor transformation, how this could be achieved by the brain, which brain structures are
involved in the different aspects and how they dynamically perform this visuomotor transformation from
visual input to the motor output.