Quantitative skills in undergraduate neuroscience education in the age of big data

Abstract

For over a half-Century, the mathematics requirement for graduation at most undergraduate colleges and universities has been one year of calculus and a semester of statistics. Many universities and colleges offer a neuroscience major that may or may not add additional mathematics, statistics, or data science requirements. Today in the age of Big Data and Systems Neuroscience, many students are ill-equipped for the future without the tools of computational competency that are necessary to tackle the large data sets generated by contemporary neuroscience research. Required courses in statistics still focus on parametric statistics based on the normal distribution and do not provide the computational tools required to analyze big data sets. Undergraduates in STEM fields including neuroscience need to enroll in the Data Science courses that are required in the social sciences (e.g., economics, political science and psychology). Contemporary systems neuroscience is routinely done by interdisciplinary research teams of statisticians, engineers, and physical scientists. Emerging “NeuroX-omics” such as connectomics have emerged along with genomics, proteomics, and transcriptomics, all of which deploy systems analysis techniques based on mathematical graph theory. Connectomics is the 21st Century’s functional neuroanatomy. Whole brain connectome research appears almost monthly in the Drosphila, zebra fish, and mouse literature, and human brain connectomics is not far behind. The techniques for connectomics rely on the tools of data science. Undergraduate neuroscience students are already squeezed for credit hours given the high-prescribed science curriculum for biology majors and premedical students, in addition to required courses in social sciences and humanities. However, additional training in mathematics, statistics, computer science, and/or data science is urgently needed for undergraduate neuroscience majors just to understand the contemporary research literature. Undoubtedly, the faculty who teach neuroscience courses are acutely aware of the problem and most of them freely acknowledge the importance of quantitative analytical skills for their students. However, some faculty members may feel that their own math and statistics knowledge or other analytical skills have atrophied beyond recall or were never fulfilled in the first place. In this commentary I suggest that this problem can be ameliorated, though not solved, through organizing workshops, journal clubs, or independent studies courses in which the students and the instructors learn and teach each other in short-course format. In addition, web-available teaching materials such as targeted video clips are plentifully available on the internet. To attract and maintain student interest, qauntitative instruction and learning should occur in neuroscience context.