Background

Understanding the function and architecture of the living human brain is among the most compelling technical challenges of our time. Brain circuitry ultimately shapes the behavior and thoughts of individuals, and preventing or repairing disruptions of that circuitry by disease or trauma comprise a major unmet societal challenge.

Techniques for imaging the structure and function of the in vivo human brain are of particular value because they can reveal connections between neurological activity within specific regions of the brain and corresponding perceptions, thoughts or behaviors. To advance imaging in humans, potentially outside a laboratory setting, there is a great need for techniques that are noninvasive, in contrast to the emerging and complementary class of brain imaging techniques that include ex vivo whole brain imaging of cleared brain structures and optogenetic imaging of cell-scale neuronal activity in animals.

New capacities to image brain function or structure noninvasively will be the main subject of this workshop. Such technologies could revolutionize the way we understand the connections between brain physiology and human behavior. Limitations of existing techniques include constraints on subject motion during imaging, requirements for elaborate electromagnetic shielding from the environment, requirements for active cooling of imaging system sensors, and system resolution that is much coarser (millimeter to centimeter scale) than that required to detect activity corresponding to individual neuronal signaling.  One of the objectives of the workshop will be to bring together global experts in the field to identify these limitations and develop scientific pathway for overcoming them.

Noninvasive neural imaging frequently entails local and selective perturbation of neural circuits combined with monitoring the neural response using optical, electrical, magnetic, chemical, and molecular modalities. Other modalities monitor response without perturbation. An important consideration is to minimize any intrusion of the brain to reduce possible damage to neuronal circuits and normal brain activities. For the purposes of this Workshop, noninvasive brain imaging technologies are defined as methods that do not require the implantation of devices. Technologies of interest may nevertheless involve molecular or genetic reporters introduced into the subject to provide externally detectable signals linked to neural activity.

Techniques that enable non-invasive imaging are desired at different spatial and temporal scales and in response to different functional modalities (e.g. electrical, chemical, degree of oxygenation, etc.), each of which will enhance information obtained at other scales and for other modalities. Large-scale imaging of global brain activity is an example at one end, whereas synapse-spike level description of local circuits is an example at the microscopic level. Only by relating activity across the various levels and their interactions can we gain a sufficiently detailed, mechanistic understanding of brain network activity to relate this activity to global function. It is also desirable to image across multiple temporal scales from neural activity associated hemodynamic and metabolic signals, neural oscillations, and neuronal action potentials that reflect communication across neuronal ensembles. Imaging resolution that is closer to cellular-size is desired because the common currency of brain communication is the action potential, which allows distant communication among the diverse neuronal types in each brain region.

One example of Noninvasive brain imaging is magnetoencephalography (MEG), which uses magnetic field sensors with ultra-high sensitivity comparable to that of commonly utilized Superconducting Quantum Interference Devices (SQUIDs). Search coils and, most recently, atomic magnetometers have both demonstrated the capability for measuring magnetic fields from the brain. Relating such activity to finer-grain neural responses, however, is challenging. The observed responses are a sum of all electrical activity in the brain, and only synchronous, spatially aligned activity can be extracted. Determining the neural sources of the observed activity is a mathematically under-constrained problem that can only be solved with appropriate a priori information. This comprises a problem of large-scale mathematical complexity requiring sophisticated signal processing. But what are the next generation magnetic sensing devices and technologies?

Advances in detection technologies may not, on their own, be sufficient to realize the goal of sensitive non-invasive neural imaging due to the weakness of intrinsic bioelectromagnetic signals. It may be necessary to enhance these signals through the use of appropriate molecular or nanostructured imaging agents, and this is another example of a topic that would be of interest to this workshop. Virtually every imaging modality in biomedicine requires the use of such agents to visualize specific aspects of anatomy and physiology (including other noninvasive techniques such as PET, CT, fMRI, and ultrasound). What agents could be developed specifically to enable more sensitive non-invasive detection of neural activity under ambient and ambulatory conditions? For example, magnetic nanostructures introduced via the circulation are already used to enhance hemodynamic signals in fMRI. New approaches are needed to provide similar enhancement in more direct electromagnetic detection of neural signaling. Reporter technologies with the potential to enable breakthrough performance include nanoparticles, hyperpolarized nuclei (enabling low-field, potentially ambulatory MRI), biologically-derived magnetic materials, and acoustically active nanostructures for transcranial ultrasound.

It is also clear that advances in noninvasive imaging of brain function is not just a hardware challenge; it is also a computational/signal-processing challenge that must use neuroscience knowledge in order to get interpretable results.