08/13/2013 12:35 pm ET | Updated Oct 13, 2013

WANTED: Neuro-quants

By Brian Caffo, Martin Lindquist and Ciprian Crainiceanu

Quantitative neuroscience is a flourishing discipline. Take as evidence President Obama's recently announced BRAIN Initiative (Brain Research through Advancing Innovative Neurotechnologies). Its ultimate goal is the mapping of the human functional "connectome," the collection of all neuronal connections. This groundbreaking initiative is creating significant new demand--and unmatched opportunity--for "neuro-quants," biostatisticians working in the neurosciences.

The term connectome is a combination of connection and genome. The parallel with the genome is no coincidence, as decoding the connectome shares many similarities. Like the sequenced genome, the staggeringly difficult task of decoding the connectome is followed by the much more daunting task of making sense of it. Consider that the number of possible connectivity patterns among only 50 neurons dwarfs the number of atoms in the known universe! Further, the human brain only can be measured with different tradeoffs, resulting in a confusing hierarchy of resolutions and technologies. Our technological focus lies in the intersection of biostatistics and neuroimaging (neurological imaging).

Biostatistics is the field of statistics applied to the biological and medical sciences (think "Moneyball" for biomedical applications). It plays a crucial role in obtaining and understanding results in the neurosciences. In particular, biostatisticians who can speak the common language of these problems promise to be central to the BRAIN Initiative.

The goals of this and related initiatives are arguably among the most complex, important and challenging issues in science today. For these endeavors to be successful, a legion of "neuro-quants" is needed to make sense of the massive amounts of data being generated.

Working on the biostatistical aspects of neuroimaging is a tremendous opportunity for college graduates who have rigorous quantitative training in statistics, mathematics, computer science or engineering. Indeed, considering the broad demand for statisticians in this area, we recall the famous quip from Jeff Hammerbacher, a former Facebook engineer, who said, "The best minds of my generation are thinking about how to make people click ads. That sucks." We contend that working in biostatistics using complex neuroimaging data to solve the biggest puzzle of them all--understanding the human brain--is a worthy alternative.

Exactly what types of neuroimaging problems do biostatisticians deal with? First, neuroimaging is an umbrella term for an ever-increasing number of techniques designed to study the brain (without actually cutting a person's head open). These include a variety of rapidly evolving technologies for measuring brain properties, such as structure, function and disease pathophysiology. These technologies are applied in a vast collection of medical and scientific areas of inquiry. Biostatistical neuroimaging is relevant in nearly every disease or disorder of the brain, such as Alzheimer's disease, autism, lead exposure and multiple sclerosis, to name only a few.

Each imaging technique and application area results in enormous amounts of data. With new studies collecting repeat measurements on thousands of subjects over multiple years, the size of data sets are becoming unimaginably large and, more importantly, complex. Complexity goes hand in hand with the modern Big Data problem. In the absence of complexity, Big Data applications are analytically routine. Given the complexity and size of neuroimaging data, simple reproducible data-analytic methods, causal thinking, data exploration, hypothesis confirmation and careful design of experiments will become increasingly important, or, in other words, will require the expertise of biostatisticians!

What should a biostatistician specializing in neuroimaging study? In our program at The Johns Hopkins University, we strive for our students to master the foundational (theoretical) aspects of statistics as well as its practical application and associated computational principles and skills. In addition, it is crucial for students to develop excellent communication skills, since working at the core of a highly dynamic scientific area requires the ability to speak multiple scientific "languages" and learn new ones rapidly. Students also must learn enough neuroscience and imaging technology to contribute in meaningful ways.

The training of the needed mass of neuro-quants will require dedicated funding from relevant government sources and private foundations. Such dedicated resources would be in a funding agency's own interest, as collecting large and complex neurological data without a serious investment in its biostatistical analysis is a recipe for a minimal return on investment. Furthermore, the scope of the problem requires academic departments to meet the challenge of providing this training and academic homes for this new kind of quantitative scientist.

It remains unclear how the larger statistical and funding communities will respond. However, their efforts in the genomics revolution provide a source for optimism. There, statisticians brought key statistical insights into the most pressing technological and scientific problems of the day. However, the dramatic demand for biostatisticians comes at the same time as a significant need for statisticians in many other disciplines, including genomics, biology, medicine, finance and public policy, to name a few.

It is a wonderful time--in the midst of a data revolution--to be a biostatistician. Quantitative neuroscience provides the opportunity for statisticians to be at the forefront of the most exciting challenge in science today by providing the critical skills needed to address it properly. We hope the discipline is up for the task and the coming generation of biostatisticians embraces the challenges of neuroimaging in the same manner previous generations embraced the genomics revolution.

Caffo, Lindquist and Crainiceanu co-lead a research working group called the Statistical Methods and Applications for Research in Technology (SMART) Group ( and are members of the department of biostatistics in the Bloomberg School of Public Health at The Johns Hopkins University.