Integration Between Different Functional Techniques
New prospects include the integration of different functional techniques to obtain more complex information about the working of cerebral structures. A major challenge for neuroscience is to understand brain function in terms of connectional anatomy and the dynamic flow of information across neuronal networks.
In white matter, water diffusion is highly directional (anisotropic), with preferential diffusion along the long axis of fibre tracts. With application of large magnetic field gradients during image acquisition, MR images can be sensitized to the diffusion water molecules within the voxel, and from these images we can compute the local direction of greatest diffusion. With these principle diffusion directions, the organization of major fibre tracts can be mapped. Furthermore, a major limitation of this method is that it does not distinguish between efferent and afferent projection.
A complementary approach is concerned with establishing the ways in which information is transmitted and integrated across brain networks. These are dynamic, context-dependent processes, in which variations in task demands lead to the preferential recruitment of some networks over others. Methods for analysis of these processes are based on the premise that functionally interacting regions will show correlated patterns of activity . The advantage of functional neuroimaging methods is that they can be used to detect activity not just in a limited set of areas but across the entire brain simultaneously. This makes it possible to examine the statistical relationship between the activities of several areas across the brain. Ramnani et al.  describe new strategies for use of functional MRI data in the analysis of functional connectivity in the human brain. They use diffusion tractography and functional mapping to highlight the possibility that future strategies for understanding interactions between regions of human brain will benefit from integrating anatomically informed models of functional interactions.
The brain is uniquely suited for functional MR imaging with the potential for extensive clinical application and benefit for the patient. Although neurological applications for functional MRI, particularly blood oxygen level dependent (BOLD) imaging, have been in experimental use since the early 1990s, its widespread clinical application is a relatively recent phenomenon.
Recent advances in functional imaging have allowed fMRI to be applied to a broad range of clinical disease processes. The combination of conventional BOLD Imaging with Diffusion Tensor Imaging and other physiologic techniques such as drug challenges holds great promise for understanding neuronal processes and the development of clinically meaningful diagnostic tests. Clinical applications of fMRI have largely relied on block style BOLD fMRI techniques to answer relatively simple questions regarding motor and language mapping in patients. These techniques have been applied most broadly in preoperative evaluations in the setting of brain tumours and epilepsy. More recent techniques such as event related paradigms and combinations of event and block style paradigms have allowed for better study of more sophisticated cognitive processes. These paradigms not only produce more complete assessments of cognitive processes presurgically, but they also open up the potential for non-invasive clinical evaluation of cognition for neurodegenerative processes such as multiple sclerosis, Parkinson's disease, and Alzheimer's disease [55, 63]. fMRI has also been recently applied to the study of the efficacy of therapy for disease processes using drug challenges during fMRI and with fMRI performed during deep brain stimulation.
The combination of fMRI with DTI provides the most comprehensive presurgical mapping available, clearly delineating the relationship of tumours or other pathology to both eloquent cortex and critical white matter pathways. The application of these combined techniques to multiple sclerosis suggests that a new pathway/network driven approach to the assessment of neurological diseases is now possible . Functional MRI has great potential both for mapping brain function and as part of a more comprehensive assessment of neural pathways and networks.
In different preliminary reports we can see the possibility of exploring the connectivity between cerebral regions. Kim et al.  use DTI fibre tracking to follow separately the connectivity of central and peripheral fields in the human visual system. Central and peripheral fibres showed different patterns of connectivity with higher visual areas. Areas showing category-specific fMRI responses showed higher connectivity with regions representing central visual field. Upadhyay et al.  used DTI fibre tracking and fMRI to understand the functional connectivity that underlies low level auditory stimuli processing in healthy subjects. They found that haemodynamic responses vary in distinct primary auditory cortex regions upon presentation of specific frequency sounds and that tonotopic organization can be revealed using fMRI. Activation was predominantly present along Heschel's gyrus and the sylvi-an fissure. Using the activation maps, grey matter ROI seeding points were created in DTI data sets to characterize axonal projections within the auditory cortex.
Schonberg et al.  used fMRI and DTI in a patient with brain lesions. The brain lesions, especially space occupying lesions, often involve the white matter and alter the known anatomical path in which the fibres pass. Nonetheless in many of these cases only a partial or functional deficit is observed, leading to the assumption that the fibres are still partially functionally intact even if deviated. In such cases, white matter mapping using seed ROI based on known normal anatomical locations might be misleading. In their work they used fMRI driven seed ROI, choosing a procedure in patients with space occupying lesions where probable deviation of white matter tracts was observed. They used analysis of the principal diffusivities to characterize the displaced white matter. Functional MRI activations were used as landmarks to choose seed regions ofinter-est for fibre tracking. Fibre tracking was used to mark areas of displaced white matter and from which DTI measures were extracted. DTI measures of the displaced fibre tracts revealed several significant changes compared to the opposite healthy hemisphere within each patient. Detailed analysis of the DTI measures showed that the displaced fibres were characterized by increased FA, decreased radial diffusivity and increased parallel diffusivity. Lowe et al.  demonstrated that the diffusivity of water transverse to the direction of the fibre was reduced in the interhemispheric pathways connecting bilateral motor regions in patients with multiple sclerosis, when compared to a control population. In this study they demonstrated that fibre tracts can be identified and isolated using DTI and fibre tracking algorithms. They adopted a method to compare the diffusion tensor in a subset of axons connecting bilateral motor regions of the brain. This was done using fMRI methods to identify cortical motor regions and DTI-based fibre tracking to select the axonal regions connecting the motor regions. The transverse diffusion was measured specifically along the tracts traversing the corpus callosum and connecting the bilateral motor regions.
One other way to use diffusion gradient is by using spectroscopy (DW-MRS). The diffusion gradient, in fact, produces a decrease in the signal intensity, which is more marked for the molecules diffusing faster. In DW-MRS, diffusion gradients are used to measure the diffusion coefficient of different metabolites. DW-MRS gives information on the size and possibly on the shape of the compartment in which the metabolite is contained and on the shape and size of cells .
In another study, Upadhyay et al.  used DTI combined with spectroscopy. They considered that in diffusion tensor imaging the degree of directed movement or anisotropic diffusion of water molecules is exploited to characterize the local neuronal microstructure and white matter fibre projections. One confounding issue of the DTI technique arises from the fact that there are intracellular and extracellular components, which contribute to the overall diffusivity, thus leading to an ambiguity in determining the underlying cause of diffusion properties in a region of interest. Of particular interest is the contribution of the intracellular and the extracellular pools to the fractional anisotropy. An-isotropic diffusion is believed to stem from restrictions imposed on diffusion bybarriers such as cell walls, and thus the FA is intimately connected with the tissue structure. The N-acetylaspartate (NAA), a solely intra-cellular constituent, has diffusion properties and it is a possible alternative for probing neuronal structure properties. Upadhyay et al.  combined DTI and NMR spectroscopy to characterize the diffusion of NAA and water molecules in a normal subject. Diffusion spectroscopy data was obtained from a voxel positioned in the anterior region of the corpus callosum. The fractional anisotropy value obtained for NAA was significantly larger than the fractional anisotropy value obtained for water in the same VOI. In their opinion this could indicate that intracellular space contributes most significantly to anisotropic diffusion in neural tissue. In a study by Tang et al. , they found that in patients with schizophrenia NAA and DTI anisotropy indices were significantly correlated. NAA was significantly reduced in the medial temporal regions and DTI-anisotropy indices were also reduced in the same regions. This implied that there was a white matter abnormality in patients with schizophrenia and the biochemical abnormality as detected in MRS was consistent with the DTI results.
Therefore DTI-MRS may become a promising tool in the investigation of tissue structure through intra-cellular diffusion properties and biochemical abnormalities.
Another possibility of multimodality imaging is the combination of high-resolution PA (phase-array coils) imaging at 3 T with the spatial localized neural activity at high temporal resolution provided by magnetoen-cephalography (MEG) and information on abnormal activity provided by the simultaneous EEG recording. Grant  describes the increased ability to detect lesions, define their extent and determine their character by phase-array coils and 3 T imaging in patients with focal epilepsy. She uses diffusion tensor imaging to study changes in white matter organization that can play a role in epileptogenesis and seizure propagation. The author also combines these structural imaging advances with coregistered neurophysiological information obtained with magnetoencephalography and with simultaneous EEG recordings. Magnetoencephalogra-phy (MEG) is an emerging methodology for detection of extracranial magnetic fields associated with the electrical current of neuronal activity. Mathematical modelling of the source of such activity, combined with anatomical registration to MR imaging, allows the forma tion of magnetic source images (MSI or functional maps). Since MEG has a submillisecond temporal resolution, these maps have the potential to display the spatiotemporal dynamics of neuronal activity and signal propagation, as well as revealing temporal signatures in evoked responses. To date, the principal clinical success of magnetic source imaging has been the identification of somatosensory cortex in patients scheduled for neurosurgical resection of mass lesions . A number of centres routinely incorporate MSI data into the neuro-navigational systems used to guide operative procedures, such that MSI is used not only for presurgical planning, but also for intraoperative guidance.
Developments in cellular and molecular biology are extending the horizon of medical imaging from gross anatomic description towards delineation of cellular and biochemical signalling processes. The emerging fields of cellular and molecular imaging aim to non-invasive-ly diagnose disease based on the in vivo detection and characterization of complex pathologic processes, such as induction of inflammation or angiogenesis. Molecular imaging can brieflybe defined as the remote sensing of cellular and molecular processes in vivo. Although molecular imaging is a biology-driven enterprise (it is the underlying biological question rather than the modality that determines which methods are used), it is currently bound by the existing techniques of optical, radionuclear, MR, and ultrasonographic imaging. These techniques are complementary in the sense that they measure processes with different sensitivities and resolutions. Each is beginning to find a niche in molecular imaging research. Each is capable of truly molecular detection in vivo, although only the radionuclear techniques and MR spectroscopy are currently applied in humans. MRI is a particularly advantageous modality for molecular and cellular imaging given its high spatial resolution and the opportunity to extract both anatomic and physiologic information simultaneously [26-76]. In molecular imaging the intrinsic contrast can be augmented by the use of targeted contrast agents in both the experimental and clinical setting. The evolving field of molecular imaging requires the development of a novel class of MR-detectable agents that are better able to provide image contrast to target specific disease processes. Strategies for generating targeted contrast can be broadly categorized  as: (1) where the CA (contrast agent) macromolecule is directed to a specific receptor using a high-affinity ligand such as a monoclonal antibody (MAb), (2) where the number of MR reporter molecules increases with enzymatic activity, resulting in signal amplification, and (3) where the presence of a targeted probe is detected through a de crease in the bulk water signal due to chemical exchange between protons of bulk water and those of the probe . The high detection sensitivity of the latter is due to numerous water molecules interacting with a single molecule of the probe .
Currently, two major classes of contrast agents exist: paramagnetic (gadolinium based) and superparamag-netic agents . The paramagnetic agents shorten both T1 and T2 relaxation, but preferentially T1. They create a hyperintense contrast on conventional T1-weighted spin-echo images. The second class of agents is based on ultrasmall superparamagnetic iron oxide (SPIO) particles. In view of the small size of these particles, the magnetic moments are unhindered by lattice orientation. In a magnetic field, the net magnetic moment is several orders of magnitude greater than that with the paramagnetic agents. This creates extremely large microscopic field gradients for dephasing nearby protons  and results in a dramatic shortening of the relaxation properties of the tissue. Recently, contrast probes have been designed to potentially demonstrate changes in gene expression or other cellular processes . The imaging protocols are critically important for SPIO detection by MRI. The susceptibility artefacts created by accumulation of SPIO are most sensitively imaged T2*-weighted imaging sequences. T1-weighted and proton density imaging with sequences are far less sensitive to the susceptibility artefacts induced by SPIO uptake in tissue . The T2*-weighted sequences typically utilize gradient echo techniques with long echo times. The long echo time accentuates signal loss due to the presence of SPIO, but these sequences also tend to suffer from a low signal-to-noise ratio (SNR). Often, highly specialized coils, such as phased array coils or application-specific surface coils, are employed in order to maximize the available SNR . In addition to the imaging protocol itself, the choice of imaging time post SPIO injection is critically important. The long circulating half-life of SPIO nanoparticles is necessary to achieve adequate loading into inflammatory cells, but it can also interfere with obtaining high quality images. Up to 24 h after SPIO administration, the blood concentration is high enough to create image artefacts. On the other hand, too long of a delay (72 h) after SPIO injection can result in no detectable susceptibility artefacts .
An alternative approach is to use polymerized vesicles conjugated to monoclonal antibodies for the highly accessible targets in endothelial receptors . These agents consist of a derivitized gadolinium diethylene-tri-aminepentaacetic acid (DTPA) polymerizable lipid, a biotinylated lipid for antibody conjugation, and a dia-cetylene phosphatidylcholine filler lipid. This approach has been successful in two areas: targeting integrin upregulation on the endothelium in tumours by using MR imaging to help define regions of abnormal angio-
genesis and in an animal model of experimental allergic encephalomyelitis [64, 65]. By incorporating avast quantity of paramagnetic complex onto each particle, the signal enhancement possible for each binding site is magnified dramatically. The increased paramagnetic influence arises from two mechanisms: the relaxivity per particle increases linearly with respect to the number of gadolinium complexes and the relaxivity of each gadolinium increases due to the slower thumbing of the molecule when attached to the much larger particle. On studying dilutions of nanoparticles in water, T1 relaxi-vity increased with increasing gadolinium payloads . The chemistry employed to bind the targeted li-gand to the particle surface can also dramatically affect the efficacy of the final contrast agent. In some cases, the active binding site of the ligand may become occupied or obscured after attachment to the nanoparticle. Obviously, such agents would yield poor molecular imaging results. In addition, the incorporation of flexible polymer spacers, i.e. polythylene glycol, between the targeting ligand and the nanoparticle surface may improve targeting efficiency. These flexible „tethers" permit a wider range of motion for the targeting ligand, potentially increasing the likelihood of encountering and binding to the target of interest . In addition to the number of gadolinium ions per particle, the number of targeting ligands per particle must also be optimized. A nanoparticle with numerous bindings ligands tends to provide a more efficacious contrast agent. In contradistinction to SPIO agents, which are designed to increase the T2* relaxation of targeted tissue, paramagnetic compounds are typically used to increase T1 relaxation. The T2* effect employed with SPIO is usually visualized as decreased image intensity with T2*-weighted MRI. Paramagnetic agents will produce increased tissue signal when interrogated with T1-weighted MRI. Therefore, the contrast effects of SPIO and paramagnetic agents are very different, and the MRI pulse sequences and parameters are also distinct for the two methods. On the T1-weighted image, the contrast between two tissues with different T1 relaxation times can be maximized by imaging at the optimum value of TR. Altering the TR by 300 - 400 ms away from the optimum value can reduce the SNR by up to 25% .
An alternative method to define inflammatory events after stroke is to tag cells with MR-detectable agents . The avidity in which ultrasmall SPIOs are phagocytosed by cells of the mononuclear phagocytic system has been used to study the role of macrophage infiltration into the brain in neurodegenerative and inflammatory diseases of the central nervous system. Studies in animal models of stroke, brain tumours, and experimentally induced allergic encephalomyelitis have demonstrated the phagocytic activity of microgli-al cells and haematogenous macrophages in the lesions
. These studies demonstrate the feasibility of labelling specific inflammatory cell lines.
The existence of a regenerating mechanism in the CNS has emphasized the potential of MR molecular imaging in particular . Although the endogenous activation and recruitment of neuronal progenitors after brain injury remains an elusive goal, an alternative approach has been the administration of transplanted cells in the form of either stem cells or neuronal progenitor cells . For stem cells to be visualized and tracked with MR imaging, they need to be tagged so that they can be detected on MR images. At present two types of contrast agents are available: the gadolinium analogues and the iron oxide nanoparticles. Several approaches have been deployed to enhance cell labelling, to allow in vivo cell tracking by conjugating MR imaging contrast agents to a range of ancillary molecules to enhance their uptake. This can be achieved by coating nanoparticles with internalizing monoclonal antibodies , HIV-Tat peptide , or transfection agents including dendrimers , poly-L-lysine, and lipofecti-on agents . An essential methodological feature of these types of studies is the validation of the imaging findings with histologic techniques.
A wealth of spatial and temporal information on tumour vasculature, metabolism, and physiology can be obtained from non-invasive MRI and MRSI methods. Recent advances in the development of targeted CAs have increased the versatility of MR molecular imaging and will make it an invaluable technique for understanding a complex disease such as cancer.
Was this article helpful?