Retinal direction selectivity encodes self-motion in vestibular coordinates


April 13, 2016 - 1:00pm
Northwest 243
About the Speaker
David Berson (Brown)

Self-motion triggers complementary visual and vestibular signals that cooperatively support image-stabilization and balance. The vestibular labyrinth biomechanically decomposes self-motion into translational (otolithic) and rotational (semicircular-canal) components. Vision presumably accomplishes a parallel decomposition, but it is unclear how. Here, through calcium imaging and recording in mouse, we show that all direction-selective neurons of the retina share a common vestibulocentric spatial geometry. Each type aligns its directional preferences with optic flow resulting from translation along a cardinal vestibular axis and responds best to that translation or to rotation around an orthogonal axis. There are 16 channels altogether: 2 eyes X 2 axes per eye X 2 directions per axis X 2 cell classes (ON-OFF and ON direction-selective). The pattern of relative activation among these channels uniquely encodes any combination of translational and rotational self-motion. The vestibulocentric reference frame for these retinal motion channels permits seamless integration with their vestibular analogs for encoding self-motion and for recruiting appropriate compensatory motor reflexes.