Philadelphia University + Thomas Jefferson University

Research Projects

Research Projects

Optimization of Diffusion-Based White Matter Mapping with Invasive Electrocorticography

optimization of diffusion-based white matter mapping

Tractography is a three-dimensional modeling technique that has been used for approximately the past two decades to help visualize the structural connections between different parts of the brain.  This method is based on diffusion magnetic resonance imaging (dMRI), which images the way in which water moves within the brain.  Since we know that water flows in the path of least resistance, we can assume that this path represents the connections between cells in the brain – or in other words, that it represents axons and white matter tracts that connect neurons.  This concept forms the basis of tractography and image-based white matter mapping.  As the only non-invasive method of visualizing white matter tracts in living human subjects, its clinical and research applications have continued to expand.  It is used clinically for planning surgeries in patients with brain tumors and epilepsy; and its potential is immense as its use as a diagnostic and prognostic tool in stroke, multiple sclerosis, Alzheimer’s disease, and spinal cord disorders has been promising.

Unfortunately, there are significant limitations to this technology.  Current MRI technology is capable of imaging down to a resolution of a 1mm cube (known as a ‘voxel’).  Since white matter tracts are smaller than 1mm, each voxel can contain many different white matter tracts, which may travel in different directions.  In an attempt to image these different tracts, more recent MRI sequences are capable of capturing a greater number of directions of water flow within a single voxel.  In order to make sense of this additional information, however, the tractography technique for modeling the fibers has become more complex.  More importantly, these models are based on statistics and complex mathematical equations – none of which have been proven to be accurate in the living human brain.  While these methods have been compared to microscopic evaluation in animals and in cadaveric specimens, there is no way for us to do the same in a living human.  As a result, the way in which we visualize the white matter tracts can be very different based on the different mathematical equations used in the process of modeling.

At the Jefferson Comprehensive Epilepsy Center, we commonly treat patients with drug-resistant epilepsy.  In the evaluation of these patients for surgery, we routinely implant a number of electrodes into the brain in order to determine where seizures are originating, where they are spreading, and how this path is related to important areas of the brain. Patients are usually in the hospital with these implanted electrodes for an average of two weeks before we have enough information to make a decision about surgery.  At the same time, this practice gives us a unique opportunity to study the electrical recordings of a living human brain.

We can use the electrical recordings from these implanted (stereoelectroencephalography) electrodes to define pathways of electrical activity and show how different brain regions communicate with each other.  With electrodes implanted both in gray matter (neurons) and white matter (axons), we will be able to see how electrical activity flows out of one group of cells, along a white matter bundle, and to another group of cells in a different part of the brain. We believe that this is the best way to see where white matter tracts lie in the living human brain. By mapping the multiple pathways of this electrical activity, we plan to create a standard against which current tractography modeling techniques can be compared.  The main idea behind this project is that maps of electrical activity in the brain can be used to refine and optimize methods of imaging white matter tracts in living human subjects. Specifically, we intend to use these maps of electrical activity to gain a better understanding of three methods of dMRI-based tractography: diffusion tensor imaging (DTI); high-angular resolution diffusion-weighted imaging (HARDI); and neurite orientation dispersion and density imaging (NODDI).  This will be the first study to use electrical recordings from the brain to refine methods of tractography.

Prediction of Patient Response to Dopamine and DBS using Advanced Imaging Methods

prediction of patient response to dopamine

Neuronal loss and dopamine depletion alters motor signal processing and propagation between the cortical motor areas, the basal ganglia and the thalamus resulting in the motor manifestation of Parkinson’s disease (PD). Abnormal neural connections and activity within these circuits have been demonstrated in animal models of disease as well as in humans suffering from movement disorders.  With the aim of better understanding changes in functional connectivity (FC) and anatomical connectivity (AC) that occur with disease progression, we are exploring how both FC and AC of cortico-basal-ganglia-thalamic circuits change in patients with advanced PD. Our findings to date have helped to translate our understanding of these networks from animal studies to human patients.  We believe that advanced MR imaging may reveal a useful imaging biomarker that can help to predict response to treatment in patients suffering from PD.

Computer Assisted Diagnostic And Planning Applications With Continuous And Non-Disruptive Data Collection

Computer assisted diagnostic

Epilepsy is a chronic condition characterized by recurrent seizures and affects over three million Americans. This disease has a spectrum of subtypes and remains a symptom with a number of underlying causes. In addition, one million (30 percent) have seizures that cannot be controlled despite maximal medical therapy. Poorly controlled epilepsy is disabling, and leads to unemployment, recurrent injury, and a high risk of death. In patients with drug-resistant epilepsy, a series of tests may be performed to help locate the origin of the seizures and provide clinicians with important information about the brain that is needed to consider a patient for surgery. Clinical teams attempt to crack the case by bringing together as many congruent clues found across different fields such as neurology, neuropsychology, neuroradiology, neuroelectrophysiology, and neurosurgery. Elements of these fields are like pieces of a jigsaw puzzle. Each element, by itself, is to no avail; but taken as a whole, the individual clues can produce a powerful diagnostic tool that will lead to an optimal treatment strategy to stop the seizures. Today research tools have been developed to address the needs within these individual fields. Yet, these solutions have been created in isolation and there is no commercial system currently available that permits the assembly of all the pieces.

Moreover, these pieces constitute a wealth of data for each patient, which are acquired at various stages of diagnosis and treatment. Yet, there is no commercial system currently available that permits the acquisition and storing of this information electronically. Consequently, data acquired at one stage of the patient's evaluation by a subset of the clinical team (e.g., the electrophysiologists) cannot easily be shared with another subset of the team (e.g., the neuroradiologists or the neurosurgeons). The lack of electronic data capture systems also prevents aggregating data acquired from large populations at multiple clinical sites. This limits the extent to which best practices at leading clinical sites can be compared as well as the transfer of expertise from leading research centers to smaller clinical sites.

Over the last decade, a research system that addresses similar issues encountered for another neurosurgical procedure, deep brain stimulation (DBS), has been developed at Vanderbilt with NIH support and integrated in the clinical flow at this institution.  In a collaborative effort with researchers at Vanderbilt, we are working on an analogous system capable of supporting and streamlining the processing of the vast amount of multimodal data that must be processed in order to make appropriate clinical decisions in patients with epilepsy.