Statisticians like to say, “All models are wrong, but some are useful.” Mapmakers could say the same thing about their maps. Mapmakers, however, could add that some maps complement each other, or have the potential to do so. Such potential is beginning to be realized for brain maps, which are still—forgive the expression—all over the map. Some brain maps illustrate neuroanatomy in exquisite detail, some capture spatial information about the flicker and glow of signaling events, and some indicate how gene expression or protein expression varies from place to place within the brain.
Unfortunately, brain maps of different types correlate hardly at all. Consider the gaps between whole-brain neuroimaging, which relies on tools such as functional magnetic resonance imaging (fMRI), and neurobiological imaging, which relies on tools such as immunohistochemistry and next-generation sequencing.
Neuroimaging may be undertaken noninvasively, in living organisms, but it suffers from poor resolution. In contrast, neurobiological imaging may achieve molecular and cellular resolution, but it is typically invasive, requiring the use of postmortem tissue.
Bridging the divides between different types of brain maps could help scientists relate large-scale brain functions to the molecular and cellular activities that support these functions. For example, processes such as perception and memory could be analyzed at a deeper level. Also, conditions such as autism, schizophrenia, and neurodegenerative disease could be better correlated with molecular phenomena. Instead of accumulating mere genotype-phenotype associations, scientists could detail pathways and mechanisms.
In this article, we’ll briefly touch on two well-developed types of brain mapping—the creation of molecular brain atlases, and the modeling of brain activities. Finally, we’ll look at nascent efforts to bridge these kinds of brain mapping. Specifically, we’ll report on recent progress in the development of molecular fMRI, an alternative form of fMRI that monitors brain activity through the use of chemical or genetically encoded probes.
Deriving insights from brain atlases
Brain mapping is an audacious pursuit, one that has inspired multiple “big science” initiatives. Among the most prominent initiatives is the Human Brain Atlas. This initiative, which was undertaken by the Allen Institute for Brain Science, is consolidating anatomical and gene expression information.
The Allen Institute’s stated mission is to provide publicly available resources that accelerate basic and clinical research of the human brain in normal and disease states. Examples of how the Allen Institute’s work is assisting researchers can be found at the organization’s website, which presents “Data Stories.” One of these stories describes how researchers at the University of Cambridge merged data from the Allen Human Brain Atlas with MRI brain scans and made an important discovery about the teenage brain.
The scientists used MRI to study the brain structure of almost 300 individuals aged 14–24 years old. By comparing the brain structure of teenagers of different ages, the scientists found that during adolescence, the outer regions of the brain, known as the cortex, shrink in size, becoming thinner. However, as this happens, levels of myelin—the sheath that “insulates” nerve fibers, allowing them to communicate efficiently—increase within the cortex.
These findings appeared in the Proceedings of the National Academy of Sciences, in an article titled, “Adolescence is associated with genomically patterned consolidation of the hubs of the human brain connectome.” The article also indicated that the “topologically focused process of cortical consolidation was associated with expression of genes enriched for normal synaptic and myelin-related processes and risk of schizophrenia.”
Essentially, the researchers found that during adolescence, brain regions that have the strongest link to the schizophrenia risk genes are developing most rapidly. “As these regions are important hubs that control how regions of our brain communicate with each other, it shouldn’t be too surprising that when something goes wrong there, it will affect how smoothly our brains work,” Edward T. Bullmore, PhD, the study’s senior author and head of psychiatry at Cambridge, noted in a statement. “If one imagines these major hubs of the brain network to be like international airports in the airline network, then we can see that disrupting the development of brain hubs could have as big an impact on communication of information across the brain network as disruption of a major airport, like Heathrow, will have on flow of passenger traffic across the airline network.”
Using machine learning to map brain activity
A recent contribution to GEN by Richard A. Stein, MD, PhD, discussed how artificial intelligence technology is helping researchers derive more information from fMRI scans. According to Stein, the technology is helping researchers clarify the brain’s notoriously obscure structure-function relationships. He even describes how an improved understanding of these relationships could help clinicians detect early signs of neurodegeneration in individual patients.
Stein’s contribution appears in full on the GEN website. A representative excerpt is presented in a sidebar (“Mapping How the Brain Organizes Semantic Activity”) that accompanies this article. The excerpt describes how researchers at the University of California are using machine learning, a form of artificial intelligence, to visualize human brain activity.
According to Stein, such work demonstrates how clinicians may use functional imaging to guide early diagnostic, therapeutic, or prognostic decisions. For example, functional imaging could help clinicians improve the management of conditions such as autism spectrum disorder or neurodegenerative diseases.
“Functional imaging,” Stein notes, “would be especially valuable if it could help clinicians distinguish between conditions that would otherwise appear to be the same condition.” He adds that making such distinctions is often difficult in neuroscience, given that “behavior is a low-dimensional reflection of the high-dimensional brain
processes that produce it.”
Extending neuroimaging with molecular probes
To generate functional images of the brain, scientists generally rely on fMRI, a noninvasive technique that tracks the magnetic signals that accompany changes in the oxygenation of hemoglobin. Essentially, hemoglobin is a natural (and convenient) contrast agent or probe of a chemical interaction that correlates with neural activity. What if other probes could be deployed, probes that would allow noninvasive neuroimaging to approach the specificity and resolution of optical neuroimaging?
This question drives researchers in the laboratory of Alan P. Jasanoff, PhD, a professor of biological engineering, brain and cognitive sciences, and nuclear science and engineering at the Massachusetts Institute of Technology. About 10 years ago, the laboratory began developing magnetic proteins that bind to dopamine. When the magnetic proteins, or probes, bind dopamine, magnetic signaling changes occur that correlate with dopamine levels. The probes allow the mapping of dopamine release in the brain.
Last year, Jasanoff and colleagues published a paper in Nature describing how they used their probes to determine how striatal dopamine release shapes local and global responses to rewarding stimulation in rat brains. “We find that dopamine consistently alters the duration, but not the magnitude, of stimulus responses across much of the striatum,” the paper’s authors reported. “Our results reveal distinct neuromodulatory actions of striatal dopamine that extend well beyond its sites of peak release, and that result in enhanced activation of remote neural populations necessary for the performance of motivated actions.”
“When dopamine was released, there was a longer duration of activity, suggesting a longer response to the reward,” Jasanoff said in a statement. “That may have something to do with how dopamine promotes learning, which is one of its key functions.”
In review articles, Jasanoff and colleagues have explained how MRI probes can be sensitized to various neurobiological processes—not just to neurotransmitter release, but also to processes such as calcium signaling and gene expression changes.
With conventional fMRI, brain activity in deep tissue is mapped by monitoring changes in blood flow. However, blood flow may be coupled to neural activity through many different physiological pathways. “As a result,” Jasanoff points out, “the signal you see in the end is often difficult to attribute to any particular underlying cause.”
In contrast, calcium ion flow can be directly linked with neuron activity. When a neuron fires an electrical impulse, calcium ions rush into the cell. “Concentrations of calcium ions are closely correlated with signaling events in the nervous system,” Jasanoff emphasized. “We designed a probe with a molecular architecture that can sense relatively subtle changes in extracellular calcium that are correlated with neural activity.”
Jasannof has also indicated that his laboratory intends to develop fMRI-detectable sensors for gene expression. To pursue this work, the laboratory will take advantage of an award it received to map how gene expression changes in the brain in response to drugs of abuse. “Our studies will relate drug-induced brain activity to longer-term changes that reshape the brain in addiction,” Jasanoff noted. “We hope these studies will suggest new biomarkers or treatments.”
Finally, Jasanoff and colleagues have expressed optimism that the kinds of fMRI probe technology that have been used in animal studies could one day be used in human subjects. In Current Opinion in Neurobiology, the scientists wrote that applications of molecular fMRI in humans would become feasible if probes could be delivered noninvasively, and if probes could be sufficiently nontoxic and stable for human use.