The Neuronal Basis of Visual Self-motion Estimation

Transcription

1 The Neuronal Basis of Visual Self-motion Estimation Holger G. Krapp What are the neural mechanisms underlying stabilization reflexes? In many animals vision plays a major role.

2 Gaze and locomotor control: a common task across phyla Photographs: Courtesy Roland Hengstenberg

3 How does the visual system estimate self-motion? Self-motion in space causes optic-flow movie clip: Courtesy BBC 2004 Animal Camera

4 Self-motion in space causes optic-flow (slow motion) movie clip: Courtesy BBC 2004 Animal Camera

5 Snapshots of optic-flow fields frames: Courtesy BBC 2004 Animal Camera

6 Snapshots of optic-flow fields from Koenderik & van Doorn, 1987 frames: Courtesy BBC 2004 Animal Camera

7 Flying insects: moderately complex experimental model systems Quantitative Behavioural Experiment Neuroanatomy, Neurophysiology & Neurogenetics Numerical and Analytical Modelling

8 Optic-flow processing tangential cells in the fly from Krapp et al. 1998

9 Measuring neuronal responses to visual motion Krapp & Hengstenberg 1997, Krapp et al. 1998

10 Receptive field organization of VS-cells from Krapp and Hengstenberg, 1996

11 Preferred rotation axes of tangential cells, motorneurons, and neck muscles Huston & Krapp, submitted

12 Multisensory contributions to gaze stabilization from Hengstenberg 1991

13 We can study inputs to compound eye haltere & compound eye ocelli from Hengstenberg 1991, Parsons et al. 2006, Huston & Krapp in prep

14 We can not study the activity in tangential cells while the fly is actually moving at least not yet fly mounted on mobile robot platform steering Implantable micro recording probe => neural signals from tangential cells closed-loop experimental platform: the fly is steering the robot Ejaz, Peterson & Krapp, work in progress photograph Peterson/Drakakis

15 Questions we will be able to address: Are the signals recorded in animals moving in space different from those recorded under restrained conditions? What impact do other sensory modalities have on opticflow processing under closed-loop conditions? Does the fly employ efference copies forward models - to facilitate self-motion estimation?

16 Expected technological Advances: Nervous-system machine interfaces: => Low-power, implantable micro recording probes Bio-inspired technologies: => Optic-flow based self-motion estimation => Sensor-rich feedback control systems based on multisensor fusion and dynamic range fractioning

17 Acknowledgements MPI f. Biologische Kybernetik: Roland Hengstenberg Bärbel Hengstenberg Matthias Franz Karin Bierig Tuebingen University: Hansjürgen Dahmen University of Oxford: Graham Taylor Bielefeld University: Martin Egelhaaf Ralf Petrowitz Katja Karmeier Rico Tabor University of Cambridge: Simon Laughlin Stephen Huston Matthew Parsons Imperial College: Manos Drakakis Simon Schultz Aman Saleem Kit Longden Daniel Wuestenberg Naveed Ejaz Kris Peterson AF Research Laboratories

18 Receptive fields of tangential cells and neck motorneurons Huston & Krapp, submitted

19 The Neuronal Basis of Visual Self-motion Estimation Holger G. Krapp What do flies and jet fighters have in common? They require powerful feedback control systems to stay airborne

20 Receptive fields of optic-flow processing cells Wuestenberg & Krapp, unpublished

21 Annual Meeting ION, Cambridge, MA April 2007 Perspectives We have started working on multisensory integration The Tangential Cells in flies set up a highly non-othogonal coordinate system which facilitates the transformation of visual self-motion information into motor commands. Tangential Cell Signals may be scaled and gated by other sensory mechanisms encoding the same selfmotion parameter. Flies vs. Jet fighter: The fly - like many other flying insects make use of sensor-rich feedback control and multisensor fusion while jet-fighters carry several super-computers on board to work out the appropriate control signals.

22 Ocellar Stimulation in the Fly Parsons et al. 2006

23 Haltere-induced signals combined with visual input induce spiking in neck motorneurons Activation of certain muscles requires simultaneous input from the halteres and the compound eye Huston & Krapp, in preparation

24 Ocellar Stimulation Modulates Spiking Activity in Identified Tangental Neurons Parsons et al. 2006

25 Annual Meeting ION, Cambridge, MA April 2007 Receptive Field Organization of the Spiking Tangential Neuron V1 Parsons et al., 2006

26 Compound eye workshop Tucson Nov 2006 work on flies: Agenda Multisensory integration & sensorimotor transformation Efference copy in decoding tangential neuron responses? Efference copy and voluntary movements Flight control in flies

27 Compound eye workshop Tucson Nov 2006 work on flies: Agenda work on locusts: Multisensory integration & sensorimotor transformation Efference copy in decoding tangential neuron responses? Efference copy and voluntary movements Flight control in flies Approach detection: plasticity in an identified neuronal circuit Biophysical mechanisms of neuronal multiplication Optic flow processing in locusts Flight control in locusts Application of biological design principles of processing optic flow in engineering and robotics

28 Compound eye workshop Tucson Nov 2006 Gaze Stabilization in the Fly Photographs: Courtesy Roland Hengstenberg

29 Receptive field organization of HS-cells from Krapp 2000, Krapp unpublished. Reconstructions, courtesy Klaus Hausen

30 Annual Meeting ION, Cambridge, MA April 2007 Preferred rotation axes of tangential cells and motorneurons Huston & Krapp, 2005

31 Annual Meeting ION, Cambridge, MA April 2007 The Fly Visual System from Hausen, 1982

32 Compound eye workshop Tucson Nov 2006 Preferred Rotation Axes of VS Neurons Krapp 2000

33 Annual Meeting ION, Cambridge, MA April 2007 Quantitative Behavioural Experiment Neuroanatomy, Neurophysiology & Neurogenetics Numerical and Analytical Modelling Photograph: Courtesy Roland Hengstenberg

34 Annual Meeting ION, Cambridge, MA April 2007 Parsons, unpublished

35 Compound eye workshop Tucson Nov 2006 Receptive Field Organization of Tangential Neurons Krapp et al., 1998

36 Annual Meeting ION, Cambridge, MA April 2007 Responses before and after Cauterization of the Ocellar Nerve Parsons et al. 2006

37 Annual Meeting ION, Cambridge, MA April 2007 The Wide-Field Tuning of the VS6-cell Flimax from Lindemann et al The receptive field organization of Tangential Cells indicates their preferred axis of rotation! Karmeier et al. 2005

38 Edinburgh 2005 Preferred Rotation Axes of Motorneurons, VS-cells, and HS-cells Huston & Krapp, unpublished

39 Compound eye workshop Tucson Nov 2006 Multisensory Integration at the Level of Descending Neurons from Hengstenberg 1991

40 Neuroinformatics Colloquium December 2005 Measuring Local Preferred Directions in Neck Motor Neurons Huston & Krapp, unpublished

41 Compound eye workshop Tucson Nov 2006 Integration of Binocular Motion Information VCH Krapp et al. 2001

42 Compound eye workshop Tucson Nov 2006 Receptive Field Organization and the Orientation of Ommatidial Rows data from Krapp et al. 1998, Petrowitz et al 2000

43 Compound eye workshop Tucson Nov 2006 Further Properties of Tangential Neurons Responses to wide-field stimuli mimicking translation and rotation to the fly (Karmeier et al., 2003) Local directional tuning is resistant to motion adaptation (Card et al., 2004) Performance of the VS-neuron population: a Bayesian approach (Karmeier et al., 2005) Velocity tuning of response fields (Baden & Krapp, in preparation) How important is the orientation of ommatidial rows for the estimation of self-motion? (Smolka & Krapp, in preparation)