Centre for Computational Neuroscience Berlin - Teilprojekt A 1: Investigating sub-millisecond dynamics through single-spike probabilities


State of the art, own contribution: Signal transduction chains provide the microscopic basis for sensory information processing and offer a rich repertoire of computational primitives that include threshold operations, low- and high-pass filtering and coincidence detection. Disentangling the elementary processing steps, assessing the functionally relevant input variables and identifying the dominant stochastic compoments is a formidable task. Unexpected progress can be made if one does not attempt to estimate the full relation between input signal and output response, but focuses on stimuli that cause the same final output (e.g., occurrence probability of single spikes). For an insect auditory system, this method of Comparing Iso-Response Stimuli (CIRS) has allowed us to disambiguate and precisely characterize two consecutive sub-millisecond processes in vivo [1,2]. Classical techniques such as poststimulus time histograms, reverse correlation or Wiener-series are constrained by spike jitter and cannot capture fine details of rapid neural processes. The resolution of the CIRS method, however, is only limited by the precision with which sensory (or electrical) stimuli can be applied. It thus allows one also to study noise sources and their consequences for neural variability and information processing. Objectives: In this project, the CIRS method will be extended and applied to other neural systems. First, we will include numerical techniques into CIRS to capture unknown intermediate non-linear processing steps. Second, we will use the technique to study synaptic signal transmission (in cooperation with project B2) and the interaction of time-structured inputs and single-cell dynamics (A2, B4). The general framework behind the CIRS-approach could also yield new insight into molecular signaling processes (B1) and the coupling between neural activity and metabolic/CBF responses (C2). Description of work: The first objective is in close reach and will be carried out using a set-up already established in the PI s lab. Apart from writing new online-software within the open-source OELprogramming environment, no conceptual or technical developments are needed. In addition, even a novice can learn the required electrophysiological techniques within a few weeks. Thus, this part of the
project is safe and well suited for a postdoc coming from physics, mathematics or computer science.
At the same time, interesting results can be expected, both regarding the specific auditory system and the extension of the CIRS framework. This will help the postdoc to get in contact with other projects (see objectives) and to start cooperations with the corresponding PIs or postdocs. Focusing on the most
promising new CIRS application(s), the postdoc should be able to write a specific grant proposal (e.g. DFG or VolkswagenStiftung), and then continue as an independent junior research group leader.

Principal Investigators
Brecht, Michael Prof. Dr. (Details) (Computational Neuroscience)

Duration of Project
Start date: 09/2004
End date: 08/2009

Publications

Machens CK, Gollisch T, Kolesnikova O, Herz AV., Testing the efficiency of sensory coding with optimal stimulus ensembles. Neuron. 2005 Aug 4;47(3):447-56.



Gollisch T, Herz AM., Disentangling sub-millisecond processes within an auditory transduction chain. PLoS Biol. 2005 Jan;3(1):e8. Epub 2005 Jan 4.


Last updated on 2020-10-03 at 16:46