November 2002 I moved to London to do motor research together with Daniel Wolpert. I address Learning and the representation of uncertainty in the Movement system. Just as I am interested in probabilistic models for hearing I am interested in probabilistic models for moving.
In addition I am interested in the loss functions used by people and more particular of their sensorimotor system. In the newest experiments I adress how people could do reinforcement learning, that is learning with delayed feedback and a continous state space.
We want to find out what cats are interested in. To find out what they
look at we mount a miniture CCD to their heads.
We then have them
explore the local park and record videos of the world from a cat's perspective.
We then analyze how cat see using these cat-cam videos.
Dual Purkinje Eyetracking is known to be one of the fastest and at the same time most precise methods for the determination of the direction of gaze. Surprisingly it has yet almost exclusively been applied to humans. We modified the method so that it is possible to get reliable data even on cats. This is used as a tool for experiments with awake behaving cats. We also built some tools to synchronize eye-tracking data with electrophysiological recordings and the presentation of natural videos.
The following image shows the eyetracker from above.
In the beginning it was very difficult to calibrate the methods since cats
are not willing to fixate a set of points. Furthermore they have a fixation gain of less then 1. So we developed an alternative method for assessing the direction of gaze of cats. We modified the tangent screen method usually used for experiments on paralyzed animals so that it is now applicable to awake animals. Brief flashes of light are used to take photographs of the picture of the animals retina projected onto a tangent screen. The following picture shows such an example:
With precise eyetracking and calibration we are now able to do very controlled eye-movement research on cats.
For some time it has been possible now to record activity in rat hippocampus using tetrodes driven by little adjustable drives. We want to change this technique so that it is also applicable to cats. The problem is that it would need to be small (so that the cat is not annoyed by it). It does not need to be very lightweight since cats are way stronger than rats. At the same time we want it to be very well encapsulated to make sure we do not get any problems with infections. The left picture shows our current state: