Research program
TEBAM: study, development and physiological-clinical validation of a multi-modal methodology for the 3D "True Electrical Brain Activity Mapping" in normal and pathological subjects
University Co-ordinator
Università di PISA -
ONCOLOGIA,DEI TRAPIANTI E DELLE NUOVE TECNOLOGIE IN MEDICINA - PISA(PI)
Research Unit Leader
Carlo BARTOLOZZI
Description
The Research Unit of Pisa will collaborate with the groups of Trieste for the test and optimisation of three-dimensional anatomical and functional models based on the MR methods that are to be integrated with three-dimensional neurophysiological data-sets, specifically with EEG data-sets.First, the appropriate technologies of the diffusion tensor imaging and fiber tracking will be implemented and new solutions will be tested for the acquisition of MR data (high angular resolution tensor, high spatial resolution tensor, anisotropic regional sampling of tensorial data), which will make it possible to build high resolution connectivistic models for the integration of other anatomical and functional data. Specifically in a first step, connectivistic tensorial methods will be avaialable from other research units for the reconstruction of three-dimensional neurophysiological models, and, in a second step, anatomical and functional volumetric models will be used as constraint for the elettrophysiological functional localizations.Alternative MR methods will be, then, tested for the integration of presently used methods for functional imaging, in order to improve the accuracy of the functional localization, avoiding the problems related to venous artifacts, and to make the method more robust with respect to the methodological errors related to the incompleteness of BOLD model.To these purposes suitable experimental paradigms will be designed that will able to maximally use the additional information that the new fMRI methods are able to gather and new physiological models will be studied to describe the cerebral metabolism during activation.An accessory function of the Research Unit of Pisa, will consist in the analysis and solution of the specific problems concerning the anatomical high resolution data processing and integration of tensorial data on to the connectivistic strucutre of the white matter, and the functional data processing obtained by innovative fMRI methods, on the basis of the several years' experience of this Unit in the field of functional data processing (15-19).While the other Reseach Unit in the coordinated project are improving the neurophysiological methods and optimising the integration of these methods with the anatomical and functional MR data, this Unit will contribute to the planning and realization of the specific activities of the other units on one side, and will produce three-dimensional multimodal anatomo-functional and connectivistic models and will quantitatively assess the potentials of the non-BOLD functional methods in terms of sensitivity, specificity and spatial resolution capabilities in comparison to analogous experiments carried on with conventional fMRI methods on the other side.It is to be expected that the problems concerning the optimisation of acquisition strategies for the time-series, the practical application of these techniques for the functional experiments, and the data processing of the collected data-sets will vary significantly across the different methods.For what concerns the perfusion-based fMRI, the method of choice of the current scientific literature is the Arterial Spin Labeling (ASL). The ASL methods are certainly to be preferred among the other methods for perfusion studies that prescribe the injection of a para- or super-paramagnetic contrast media, for the superior feasibility of execution of the protocols, the absence of biological invasiveness and the possibility of repeating many times the stimulation paradigm. Among the main problems concerning the ASL techiniques, it is notable the fairly modest signal-to-noise ratio, the limited spatial resolution and the quite long acquisition times.This Research Unit will then be responsive of analysing the operation modalities that will gather the best spatial and temporal resolutions, and contrast-to-noise ratios of the functional images obtained through the ASL sequences. In a second step, the best strategies of processing will be investigated for the acquired functional information in order to adapt the analysis tools to the data collected with the perfusion sequences. The strategy that this Unit will propose consists mainly in the use of methods of statistical analysis of "non-inferential" type and specifically the use of the decomposition of data-sets into Independent Components (Independent Component Analysis, ICA). This technique is able to analyse the spatio-temporal changes in the fMRI time-series without using any prior information and by decomposing them in a number of spatio-temporal components characterised by a the maximum possible mutual statistical independence. After a subsequent analysis step of the produced spatio-temporal modes of activity, also using methods for the automatic modal ranking (rank-ordering) on the basis of the statistical parameters of each singe mode (RMS value, Kurtosis, auto-correlation), specific spatio-temporal components will be represented by activation maps and associated time-courses. While the maps will display the spatial distribution of the activation phenomenon in the anatomical space, the associated time-courses will gather information about the operation in the time domain of the region that active in the corresponding map: practically from the time-courses it is possible to extract a rough model of cerebral activity as it is observed in the sequence of activation. In this sense the "non-inferential" methods for data analysis are also defined as "model generators" and their use is perfectly suitable for the application of fMRI acquisition methods that aim to detect "new" phenomena for which a detailed or simplified model is not yet available. In a subsequent phase of the work, this Research Unit will evaluate analytically the activation maps built by means of the ASL sequences, in order to show similarities and differences with respect of the more common BOLD maps and to study the relative advantages and the possibility of fusing BOLD and perfusion information.It is to be expected that the accurate comparative analysis of perfusion and BOLD data can produce tremendously important information about the regional variations in the cerebral neurovascular physiology.The practical implementation of the methods for perfusion studies will be made more feasible by the availability of ASL sequences for the cerebral perfusion in the MR equipment of this Research Unit. The already existing collaboration with other research groups particularly specialised in the design of MR acquisition sequences (University of Basel, Prof. Klaus Scheffler) for equipments that are compatible with those of this Unit, will guarantee the necessary technological knowledge.For what concerns the application of the spectroscopy, the operational difficulties consist mainly in the scarce temporal resolution of the method. In order to make the spectroscopy compatible with the current normal protocols for functional imaging, suitable methodological solution will be evaluated that are able to gain temporal resolution through the use of "time-locked" spectroscopic acquisition that will be synchronised to the stimulation, and will be able to decompose the acquisition in several subsequent phases, each of which sample on a specified period after the stimulation (14). In the field of Spectroscopy the effort of this Unit will consist in the processing of spectroscopic data in order to optimise the information content of the signal, balancing the low signal-to-noise ratio induced by the executive conditions of the experiments of functional spectroscopy. Also for functional spectroscopy, the already existing collaboration with the Department of Physics of the University of Rome (Prof. Maraviglia) on these experimental issues guarantees the experimental basis that will favour significantly the practical implementation of the methods.The study of the functional diffusion is probably the functional non-BOLD method that will require the maximum effort in the organization and planning of the experiments. The poor literature existing about the argument (4) requires the need of a number of pilot experiments in order to choose optimally the characteristic of the sequences that can produce the strangest function effects. Among these targets of the Unit, the precise and systematic study of the fine characteristics of the phenomena of functional diffusion will be of extreme importance, since these phenomena are presently not know for the major part. Again, as for the other functional non-BOLD methods, it will be necessary a preliminary phase for the generation of a model for the neuronal activation according to how this phenomenon is mirrored by the functional diffusion. To this purpose, the non-inferential methods for the decomposition of the acquired time-series will be very much useful in order to detect functional phenomena of which little is known at present. The possibility of using specialised acquisition sequences for the functional diffusion designed by foreign academic Institute within a current collaboration with this unit is of extreme usefulness. In fact an acquisition sequence for function diffusion studies designed at the University of Basel for the use in the MR equipment in Pisa.For what concerns the studies on the diffusion tensor, this Unit will make use of the collaboration with other Research Groups that have important experience in the use of the method. Particularly, the collaboration with the University of Boston (Prof. Dae-Shik Kim, Prof. Susumo Mori) and of Maastricht (Prof. Rainer Goebel, Prof. Elia Formisano) will help the implementation of the method on the equipment in Pisa, and will optimise the specific software for the visualization of the information relative to the diffusion tensor.Experiments on the MR-based detection of neuronal currents will be performed as well. Also for this project an already existing experimental work carried on by the Unit of Pisa in collaboration with the Unit of Basel, has already developed specific knowledge about innovative specific sequences for the study of the neuronal electrical phenomena.Finally, the most favourable conditions for a "constructive" integration of the potentials of the single methods will be determined by adapting the contribution of each method to the diverse experimental conditions and specific features of the regional anatomy of the brain.
|
