Funded Projects (as of July 2008)

Nanowires: Development of polymer-based electrodes for chronic neural recordings in vivo and in vitro.

PI: Ian Hunter (MIT Dept. of Mechanical Engineering)

Co-PIs: Emilio Bizzi and Martha Constantine-Paton (McGovern Institute)

Other personnel: (Andrew Taberner, Bryan Ruddy, Giovanni Talei-Franzesi, Dept. of Mechanical Engineering; Woong-Jin (Chris) Bae, McGovern Institute)

Conducting polymers have great potential as new materials for electrode construction. Compared to traditional metal or glass electrodes, polymers such as polypyrrole are flexible and highly biocompatible, and they can be made extremely thin. These properties will be especially valuable for the construction of high-density electrode arrays that can be implanted chronically in the brain. We are fabricating polypyrrole nanowires and exploring their use as intracortical recording electrodes. Ultimately we hope to develop a new generation of electrodes that can be stably implanted in the brain for long periods of time, both for research and clinical applications.

We are also working to produce polypyrrole electrode arrays on a tissue culture surface. These arrays will be used to deliver patterned stimulation to neurons in culture, in order to explore synaptic plasticity rules and neural network effects in vitro. Such a system may ultimately form the basis of a new platform for drug discovery, by allowing high-throughput screening of chemicals that modify the properties of neural networks in therapeutically useful ways.

Alternate Indicators in MRI: Design and synthesis of a new membrane-permeable MRI sensor for functional imaging of neuronal calcium

PI: Stephen Lippard (MIT Dept. of Chemistry)

Co-PI: Alan Jasanoff (McGovern Institute)

Other personnel: Xiao-an Zhang (Dept. of Chemistry)

Functional magnetic resonance imaging (fMRI) is a well-established noninvasive neuroimaging method that measures brain activity indirectly through local changes in blood flow and oxygenation. fMRI is limited, however, by the slow time-course and coarse spatial scale of the hemodynamic response, and by confounding factors unrelated to neural activity. It is thus of great interest to develop more direct fMRI methods that combine the cellular specificity of classical neurophysiology techniques with the noninvasiveness and whole-brain coverage of MRI. Such a technique would give us a greatly expanded view of brain function and, potentially, of human brain disorders.

We have recently developed a zinc MRI sensor based on water-soluble manganese porphyrin. This agent is the first reported cell membrane-permeable contrast agent for metal ion sensing. The establishment of this chemical platform allows us to design a novel calcium sensor for MRI, based on a mechanism of Ca-dependent ligand gating of a water coordination site on the metalloporphyrin complex. We plan to synthesize this molecule and test its utility as a MRI sensor in the rodent brain.

Use of virally encoded genes to optically record and manipulate neural activity in the songbird brain

PI: Michale Fee (McGovern Institute)

Co-PI: Carlos Lois (Picower Institute for Learning and Memory)

Other Personnel: Tim Gardner (McGovern Institute)

To understand brain function at the level of neural circuits, we wish to visualize the activity of many neurons simultaneously, with high temporal resolution. We are developing a genetic approach to this goal based on a new calcium indicator, GCaMP2, that has recently been developed by Dr Junichi Nakai at RIKEN Brain Science Institute. This indicator, consisting of a circularly permuted GFP fused to calmodulin, has been used to image Ca signaling in transgenic mouse heart. We will use this system for monitoring neural activity in vivo, using viral vectors to deliver the transgene reporter to neurons in the song bird nucleus known as HVC. Our goal is twofold: to examine spatio-temporal patterns of activity in HVC during bird song learning (a model for the acquisition of complex motor skills such as speech), and more generally to demonstrate the utility of GCaMP2 as a genetic indicator of neural activity. If successful, we also plan to combine this approach with the use of channelrhodopsin-2 and/or halorhodopsin, which should allow simultaneous optical recording and manipulation of neural activity in vivo.

Application of Optical Control of Neural Activity to Analysis of the Neural Substrates Causally Mediating Cognition

PI: Ed Boyden (MIT Media Lab)

Co-PI: Robert Desimone (McGovern Institute)

Other Personnel: Xue Han (McGovern Institute)

We have recently shown that the light-activated proteins channelrhodopsin-2 and halorhodopsin can be used to activate and inhibit neurons in response to light of different wavelengths. We are now developing precisely-targetable fiber arrays and in vivo-optimized expression systems to enable the use of this tool in awake, behaving primates. By combining these technologies with behavioral and physiological experiments, we hope to open up new horizons on the analysis of cognition. In the longer term, it may be possible to apply a similar approach to the human nervous system; potential clinical applications include the suppression of epileptic seizures, restoration of visual perception in patients with retinal degeneration, or deep brain stimulation for conditions such as Parkinson's disease.

Ultra-Low-Power Electronics for Chronic Wireless Brain-Machine Interfaces

PI: Rahul Sarpeshkar (MIT Dept. of Electrical Engineering & Computer Science)

Co-PI: Michale Fee (McGovern Institute)

The development of brain-machine interfaces holds great promise as a new therapeutic approach to neurological disease and injury. An implanted prosthetic device could, for example, enable a paralyzed patient to control a computer or mechanical device directly through signals recorded from the brain, without any need for manual controls. Successful implementation of this concept will require local processing of information from implanted electrode arrays followed by transmission (preferably wireless) across the skull. This must be accomplished with minimal power consumption in order to prolong battery life and to avoid excessive dissipation of heat within the cranium. We are developing an ultra-low power neural amplifier along with circuits and algorithms for compression of neural recording data. These systems, which exploit a combination of analog and digital processing, will be tested in the songbird brain, with the aim of establishing a chronic interface for intracranial recording and telemetry.

Carbon nanotubes

PI: Jing Kong (MIT Department of Electrical Engineering and Computer Science)

Co-PI: Emilio Bizzi (McGovern Institute)

Bizzi and Kong are exploring the use of carbon nanotubes as an alternated, biocompatible alternative material. To study long-term changes in the brain, researchers need to make chronic recordings of neural activity. But the standard electrodes can damage brain tissue and lose their electrical contacts over time, so there is a need for alternative materials. Bizzi, who researches the control of movement, wants to use long-term recordings initially for basic research but ultimately for prosthetic devices in human patients. Such devices might, for example, allow a paralyzed patient to control a robotic arm or a computer directly from the brain.

Rigid sheaths for flexible nanowires

PI: Robert Langer (MIT Department of Chemical Engineering)

Co-PI: Emilio Bizzi (McGovern Institute)

Bizzi is also exploring another alternative electrode material made from thin flexible strands of conducting polymers, called nanowires. These polymers are expected to produce less damage to brain tissue, but they are difficult to insert into the brain. Langer and Bizzi will explore one potential solution: a biodegradable coating that can provide temporary stiffness but disappears after insertion.

Optical control of signaling molecules

PI: Shuguang Zhang (MIT Center for Biomedical Engineering)

Co-PI: Ed Boyden (McGovern Institute)

Ed Boyden, a member of the MIT media lab and an associate member of McGovern Institute, is a pioneer in the development of optical tools for manipulating electrical activity in neurons. He plans to extend this approach to manipulate intracellular signaling, in collaboration with Shuguang Zhang, director of the MIT Center for Biomedical Engineering and an expert on protein engineering. If successful, this could be a valuable method for determining the function of signaling pathways in vivo and for identifying potential targets for drug development.

Analyzing MRI data

Nancy Kanwisher in the McGovern Institute and Polina Golland, Associate Professor, EECS and Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT.

One challenge for neuroscience research is analyzing the very large datasets produced by brain imaging studies. Two new MINT projects will explore different computational approaches, using data from Nancy Kanwisher at McGovern Institute.

In the first, Polina Golland in the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) will use Kanwisher's fMRI data to search for brain areas that respond to specific categories of visual objects. Several such areas are known to exist, but the current methods for identifying them only work if they are in the same location in every person. Golland has developed a computational method that avoids this assumption. She and Kanwisher plan to test the method on a large body of brain scanning data to determine whether it can reveal the existence of new brain areas that cannot be found with existing methods. Golland also collaborates with neurosurgeons to analyze imaging data to help surgical planning for brain tumor surgery.

Detecting patterns in MRI

PIs: Navia Systems

Co-PI: Nancy Kanwisher (McGovern Institute)

In a related project, Kanwisher will work with Navia Systems Inc, a startup company founded by former MIT students in the Department of Brain and Cognitive Sciences. Navia uses proprietary computational methods, licensed from MIT, to identify patterns in large complex datasets and assign significance to them Navia to develop algorithms for detect patterns in large, complex data sets, such as neural imaging studies. The algorithms look for clusters that exist within the data without any prior assumptions. For example, does the category 'animal' activate one tight cluster or a million scattered points? How do patterns for bats, birds and cats differ? These algorithms will help reveal how the brain classifies the world, and could enhance the study of brain disorders, for example by identifying relationships between brain activity, genetics and clinical diagnostic categories. It could also be adapted to other fields, such as seismic data to locate oil fields.

Precise detection of brain cells

PI: Mehmet Fatih Yanik (MIT Department of Electrical Engineering)

Co-PI: Ann Graybiel (McGovern Institute)

Researchers often use a method called laser capture microdissection (LCM) to analyze single cells within a tissue -- for example to identify genetic abnormalities that distinguish tumor cells from their healthy neighbors. The ability to analyze single cells is especially important in the brain, where cells of many different types are closely intermingled. However, because of the brain's dense meshwork of connections, it is often impossible to cleanly remove a single cell without contamination from adjacent cells. To solve this problem, Ann Graybiel of the McGovern Institute will collaborate with Mehmet Fatih Yanik in the MIT department of Electrical Engineering and Computer Science, who developed a laser producing extremely local concentrations of very high energy pulses lasting just a femtosecond (1 millionth of a nanosecond, or 10-15 of a second). Yanik plans to develop a 3-dimensional laser-based cutting method that can dissect a single cell from its neighbors. Graybiel hopes to apply these new methods to her studies on the basal ganglia, brain regions implicated in Parkinson's disease, addictive behaviors and mood disorders.

Back to MINT Program page.