Background and state of the art
Diffusion Tensor Imaging (DTI) is a new 3-D imaging technique that measures the diffusion of water molecules in human tissues. The directional dependence of water diffusion rates is closely related to the structural anisotropy of the medium, thus it can be used to infer the organisation and orientation of tissue components. For example, the DTI measurements allow one to track neural connections in the white matter of the brain.

This relatively new technique has generated much enthusiasm and high expectations, because it presently is the only approach available to non-invasively study the three-dimensional architecture of white matter tracts, and quantify physical and geometric properties of neuronal fibres in vivo.

Unfortunately, current image processing and computer vision algorithms are unable to take full advantage of what DTI offers. In addition, the validity of new DTI discoveries and the precision of tract estimates have yet to be established and a comparison and integration of DTIs with other imaging techniques is yet lacking. This is particularly due to the fact that advanced registration techniques needed to compensate for spatial distortion are under-developed.
During the last years, the Department of Mathematics at Lübeck has become one of the leading centres for image registration worldwide. The group has large expertise not only in the mathematical basis of registration, but also in design and software engineering and in a large variety of applications. Applications include CT, MRI, fMRI, non-invasive surgery and histology. Emphasis is placed on multi-modal registration in collaboration with the "Stanford Vision and Imaging Science and Technology" group of Prof. Brian Wandell on fMRI images and diffusion tensor imaging.

The project
In this project it is planned to generalise registration methods to DTI images and register DTI images to other modalities, in order to gain insight into DTI imaging, validate new results obtained with DTI, and to make DTI to a more accessible imaging modality.

Investigators
Clinical	Life Sciences	Informatics
C. KochInstitut für Neuroradiologie Contact: koch (at) neuroradiologie.uni-luebeck.de	—	B. FischerInstitut für Mathematik Contact: fischer (at) math.uni-luebeck.de
		J. ModersitzkiInstitut für Mathematik Contact: modersit (at) math.uni-luebeck.de

Key Literature

• Modersitzki, J.: Numerical Methods for Image Registration. Oxford University Press, 2004
• Fischer, B. and Modersitzki, J.: A unified approach to fast image registration and a new curvature based registration technique. Linear Algebra and its Applications 380:107-124, 2004

Magnetic Resonance Imaging (MRI) and Functional Magnetic Resonance Imaging (fMRI) provide the basis for a very large number of clinical applications. In both cases images are subject to distortions caused by the magnetic susceptibility of the patient.
Recently a research group at the host university has discovered a method for computing the distortions of the static field in MRI. This method is based on multi-grid methods, and has been successfully applied to the correction of the geometric distortions in T1-weighted images.

Based on new registration methods, it was shown that the statistic measure of similarity between CT images and T1-MR images of the same anatomy increased after applying this new method. The goal of this project is to apply the techniques developed to fMRI images, in order to allow for the use of fMRI in navigation. In the case of fMRI, distortions are much larger and more complex than in T1-MRI.

Investigators
Clinical	Life Sciences	Informatics
C. KochInstitut für Neuroradiologie Contact: koch (at) neuroradiologie.uni-luebeck.de	—	A. SchweikardInstitut für Robotik und Kognitive Systeme Contact: schweikard (at) rob.uni-luebeck.de
D. PetersenInstitut für Neuroradiologie Contact: dirk.petersen (at) uni-luebeck.de

Key Literature

• Burkhardt, S., Schweikard, A. and Burgkart, R.: Numerical Determination of the Susceptibility Caused Geometric Distortions in Magnetic Resonance Imaging. Medical Image Analysis, 7(3):221-236, Elsevier, Amsterdam, 2003.

Background and state of the art

Optical Coherence Tomography (OCT) is a new high-resolution, real-time imaging modality. Institutes of Lübeck University have been instrumental in developing this emerging technology, which is now beginning to gain recognition in laboratories world-wide. We have shown that OCT is able to capture the difference between white and grey brain matter.

Stereotactic implantation of micro-electrodes into the brain is subject to inaccuracies. Minimally-invasive imaging based on optical coherence tomography can improve the accuracy of electrode placement, by giving feedback imaging information in vivo. Based on optical coherence tomography, online identification of tumours during biopsy will be possible. Additionally, information about vessels in the planned approach can reduce the risk of intraoperative bleeding.

The project
The goal of the project is to investigate Optical Coherence Tomography (OCT) as a minimal-invasive imaging modality for deep brain navigation and diagnosis in neurological disorders.
The goals of the project are:

1. Development of endoscopic OCT, thus OCT that can be placed inside a thin endoscopic needle.
2. Development of signal processing methods for tissue identification and classification.
3. Integration of OCT in a stereotactic environment.
4. Integration of OCT into neuronavigation systems.

Investigators
Clinical	Life Sciences	Informatics
M. BonsantoKlinik für Neurochirurgie Contact: matteo.bonsanto (at) uk-sh.de	—	Y. BaillyInstitut für Robotik und Kognitive Systeme Contact: bailly (at) gradschool.uni-luebeck.de
V. TronnierKlinik für Neurochirurgie Contact: volker.tronnier (at) uk-sh.de		R. BirngruberInstitut für Biomedizinische Optik Contact: bgb (at) bmo.uni-luebeck.de
		G. HüttmannInstitut für Biomedizinische Optik Contact: huettmann (at) bmo.uni-luebeck.de
		A. SchweikardInstitut für Robotik und Kognitive Systeme Contact: schweikard (at) rob.uni-luebeck.de

Key Literature

• Ramrath, L., Füllgraf, H., Winter, C., Hofmann, U., Hüttmann, G., Moser, A. and Schweikard, A.: 3D Visualization and Identification of Rat Brain Structures using Optical Coherence Tomography, 5. Jahrestagung der Dt. Gesellschaft für Computer- und Roboterassistierte Chirurgie, Hannover, 2006
• Ramrath, L., Hofmann, U., Hüttmann, G., Moser, A. and Schweikard, A.: Towards Automated OCT-based Identification of White Brain Matter, Bildverarbeitung für die Medizintechnik, München, 2007

Background and state of the art
In combination with miniaturised PET and SPECT scanners, micro CT has recently become an important modality for small animal molecular imaging studies. Very recently, X-ray tubes with a nanometre focus have reached the market. The development of nano CT scanners is a thus consequent step following the development of micro CT. However, currently the application of nano CT is restricted to the analysis of in vitro specimens because the X-ray dose exposed to a living animal would be unacceptably high, if the same signal-to-noise ratio is expected as for clinical CT. Indeed, the miniaturisation of the X-ray detector elements in nano CT requires a 4th power increase of X-ray dose.

The project
In commercially available micro and nano CTs, the Feldkamp reconstruction technique is based on the assumption that the projection data is related to the Radon transform of the object. Unfortunately, such modified filtered back projection works with a simplified model of the underlying physical X-ray absorption processes, which often leads to undesired image artefacts. In this project, alternative reconstruction principles are developed. Due to the poor photon statistics measured with the extremely small detector elements, reconstruction principles must be implemented that take the Poisson static into account. Additionally, scatter processes and beam hardening should be modelled within a maximum likelihood approach. It is expected that the integration of a physically correct model of the X-ray absorption and detection process into the reconstruction algorithm will improve the image quality significantly. This will allow for reducing the dose such that living animal nano CT will become feasible.

Investigators
Clinical	Life Sciences	Informatics
S. GottschalkInstitut für Neuroradiologie Contact: stefan.gottschalk (at) uk-sh.de	—	T. BuzugInstitut für Medizintechnik Contact: buzug (at) imt.uni-luebeck.de

Key Literature

• Oehler, M., Weber, S., Musmann, P., and Buzug, T.M.: Image Reconstruction based on Fourier-Rebinning for a Novel High Resolution ClearPET-Neuro Scanner, in: H. Lemke et al. (Eds.), Computer Assisted Radiology, 131-136, Elsevier, Amsterdam, 2005
• Buzug, T.M.: Computed Tomography: From Photon Statistics to modern Cone-Beam CT, Springer, 2007, to appear

Background and state of the art
Despite its normal appearance on CT and MRI scans, the substantia nigra (SN) shows a distinct hyperechogenic pattern on transcranial sonography (TCS) in about 90% of patients with Parkinson�s disease (PD). In atypical Parkinsonism, such as multiple system atrophy and progressive supranuclear palsy, the SN is not different from healthy volunteers but the lentiform nucleus (LN) shows an abnormal hyperechogenic ultrasound pattern. The combination of clinical characteristics and the ultrasound pattern assists in establishing the correct diagnosis of a specific movement disorder. Monogenic forms of Parkinsonism may be clinically indistinguishable from PD. We described SN hyperechogenicity even in asymptomatic carriers of single heterozygous Parkin mutations with or without PET abnormalities. This interesting finding leads to the hypothesis that SN and LN ultrasound patterns may be a potential preclinical marker to detect PD susceptibility. To date, for quantitative analysis of SN and LN hyperechogenicity, only the area of SN (aSN) and the detection of LN hyperechogenicity but not other signal characteristics have been considered.

The project
There are several open questions regarding ultrasound technology in genetic and nongenetic forms of parkinsonism:

1. Is there information other than aSN or LN hyperechogenicity in the ultrasound signal from the mesencephalic and diencephalic ultrasound images to characterise distinct forms of parkinsonism/PD?
2. Is the ultrasound investigation useful to screen a large population for genetic and other forms of PD?
3. Can SN hyperechogenicity serve as a preclinical marker?

Computer methods for automated investigator-independent detection of the SN and LN will be developed. After development of this tool for the detection of aSN and LN hyperechogenicity, a statistical image processing tool will be created to detect specific characteristics of the images.

Investigators
Clinical	Life Sciences	Informatics
C. KleinKlinik für Neurologie Contact: christine.klein (at) neuro.uni-luebeck.de	—	U. HofmannInstitut für Signalverarbeitung und Prozessrechentechnik Contact: hofmann (at) isip.uni-luebeck.de
G. SeidelKlinik für Neurologie Contact: guenter.seidel (at) neuro.uni-luebeck.de		C. KierInstitut für Signalverarbeitung und Prozessrechentechnik Contact: kier (at) isip.uni-luebeck.de

Key Literature

• Walter, U., Behnke, S., Eyding, J., Niehaus, L., Postert, T., Seidel, G. and Berg, D.: Transcranial brain parenchyma sonography in movement disorders: State of the art. Ultrasound Med Biol 33:15-25, 2007
• Hagenah, J.M., Hedrich, K., Becker, B., Pramstaller, P.P., Seidel, G. and Klein, C.: Distinguishing early-onset PD from dopa-responsive dystonia with transcranial sonography. Neurology 66:1951-1952, 2006
• Kier, C., Cyrus, C., Seidel, G., Hofmann, U.G., and Aach, T.: Segmenting the Substantia nigra in ultrasound images for early Parkinson diagnosis. Accepted with Computer Assisted Radiology and Surgery, Berlin, 2007

Background and state of the art
Motion artefacts are inevitable in most medical imaging modalities, and deteriorate image quality. Patient movement and pulsation constitute serious problems for reconstruction. The movements cause misalignment of the projection frames, which degrades the reconstructed image and may introduce artefacts. Currently, little is known on general schemes which both detect a movement and correct for such artefacts. We will use single photon emission computed tomography (SPECT) as an example for new correction methods. In SPECT, the imaging time is in the range of 5-30 minutes, and large motion artefacts are typical.

The project
We will investigate a new promising methodology for correcting motion artefacts. The correction is performed within the image space. The aim is to come up with a flexible framework for the incorporation of motion information in the projector/back projector pair of the reconstruction. In addition, the framework should be independent of the used camera system and/or the actual motion states in the measured data. Furthermore, it is planned to develop an overall scheme, which combines both motion estimation and reconstruction. The main driving engine will be a mathematical optimisation approach based on a multi-scale and multi-level strategy.

Investigators
Clinical	Life Sciences	Informatics
S. GottschalkInstitut für Neuroradiologie Contact: stefan.gottschalk (at) uk-sh.de	—	B. FischerInstitut für Mathematik Contact: fischer (at) math.uni-luebeck.de
		J. ModersitzkiInstitut für Mathematik Contact: modersit (at) math.uni-luebeck.de

Key Literature

• Röhl, E, Schumacher, H and Fischer, B: Automatic detection of abrupt patient motion in SPECT data acquisition, to appear in Proceedings of SPIE 2007, Medical Imaging, San Diego, 2007
• Fulton, R and Meikle, S: Reconstruction of projection data corrupted by rigid or non-rigid motion, in Conference Record of the 2005 IEEE Nuclear Science Symposium and Medical Imaging Conference, San Juan, 2005

Background and state of the art

Glutamate and GABA are the two most important neurotransmitters, distributed across the whole brain. Imbalances in regional glutamatergic and GABAergic neurotransmission lay at the core of several important psychiatric and neurological disorders, including dementia, schizophrenia, depression, anxiety disorders or Morbus Parkinson. Using in-vivo brain magnetic resonance spectroscopy (MRS) at 3 Tesla, we were able to reliably quantify glutamate concentrations and, with some restrictions, GABA concentrations and their dynamic changes in the course of treatment in patients with obsessive-compulsive disorder. Our longitudinal reliability measurements in healthy subjects show a high stability of glutamate levels in cortical and subcortical regions. The current standard of MRS measures of the brain is single voxel spectroscopy. The distribution of concentrations of neurotransmitters and metabolites across the whole brain is unknown in humans, if derived from single-voxel measurements. In the present project, whole-brain maps of glutamate and GABA will be acquired, while addressing the abovementioned shortcomings of MRS glutamate measurements.

The project

In this study, MRS measurements will be performed in healthy subjects in order to acquire a neurotransmitter distribution map of the brain using in-vivo MRS. Repeated measurements at 3 Tesla will be performed in the same subjects in order to address intra-individual and inter-individual variability. Spectrums will be acquired at a spatial resolution as high as 10 mm, using 10x10x10 mm as voxel size. GABA-Spectrums will be acquired using an additional sequence with GABA editing, as it is usually covered by other prominent peaks. Average concentration distributions will be projected onto a standard brain and visualised via colour codes, similar to that known from functional neuroimaging. The technique of whole-brain MRS mapping in healthy subjects will be used in subsequent basic neuroscience and clinical projects. Comparisons of healthy subjects and psychiatric patients as well as measurement of dynamic changes of the glutamate (and GABA) distribution due to pharmacological interventions will be based on image registration and analysis tools.

Investigators
Clinical	Life Sciences	Informatics
B. ZurowskiKlinik für Psychiatrie und Psychotherapie Contact: bartosz.zurowski (at) psychiatrie.uk-sh.de	—	F. BinkofskiKlinik für Neurologie Contact: ferdinand.binkofski (at) neuro.uni-luebeck.de
		A. SchweikardInstitut für Robotik und Kognitive Systeme Contact: schweikard (at) rob.uni-luebeck.de
		W. Weber-FahrInstitut für Systemische Neurowissenschaften Contact: w.weber-fahr (at) uke.uni-hamburg.de

Key Literature

• Weber-Fahr, W., Ende, G., Braus, D.F., Bachert, P., Soher, B.J., Henn, F.A., and Buchel C.: A fully automated method for tissue segmentation and CSF-correction of proton MRSI metabolites corroborates abnormal hippocampal NAA in schizophrenia. Neuroimage 16(1):49-60, 2002.

Background and state of the art
Europe currently has over 1 million patients suffering from stroke each year. The incidence rate is expected to grow due to demographic development. After initial rehabilitation, most patients with impairments of movement are released to daily life, often without regaining their original mobility. One option to improve this situation, without increasing the costs for health care, is to design and develop small, low-cost reproducible robotic systems to be deployed with the stroke patient at home. The robots under investigation must have novel sensory capabilities to adapt to the patient's individual needs. Interdisciplinary tracks of developments would aim at cheap, yet precise sensor-actuator combinations, high acceptability by ergonomic and psychologically optimised appearance, interactivity by smart sensors and signal analysis as well as therapeutically optimised movement paradigms.

Investigators
Clinical	Life Sciences	Informatics
F. BinkofskiKlinik für Neurologie Contact: ferdinand.binkofski (at) neuro.uni-luebeck.de	—	E. MaehleInstitut für Technische Informatik Contact: maehle (at) iti.uni-luebeck.de
P. TrillenbergKlinik für Neurologie Contact: peter.trillenberg (at) neuro.uni-luebeck.de

Key Literature

• Maehle, E., Brockmann, W. and Walthelm, A.: Intelligente autonome Roboter: Möglichkeiten und Grenzen. Zentralblatt für Chirurgie, 127:1-7, 2002

Background and state of the art
Haptic interfaces with realistic force/ texture feedback have potential applications in surgery and virtual training systems. To understand the mechanisms governing the human motor system would help to develop better algorithms for solving the underlying computational problems. The notion that the motor system is organised hierarchically has been proposed already by Sherrington in the first half of the 20th century. This hierarchy implies that the motor system can be viewed as a "system of systems", in which each level consists of a collection of subsystems, which are in turn composed of smaller units. In such a system, higher-level systems can modulate the activity of lower-level mechanisms. With respect to motor planning, the highest levels of the hierarchy are concerned with generating commands to achieve an action goal, while lower-level mechanisms translate the commands into a movement. Thus, action selection would involve activation of increasingly smaller elements, the activation of these elements can occur in serial as well as in parallel manner. On the other hand, skilled motor behaviours rely on accurate predictive models of both our own body and the interaction with external objects and the environment. Based on computational studies it has been suggested that the CNS internally simulates aspects of sensorimotor processing during planning, control and learning of actions. The neural correlates that perform such transformations are termed internal models as they are internal to the CNS and model aspects of the sensorimotor representation. These models are represented within the CNS on the different hierarchical levels, depending on the action context. They are also the basis of the acquisition of simple and skilled motor behaviour.

The project
The goal of the project is to transfer knowledge about the function of the motor system into haptic interface systems and stroke rehabilitation robotics. We will first consider the well established task of reaching for / grasping an object. This complex action consists of two main components, which are processed in parallel channels. The phylogenetically "older" reaching component is processed in the fast dorso-dorsal channel and the "younger" component of grasping requires more programming of the hand-object interaction and is processed in the slower ventro-dorsal channel. Neuroimaging studies with fMRI, MEG and DTI will be used to validate models. Here, fMRI will be used to localise the networks of areas engaged in the performance of haptic tasks. MEG will be used to test the temporal order of activation and effective connectivity between the areas in question. Diffusion tensor tractography (DTI) will be used to identify existing anatomical connections between the activated areas. In this way, effects of stroke rehabilitation can be verified, and robotic interfaces can be adapted.

Investigators
Clinical	Life Sciences	Informatics
F. BinkofskiKlinik für Neurologie Contact: ferdinand.binkofski (at) neuro.uni-luebeck.de	—	A. SchweikardInstitut für Robotik und Kognitive Systeme Contact: schweikard (at) rob.uni-luebeck.de

Key Literature

• Binkofski, F. and Buccino, G.: The role of ventral premotor cortex in action execution and action understanding. J. Physiol. 99:396-405, 2006.
• Pisella, L., Binkofski, F., Lasek, K., Toni, I. and Rossetti, Y.: No double-dissociation between optic ataxia and visual agnosia: Multiple sub-streams in the posterior parietal lobe for multiple visuo-motor transformation processes. Neuropsychologia 44:2734-2748, 2006.

Background and state of the art
Experienced radiologists scan images in a way that substantially differs from inexperienced viewers. It is known that patients suffering from hemineglect after a stroke show different scanning behaviours.

The project
We develop gaze-contingent interactive displays on which information can be displayed such as to guide the gaze of the viewer. By recording the gaze pattern of experts and applying it to novices who view the interactive display, we can evoke a sub-conscious learning effect. We plan to use this technology to train novices how to efficiently view medical images and improve diagnosis, and to aid children with reading disabilities to provide them with a display on which they can read with high performance. Furthermore, we will derive strategies for reducing the attentional deficits of neglect patients, and to improve the patients' exploration in the neglected field. Finally, we aim at "visual aids" such as head-mounted displays that show both the real world and guiding cues.

Investigators
Clinical	Life Sciences	Informatics
C. HelmchenKlinik für Neurologie Contact: christoph.helmchen (at) neuro.uni-luebeck.de	—	E. BarthInstitut für Neuro- und Bioinformatik Contact: barth (at) inb.uni-luebeck.de
H. KimmigKlinik für Neurologie Contact: hubert.kimmig (at) neuro.uni-luebeck.de		T. MartinetzInstitut für Neuro- und Bioinformatik Contact: martinetz (at) informatik.uni-luebeck.de
A. SprengerKlinik für Neurologie Contact: andreas.sprenger (at) neuro.uni-luebeck.de

Key Literature

• Barth, E., Dorr, M., Böhme, M., Gegenfurtner, K. and Martinetz, T.: Guiding eye movements for better communication and augmented vision. In Perception and Interactive Technologies, Lecture Notes in Artificial Intelligence, 4021:1-8, Springer, 2006
• Machner, B., Sprenger, A., Kömpf, D. and Heide, W.: Cerebellar infarction affects visual search. NeuroReport 16:1507-1511, 2005

Background and state of the art

Project
The goal of the project is to precisely locate the different pacemaker centres such as the sinoatrial node and the atrioventricular node using noninvasive techniques.

Investigators
Clinical	Life Sciences	Informatics
—	—	R. BruderInstitut für Robotik und Kognitive Systeme Contact: bruder (at) rob.uni-luebeck.de
		F. ErnstInstitut für Robotik und Kognitive Systeme Contact: ernst (at) gradschool.uni-luebeck.de
		A. SchlaeferInstitut für Robotik und Kognitive Systeme Contact: schlaefer (at) rob.uni-luebeck.de
		A. SchweikardInstitut für Robotik und Kognitive Systeme Contact: schweikard (at) rob.uni-luebeck.de

Key Literature

• Ueda, T. et al.: Visualization of Source Current Distribution in Human Heart based on Magnetocardiogram Data, 9th Digital Signal Processing Symposium, Nov. 10-11, 1994, pp. 307-312.

Background and state of the art
In humans, the formation of hippocampus-dependent declarative memories has been shown to benefit from slow wave sleep (SWS). Based on evidence from animal and human studies, the concept has been put forward that a "replay" of recently acquired materials takes place in hippocampal neuronal networks during SWS which leads to a gradual hippocampo-neocortical transfer of the representation and eventually, within weeks and months, to a preferential storage of the representations in neocortical networks independent of the hippocampus. Memory replay activity during SWS is associated with so-called sharp-wave ripple activity in the hippocampal EEG.

The Project

On this background, several study projects are planned which selectively aim at the induction of slow oscillations by tDCS during human SWS. Effects of local electrical stimulation by externally applied weak oscillating currents that are "in phase" and "out of phase" with endogenous oscillations will be compared. Techniques shall be developed that allow for separate measurement of endogenous brain activity during the period of stimulation. Eventually this research could lead to the development of brain stimulation devices that can be used to enhance sleep-dependent memory formation during neurological rehabilitation therapy.

Investigators
Clinical	Life Sciences	Informatics
J. BornInstitut für Neuroendokrinologie Contact: born (at) kfg.uni-luebeck.de	—	E. BarthInstitut für Neuro- und Bioinformatik Contact: barth (at) inb.uni-luebeck.de
L. MarshallInstitut für Neuroendokrinologie Contact: marshall (at) kfg.uni-luebeck.de		T. MartinetzInstitut für Neuro- und Bioinformatik Contact: martinetz (at) informatik.uni-luebeck.de

Key Literature

• Wagner, U., Gais, S., Haider, H., Verleger, R. and Born, J.: Sleep inspires insight. Nature 427:352-355, 2004
• Marshall, L., Molle, M., Hallschmid, M. and Born, J.: Transcranial direct current stimulation during sleep improves declarative memory. J. Neurosci. 24:9985-9992, 2004.
• Marshall, L., Helgadottir, H., Mölle, M. and Born, J.: Boosting slow oscillations during sleep potentiates memory. Nature 444:610-613, 2006.

Background and state of the art
Despite recent progress, the mechanisms involved in brain plasticity are currently not completely understood. Theoretical models and simulations could provide insights into the underlying processes. Self-organising systems have been studied independently in the area of robotics and computer networks. Recent progress in brain imaging such as Diffusion Tensor Imaging allows for mapping pathways in the brain's white matter with a resolution that would have seemed impossible only a few years ago. This progress can be applied to achieve more detailed and realistic modelling.

The Project
Our goal is to find new models for plasticity drawing from our experience in Organic Computing, self-organising neural networks and learning theory. In previous work, a general-purpose simulation tool for modelling the dynamics of highly interconnected non-linear systems has been developed. In close cooperation with sample project I3a, the experimental results gained from sample project I3a will be translated into this modelling tool.

Investigators
Clinical	Life Sciences	Informatics
J. BornInstitut für Neuroendokrinologie Contact: born (at) kfg.uni-luebeck.de	—	E. BarthInstitut für Neuro- und Bioinformatik Contact: barth (at) inb.uni-luebeck.de
L. MarshallInstitut für Neuroendokrinologie Contact: marshall (at) kfg.uni-luebeck.de		J. ClaussenInstitut für Neuro- und Bioinformatik Contact: claussen (at) inb.uni-luebeck.de
		T. MartinetzInstitut für Neuro- und Bioinformatik Contact: martinetz (at) informatik.uni-luebeck.de

Key Literature

• Ritter, H., Martinetz, T. and Schulten, K.: Neural Computation and Self-Organising Maps: An Introduction. Addison-Wesley, Massachusetts, revised and translated edition, 1992
• Stoerig, P. and Barth, E.: Phenomenal vision despite unilateral destruction of primary visual cortex. Consciousness and Cognition 10:574-587, 2001
• Mayer, J., Schuster, H.G., Claussen, J.C. and Mölle, M.: Corticothalamic projections control synchronization in locally coupled bistable thalamic oscillators. Phys. Rev. Lett. 99:068102, 2007
• Mayer, J., Schuster, H.G. and Claussen, J.C.: The role of inhibitory feedback for information processing in thalamocortical circuits. Phys. Rev. E 73:031908, 2006

Background and state of the art

Regulation of human glucose metabolism is tightly linked to cerebral energy supply which is sensed by adenosine triphosphate (ATP) -sensitive potassium channels. 31Phosphor-magnetic resonance spectroscopy (31P-MRS) allows to perform non-invasive, in vivo measurements of brain metabolites that are centrally involved in energy metabolism such as ATP or phosphocreatine (PCr), the buffer substance to maintain stable ATP levels. A decrease in PCr concentrations signifies energy consumption and thereby mirrors local neuronal activation. Mathematical models such as the well-known "minimal model" are able to describe the regulation of peripheral glucose metabolism; moreover mathematical neural systems are adequate to reflect energy consumption by neuronal activation.

The Project
The goal of this project is to investigate the close link between neuronal brain activity and metabolic responses of the human organism. The project includes:

1. Development of a mathematical model which combines brain activity and subsequent peripheral glucose response.
2. Experimental data acquisition by 31P-MRS upon neuronal activation (e.g. visually evoked potentials) or depletion (hypoxia-induction) with concomitant monitoring of the peripheral glucose metabolism (by euglycaemic glucose clamping).
3. Verification of the model via experimentally acquired data and development of a parameter identification procedure to gain specific information about the relationship between brain activity and peripheral glucose metabolism.

Investigators
Clinical	Life Sciences	Informatics
K. OltmannsKlinik für Psychiatrie und Psychotherapie Contact: oltmanns (at) medinf.mu-luebeck.de	—	M. ConradInstitut für Mathematik Contact: conrad (at) mathcs.emory.edu
		D. LangemannInstitut für Mathematik Contact: langemann (at) math.uni-luebeck.de

Key Literature

• Oltmanns, K.M., Melchert, U.H., Scholand-Engler, H.G., Howitz, M., Schultes, B., Born, J., and Peters, A.: Differential energetic response of brain vs. skeletal muscle upon hyperglycemia in humans, Exp Clin Endocrinol Diabetes 115:S64, 2007.
• Peters, A., Schweiger, U., Pellerin, L., Hubold, C., Oltmanns, K.M., Conrad, M., Schultes, B., Born, J. and Fehm, H.L.: The selfish brain: competition for energy resources, Neurosci Biobehav Rev 28:143-180, 2004.

Background and state of the art
Robotic assistance in medical applications is becoming more and more important. In most cases, the robot assisted interventions are performed under the control of navigation tools to assist the surgeon with spatial information on the location of the tool within the area of interest. Common approaches to medical navigation are based on optical systems (infrared) or electromagnetical sensors and rely on preoperative image data. The accuracy of these techniques is inherently limited. This is due to external influences (e.g. temperature), image-based inaccuracies, errors in the reference of tool tip to the respective tracking system, and morphological changes during the intervention. Current navigation technologies do not allow highly precise manipulation and do not provide information on the real local neighbourhood of the instrument.

The Project

The goal of this project is a highly precise navigation method suitable for the brain. The disadvantages of conventional navigation can be compensated by a direct integration of sensors into the tool-tip of the respective probe. Analysis of the respective sensor signals allows for identification of tissue properties in the local neighbourhood of the tool tip. The measured signals can be used for navigation and diagnosis purposes. A large research effort on campus addresses optical coherence tomography (OCT). We have placed an OCT sensor onto the needle tip of the micromanipulator. The goal of the project is to include further sensors, such as electrophysiology (e.g. microelectrodes), force measurements and biochemical engineering (micro dialysis) into the needle tip. The micromanipulator allows for steering the needle. Thus, the signals read at the needle tip, and along the needle path can be used to steer the needle, thereby navigate to a goal region in the brain. Explicit goals of the project are (a) Development of new sensor technologies (b) Integration into a minimal-invasive environment (c) Development of respective signal processing methods (d) Integration into the field of navigation and diagnosis.

Investigators
Clinical	Life Sciences	Informatics
M. BonsantoKlinik für Neurochirurgie Contact: matteo.bonsanto (at) uk-sh.de	—	Y. BaillyInstitut für Robotik und Kognitive Systeme Contact: bailly (at) gradschool.uni-luebeck.de
A. GieseKlinik der Neurochirugie Contact: alf.giese (at) med.uni-goettingen.de		R. BirngruberInstitut für Biomedizinische Optik Contact: bgb (at) bmo.uni-luebeck.de
A. MoserKlinik für Neurologie Contact: andreas.moser (at) neuro.uni-luebeck.de		U. HofmannInstitut für Signalverarbeitung und Prozessrechentechnik Contact: hofmann (at) isip.uni-luebeck.de
V. TronnierKlinik für Neurochirurgie Contact: volker.tronnier (at) uk-sh.de		A. SchweikardInstitut für Robotik und Kognitive Systeme Contact: schweikard (at) rob.uni-luebeck.de

Key Literature

• Böhringer, H.J., Lankenau, E., Rhode, V., Huettmann, G. and Giese, A.: Optical Coherence Tomography for Experimental Neuroendoscopy, Minm. Invasive Neurosurg 49:269-275, 2006.

Background and state of the art
Navigation methods for a variety of surgical applications have been studied in the past 10 years. In many cases the developments have originated in university institutes, have then reached the industry via initial tests in university clinics. The new technologies are now reaching the clinic again, in this case as commercial products. While image-guided navigation in bones is now fairly well established for bone surgery (hard-tissue), the navigation problem is by no means solved for soft tissue. Soft-tissue navigation requires elastic registration, based on a variety of complex mathematical models. A major challenge in the field of soft tissue navigation remains for the case of navigation in the brain. Here the requirements with respect to accuracy and temporal resolution are much higher than in the case of classical navigation. Instead of respiration artefacts, artefacts from pulsation, brain shift, and distortions must be addressed. In this project, methods for navigated insertions of DBS probes will be developed.

One of the most important recent innovations in the treatment of movement or psychiatric disorders is Deep Brain Stimulation (DBS), suppressing some of the most adverse side effects of neuropsychiatric diseases. In this project ways for navigated insertions of DBS probes will be developed for different target regions.

One target region will be -besides the well known Nucleus Subthalamicus for Parkinson's treatment -the Nucleus accumbens. It consists of the functionally distinct shell and a core region is highly relevant for the understanding of psychiatric disorders like depression, obsessive-compulsive disorder or substance abuse. Consequently, the Nucleus accumbens is also a promising target region for currently developed therapeutic DBS for otherwise intractable patients with obsessive-compulsive disorder or depression. For optimal placement of stimulation electrodes, the definition of regionally specific firing patterns of cells is of great importance. However, to date, only anatomical criteria have been used for therapeutic electrode placement because the electrophysiological characteristics of core and shell region have not been described in humans.

The Project
Classical navigation for brain surgery relies on two or three-dimensional image data. For DBS a macroscopic stimulating probe must be implanted into a minute area of the basal ganglia (see figures one and two), usually via a stereotaxic procedure based on pre-operative planning. Although this problem is highly complex in the light of the accuracy requirements, a number of possibilities exist, which have currently not been studied in the literature. Most prominently, electrophysiological data, acquired during the insertion have not been fused with the pure imaging data. Experiments show that distinct structures of the brain are represented by distinct neuronal activity patterns (see figure three), Moll et al 2005)). The project intends to overcome this low precision of current DBS-implantation (currently within 5mm). Online recognition of the typical spike pattern and display on the navigational computer screen should alleviate electrode placement for movement and psychiatric disorders.
In a randomised, controlled study of DBS in therapy refractory severe obsessive-compulsive disorder, electrophysiological single-cell recordings will be performed for optimal placement of the stimulation electrodes in the Nucleus accumbens. Referring to our and other's measurements from non-human primates and rodents, the acquired single-cell recordings will be used to reliably distinguish core and shell tissue in human Nucleus accumbens upon electrophysiological characteristics.

Investigators
Clinical	Life Sciences	Informatics
V. TronnierKlinik für Neurochirurgie Contact: volker.tronnier (at) uk-sh.de	—	U. HofmannInstitut für Signalverarbeitung und Prozessrechentechnik Contact: hofmann (at) isip.uni-luebeck.de
B. ZurowskiKlinik für Psychiatrie und Psychotherapie Contact: bartosz.zurowski (at) psychiatrie.uk-sh.de

Key Literature

• Krause, M., Fogel, W., Kloss, M., Rasche, D., Volkmann, J. and Tronnier, V., Pallidal stimulation for dystonia. Neurosurgery 55:1361-1370, 2004.

Background and state of the art
Aggressive surgical resection results in a prolonged survival of patients with malignant gliomas and may possibly lower the risk of anaplastic progression in low-grade glial tumours. However, the intra-operative detection of residual tumour tissue especially in low-grade tumours is difficult because of a low inherent optical contrast between tumour and adjacent brain. Furthermore, gliomas lack a true tumour to brain interface, because these tumours show invasion of adjacent brain with single invasive cells migrating along preformed anatomical structures several centimetres beyond the highly cellular margin of the tumour.
We have recently demonstrated that optical coherence tomography (OCT) of human brain in vivo can be done and that this optical tissue imaging delineates adjacent brain, tumour invaded brain and the highly cellular part of the tumour. Because optical coherence tomography can be performed as a non-contact and non-invasive imaging procedure this technology lends itself for integration into operating microscopes.

The Project
In such a scenario, a robot-assisted and navigated microscope would be able to obtain consecutive images of a tumour resection cavity. These imaging data could be scored for residual tumour. The goals of this project are: (1) Robotic microscope, (2) Integration of optical coherence tomography into the optical pathway of a neurosurgical operating microscope (3) 3D mapping data of the OCT tissue analysis and imaging data for navigation.

Investigators
Clinical	Life Sciences	Informatics
M. BonsantoKlinik für Neurochirurgie Contact: matteo.bonsanto (at) uk-sh.de	—	M. FinkeInstitut für Robotik und Kognitive Systeme Contact: finke (at) rob.uni-luebeck.de
V. TronnierKlinik für Neurochirurgie Contact: volker.tronnier (at) uk-sh.de		G. HüttmannInstitut für Biomedizinische Optik Contact: huettmann (at) bmo.uni-luebeck.de
		A. SchlaeferInstitut für Robotik und Kognitive Systeme Contact: schlaefer (at) rob.uni-luebeck.de

Key Literature

• Tronnier, V.M., Bonsanto, M.M., Staubert, A., Knauth, M., Kunze, S. and Wirtz, C.R.: Comparison of intraoperative MR Imaging and 3D-navigated ultrasonography in the detection and resection control of lesions. Neurosurg. Focus 15:10 E3, 2001.
• Böhringer, H.J., Boller, D., Leppert, J., Knopp, U., Lankenau, E., Reusche, E., Huettmann, G. and Giese, A.: Time-domain and spectral-domain optical coherence tomography in the analysis of brain tumour tissue. Laser Surg. Med. 38:588-597, 2006.

Background and state of the art
One of our main tools to communicate with the environment is the coordination of optical information and manipulation of objects in space by eye and hand movements. The differential influence of cortical areas like the prefrontal, motor, or parietal areas in cooperation with the cerebellum on the control of isolated and combined eye and hand movements in humans is difficult to examine in the functional context of ongoing movements. Using functional Magnetic Resonance Imaging (fMRI), we focus on the investigation of the central representation of eye and hand movements within the cortical network and the impact of a differential cognitive preload on the activation patterns. Our group has demonstrated maps of cerebro and cerebellar activations representing the calculation of motor and oculomotor processes (see figure one)

The Project
One of our topics will be to define the possible modulation of these cerebro-cerebellar activation patterns depending on a variable context of the movement. This will be investigated by using different levels of complex tasks, or, as in "normal" life situations, different levels of attention to the tasks especially due to different levels of external distraction.
This definition of certain cerebro-cerebellar networks allows a possibly more precise interaction with technical devices such as EEG or ECoG, for instance to control a computer or an external technical device in a more natural way.

Investigators
Clinical	Life Sciences	Informatics
F. BinkofskiKlinik für Neurologie Contact: ferdinand.binkofski (at) neuro.uni-luebeck.de	—	U. HofmannInstitut für Signalverarbeitung und Prozessrechentechnik Contact: hofmann (at) isip.uni-luebeck.de
		A. SchweikardInstitut für Robotik und Kognitive Systeme Contact: schweikard (at) rob.uni-luebeck.de

Key Literature

• Nitschke, M.F., Kleinschmidt, A., Wessel, K. and Frahm, J.: Somatotopic motor representation in the human anterior cerebellum. A high resolution functional MRI study. Brain 119:1023-1029, 1996.
• Nitschke, M.F., Binkofski, F., Buccino, G., Posse, S., Erdmann, C., Kömpf, D., Seitz, R.J. and Heide, W.: Activation of cerebellar hemispheres in spatial memorization of saccadic eye movements - an fMRI study. Hum. Brain Mapp. 22:155-164, 2004.

Background and state of the art
One in one hundred people in the western world develop epileptic seizures. In recent years, we gained new insight in the dynamics of seizure development using Wavelet Analysis, a signal processing method of decomposing EEG readings in time localised frequency fractions. We were able to diagnose a special form of adolescent epilepsy in background EEG signals.
Transcranial Magnetic Stimulation (TMS) is a method of exciting nerve tissue by electromagnetic waves. When applied to the head, parts of the brain underneath the coil are stimulated, exciting for example limp muscles or inducing visual sensations. Recently, a robotic TMS device has been developed at our university, allowing for treatment planning, navigation support and motion compensation.

The Project
The goal of the project is to extend pre-seizure diagnosis to other forms of epilepsies. Being able to diagnose epileptic disorders before first seizures occur would allow for new treatment methods. The basic idea for the project is to combine rTMS and EEG recordings. By comparing recordings prior and after stimulation and between epileptic patients and healthy subjects we aim to find markers specific for epilepsy. The main tool for analysing the data will be Wavelet Analysis. In particular, the wavelet based quantifiers Relative Wavelet Energy (RWE), Normalised Total Wavelet Entropy (WS) and Wavelet Statistical Complexity (WSC) will be applied to the EEG data. Besides its medical and mathematical challenges, the project also addresses challenging robotic, technical and signal processing research.

Investigators
Clinical	Life Sciences	Informatics
P. TrillenbergKlinik für Neurologie Contact: peter.trillenberg (at) neuro.uni-luebeck.de	—	K. KellerInstitut für Mathematik Contact: keller (at) math.uni-luebeck.de
		J. PrestinInstitut für Mathematik Contact: prestin (at) math.uni-luebeck.de
		A. SchweikardInstitut für Robotik und Kognitive Systeme Contact: schweikard (at) rob.uni-luebeck.de

Key Literature

• Rosso, O.A., Martin, M.T., Figliola, A., Keller, K. and Plastino, A.: EEG analysis using wavelet-based information tools. J. Neurosci. Methods 153:163-182, 2006.
• Matthäus, L., Giese, A., Wertheimer, D., and Schweikard, A.: Planning and analyzing robotised TMS using virtual reality. Stud: Health Technol: Inform, 119:373-378, 2006.

Background and state of the art
Over the last decade three-dimensional (3D) recordings of eye, head and limb movements have made it possible to assign specific functions to certain brain structures. In addition, coordinate systems that biological systems involved in sensori-motor control use to optimise efficiency and reduce redundancy were investigated. The oculomotor system plays a pivotal role since eye movements can be recorded and analysed better than any other motor system. By relating 3D rotation vectors of vestibularly induced eye movements to the geometry of the semicircular canals of the peripheral vestibular system or the rotation axes of individual eye muscles we identified peripheral and central vestibular lesions (Helmchen et al., 2005; Rambold and Helmchen, 2005). By using oculomotor modelling, additional lesion sites can be predicted which clinicians have not encountered so far (Glasauer et al., 2001). Vestibular oculomotor deficits are inevitably related to perceptional (perception of gravity and spatial constancy) and postural deficits and may interfere with head and hand movements. Moreover, eye, head and hand have different centres of rotation which the brain needs to incorporate into the planning of visually guided arm movements. In addition, eye movements elicited by the vestibular system are organised in 3D coordinates while rapid saccadic eye movements are organised in 2D coordinates with the head stationary. The neural implementation of this 3D-2D transformation has become an important research tool to investigate the oculomotor system in health and disease. An example to use sensorimotor (vestibulo-ocular) commands for technical devices is the recent development of a head-mounted camera system (Schneider et al., 2006) which can be used by surgeons while the oculomotor reflexes naturally stabilise the camera on target during head and target movements.

The Project
The project addresses two hypotheses:

• The central processing of convergent semicircular canal information in the human brain as assessed by head impulse tests is consistent with its prediction obtained from the alignment of 3D rotation vectors of the spontaneous nystagmus in patients with brain stem lesions and the vectors resulting from individual semicircular canal stimulation.
• Eye-centred rather than body-centred mechanisms are governing the integration of target and hand position in programming reaching movements in patients with a vestibular lesion with the head unrestrained vs. restrained.

Methods:

• Recording of 3D eye-and head movements requires a precise technique with a high resolution (0.05 deg). In recent years we have used the "scleral search coil technique" which uses contact lenses containing small coils (figure one) which induce a current while the subject is sitting in a frequency modulated magnetic field. Similar coils can be used to record head movements while arm movements will be recorded with an ultrasound-based system (Zebris).
• Small amplitude, high accelerating head impulses (8000-10000/s2) are passively performed in the appropriate plane of the semicircular canals (figure one). The vestibulo-ocular reflex (VOR) elicits 3D eye movements in the spatial coordinates of the stimulated planes. This exceptional technique allows - as the only available method - to identify individual semicircular canal (dys)function in the peripheral and central nervous system. Figure one illustrates decreased eye vs. head velocity when the anterior and horizontal canal is stimulated.
• Patients with peripheral and central vestibular disorders will be investigated for VOR deficits in the geometric planes of the semicircular canals. 3D rotation vectors of pathological spontaneous eye movements, which are caused by the vestibular innervational imbalance will be analysed and related to eye muscle rotation vectors known from individual semicircular canal stimulation.
• Hand pointing movements will be recorded with the ZEBRIS system (Trillenberg et al., 2006) and patients with peripheral vestibular lesions will be tested in four conditions: a) without visual information about hand position (unseen-hand condition) vs. b) when hand and target position is simultaneously visible before movement onset (seen-hand condition) (n=8 for each condition), c) with the head fixed vs. d) unrestrained.
• Perceptual deficits will be assessed by the subjective visual vertical and postural imbalance by posturography.

Investigators
Clinical	Life Sciences	Informatics
C. HelmchenKlinik für Neurologie Contact: christoph.helmchen (at) neuro.uni-luebeck.de	—	U. HofmannInstitut für Signalverarbeitung und Prozessrechentechnik Contact: hofmann (at) isip.uni-luebeck.de
P. TrillenbergKlinik für Neurologie Contact: peter.trillenberg (at) neuro.uni-luebeck.de

Key Literature

• Rambold, H. and Helmchen, C.: Spontaneous nystagmus in dorsolateral medullary infarction indicates vestibular semicircular canal imbalance. J. Neurol. Neurosurg. Psychiatry 76:88-94, 2005.
• Trillenberg, P., Fuhrer, J., Sprenger, A., Hagenow, A., Kömpf, D., Wenzelburger, R., Deuschl, G., Heide, W. and Helmchen, C.: Eye-hand coordination in essential tremor. Mov. Disord, 21:373-379, 2006.
• Burmeister, O., Litza, M. and Hofmann, U.G.: Synchronous stereo-video and biosignal recording - a basic setup for Human-Computer-Interface applications , IEEE Proceedings of the 2nd Int'l Conference on Neural Engineering, 501-505, 2005.

Background and state of the art
Recent progress in analysis of brain signals from surface scalp electrodes (EEG), cortical surface electrodes (ECoG) and intra-cortical depth electrodes (DBS) has rendered the brain accessible as source of signals, which can be used to control robotic actuators. This field is generally called Human-Computer-Interfacing, capitalising either on slow macroscopic potentials or on fast neuronal population code.

The goal of this project is to develop stationary medical grade signal processing hard- and software to record from 256 wideband channels while ignoring synchronous magnetic disturbances (TMS). This system is meant to record intra-operatively from surface scalp electrodes (EEG), cortical surface electrodes (ECoG) and intra-cortical depth electrodes (DBS) during scheduled surgical procedures (tumour resection and epileptic focus resection) The second step will deal with the improvement of the stationary system to a mobile platform, based on embedded signal processing. The system will then be set up for real-time use in the analysis of brain signals. The goal is to develop optimal classifiers to interpret brain activity and utilise this analysis to provide a scarce command set for output devices. Methods to be used will include advanced types of Blind-Source Separation and Spike detection based on wavelet analysis

Investigators
Clinical	Life Sciences	Informatics
A. MoserKlinik für Neurologie Contact: andreas.moser (at) neuro.uni-luebeck.de	—	U. HofmannInstitut für Signalverarbeitung und Prozessrechentechnik Contact: hofmann (at) isip.uni-luebeck.de
V. TronnierKlinik für Neurochirurgie Contact: volker.tronnier (at) uk-sh.de		A. MertinsInstitut für Signalverarbeitung und Prozessrechentechnik Contact: mertins (at) isip.uni-luebeck.de

Key Literature

• Folkers, A., Mösch, F., Malina, T. and Hofmann, U.: Realtime bioelectrical data acquisition and processing from 128 channels utilising the Wavelet-Transformation, Neurocomputing 52-54:247-254, Elsevier, 2003.

Background and state of the art
Nonlinear EEG analysis techniques play an increasing role in clinical research and practice. In particular, methods based on information theory and nonlinear dynamics have been introduced into EEG analysis. Recent developments provide new insights into the organisation of bioelectrical activity in the brain, e.g. in relation to epileptic activity and to cognitive processes. One new promising approach to the analysis of large and high dimensional EEG data is ordinal time series analysis which is based on investigating the distribution of order patterns in a time series for different time scales, instead of focusing on the distribution of rhythms. This approach results in very fast and flexible algorithms and allows for robustness towards added noise and monotonic scale changes.

The Project

The goal of this project is to develop algorithms for identifying, classifying and discriminating long and high dimensional EEG time series and parts of them and the related brain states, mainly on the base of ordinal time series analysis, but also by combining ordinal time series analysis with other methods as Fourier and Wavelet analysis. We focus to the analysis of epilepsy and sleep data, in particularly to the automatic detection of different stages of epileptic activity and sleep stages, respectively. Special emphasis is put on extending ordinal time series analysis from one-channel to multichannel analysis, in order to be able to quantify similarity and coupling between the components of a multi-channel EEG. One of the central aspects is a generalisation of the permutation entropy which is defined as the Shannon entropy of the ordinal pattern distribution obtained from a time series and is used for quantifying the complexity of the time series and the system behind. On the one hand we use different entropies, e.g. Renyi entropies and Tsallis entropies in order to have more flexibility in distinguishing ordinal pattern distributions and on the other hand we generalise the concept of permutation entropy to higher dimensions. The use of different entropies has been successful in the context of wavelet entropies in the investigation of special epilepsy types. One aim is a better understanding of the relationship of wavelet entropies and (generalised) permutation entropies.
Ordinal pattern distributions are not completely characterised by their entropies, requiring the consideration of other quantities, such as related to the mean length of monotone parts and the frequency of monotonic changes in a time series with interesting relationships to the Hurst parameter in the case of self-similar time series. We consider ordinal similarity and coupling measures. Such quantities and the permutation entropies obtained from a sliding window analysis are considered for a model based time series analysis.
Our recent work demonstrates that Cluster Analysis of distributions of ordinal patterns in EEG time series is a very useful and flexible method for analyzing and visualising the structure of long and high dimensional EEG data set and detecting different brain stages. One of the advantages of this method is that it allows for a simultaneous inspection of large collections of EEG data sets.

Investigators
Clinical	Life Sciences	Informatics
A. SprengerKlinik für Neurologie Contact: andreas.sprenger (at) neuro.uni-luebeck.de	—	K. KellerInstitut für Mathematik Contact: keller (at) math.uni-luebeck.de
A. VerlegerKlinik für Neurologie Contact: rolf.verleger (at) neuro.uni-luebeck.de		A. MertinsInstitut für Signalverarbeitung und Prozessrechentechnik Contact: mertins (at) isip.uni-luebeck.de
		J. PrestinInstitut für Mathematik Contact: prestin (at) math.uni-luebeck.de

Key Literature

• Keller, K. and Lauffer, H.: Symbolic analysis of high-dimensional time series. Int. J. Bif. Chaos 13:2657-2668, 2003.
• Bandt, C., and Pompe, P.: Permutation entropy: A natural complexity measure for time series, Phys. Rev. Lett. 88:174102, 2002.

Background and state of the art
Deep brain stimulation (DBS) by means of implanted electrodes has become a well accepted technique for treatment of several movement disorders. The mechanisms of the molecular and pharmacological actions of electrical high frequency stimulation (HFS) with 130 Hz are, however, still unknown. The effectiveness of HFS has been described by various mechanisms, ranging from blockage of depolarisation to stimulation-evoked release of Γ-aminobutyric acid (GABA) or stimulation of inhibitory receptors in the nuclei.

The Project
Data of our in vitro studies provide initial evidence that HFS not only requires GABAA-receptors but also intact GABAergic nerve terminals to achieve an inhibitory effect. These results also indicate that HFS has a specific effect on GABAergic neuronal terminals resulting in an enhancement of extracellular GABA. Therefore, we developed an in vivo model that allows for simultaneous and collocated micro dialysis and HFS in the caudate nucleus of alert and active rats, which are stimulated by electrical pulses of 124 Hz, to analyse neurotransmitters under HFS (figure one).

Although for DBS, the GPi and STN are the most important structures, these two regions are very small in rats and therefore not suitable for our in vivo experiments. The large-sized caudate nucleus was employed due to the high density of striatal GABAergic neurons, but not because its linkage to Parkinson's disease. To achieve high precision in probe placement (e.g. in the STN), a prototype of a robotic stereotactic manipulator for small animals has been developed. For guide cannula implantation, Wistar rats were pre-anesthetised with CO2 followed by full anaesthesia with sodium pentobarbital. Using standard stereotaxic techniques, an intracerebral double guide cannula was placed just above the right caput nuclei caudati and fixed to the skull. The double guide cannula consisted of a micro dialysis cannula (CMA/11) to which a second cannula was glued by two component epoxy resin under microscopic and micromanipulator control. The second cannula served as guide for a concentric bipolar stimulation electrode with 250 µm outer diameter and 30 mm total length.
GABA and glutamate outflow were sampled by micro dialysis technique and quantified after pre-column o-phthaldialdehyde sulphite derivatisation using HPLC with electrochemical detection. Our preliminary experiments allow for quantitative neurotransmitter determination of GABA and glutamate in an in vivo animal model with simultaneous micro dialysis and HFS. We could demonstrate that high frequency stimulation significantly increased basal GABA outflow without affecting glutamate levels in freely moving rats (figure two).

The aims of this project are: (1) To analyze the GABAergic neuronal target system for electrical high frequency stimulation (HFS) in our in vivo animal model. (2) To develop stimulation on demand controlled by a GABA input impulse to optimise DBS. Here, signals delivered from GABAergic neurons should be employed as endogenous trigger. (3) To optimise micro dialysis probe and stimulation electrode positioning (e.g. in the STN or Gpi), we develop an automatic stereotactic manipulator integrating Optical Coherence Tomography (OCT) as an imaging modality for in situ control, see Figure 14 above).

Investigators
Clinical	Life Sciences	Informatics
A. MoserKlinik für Neurologie Contact: andreas.moser (at) neuro.uni-luebeck.de	—	U. HofmannInstitut für Signalverarbeitung und Prozessrechentechnik Contact: hofmann (at) isip.uni-luebeck.de
		G. HüttmannInstitut für Biomedizinische Optik Contact: huettmann (at) bmo.uni-luebeck.de

Key Literature

• Hiller, A., Loeffler, S., Haupt, Ch., Litza, M., Hofmann, U. and Moser, A.: Electrical High Frequency Stimulation Induces GABA Outflow in Freely Moving Rats. J. Neurosci. Meth. 159:286-290, 2007.
• Li, T.L., Thümen, A. and Moser, A.: Modulation of a neuronal network by electrical high frequency stimulation in striatal slices of the rat in vitro. Neurochem. Int. 48:83-86, 2006.
• Gritsun, T., Litza, M., Hiller, A., Moser, A. and Hofmann, U.G.: Semichronic, collocated deep brain stimulation and multisite recording in rats, ID WP2, Int'l Conf on Microtechnologies in Medicine and Biology. Bankoku-Shinryokan, Okinawa, Japan, 2006.

Background and state of the art
It has been a long way from the first evidence of electro-stimulated auditory perception to state-of-the-art cochlear implants. However, even with the best-known systems, the cochlear ganglion-cells are still the primary targets for multichannel stimulation and therefore prone to degradation and long-term decay. Consequently, it is desirable to get access to an alternative stimulation paradigm to elicit auditory perception.

The Project

The primary goal of this project is to develop chronically implantable, long-term stimulation electrodes for a brain-stem auditory prosthesis. This goal may be achieved by mechanically compliant, flexible multisite microelectrode arrays, which are coated with tissue-engineered and pre-seeded bio reactive cells. These cells are locked to micro cages in the probe and will form a highly specific synaptic contact to the surrounding brain tissue, thus not only fixing the probes, but fostering synaptic connections over time. For this purpose flexible probes from PDMS or polyimide with micro wells are to be colonised with tissue engineered and optimised intermediator cells (neurons) and implanted in an in vivo model of the brain stem.

Investigators
Clinical	Life Sciences	Informatics
B. WollenbergKlinik für Hals-, Nasen- und Ohrenheilkunde Contact: Barbara.Wollenberg (at) uk-sh.de	C. KruseInstitut für Medizinische Molekularbiologie Contact: charli.kruse (at) emb.fraunhofer.de	U. HofmannInstitut für Signalverarbeitung und Prozessrechentechnik Contact: hofmann (at) isip.uni-luebeck.de

Key Literature

• Hartmann E., Graefe H., Hopert A., Pries R., Rothenfusser S., Poeck H., Mack B., Endres S., Hartmann G. and Wollenberg B.: Analysis of plasmacytoid and myeloid dendritic cells in nasal epithelium. Clinical and Vaccine Immunology 2006, 1278-1286.
• Hofmann, U.G. and M. Es-Souni, Bi-Stabile Mikrodraht-Matrix für die neuronale Ableitung. Deutsches Patent, 2003 (DE 10355815).