The Program in Neural Computation (PNC) is a graduate training program in computational neuroscience for students seeking training in the application of quantitative approaches to the study of the brain. Specifically the program is designed to take advantage of the world class strengths of Carnegie Mellon University and our partner institution, the University of Pittsburgh, in areas including computer science, machine learning, statistics and dynamical systems and to train students to apply these tools to critical problems in neuroscience.
The PNC is based on the highly successful, but non degree granting training program of theCenter for the Neural Basis of Cognition (CNBC), which was established in 1994 to foster interdisciplinary research on the neural mechanisms of cognitive function. The CNBC now comprises 139 faculty having appointments in at least 17 departments (CMU: computer science, machine learning, robotics, statistics, biomedical engineering, biological sciences, psychology; PITT: mathematics, neurobiology, neuroscience, neurology, psychiatry, communication science, psychology, radiology, bioengineering, history and philosophy of science). Both these programs require training in neuroscience ranging from the cellular to cognitive levels. The PNC also requires extensive training and a research focus in at least one quantitative area. Thus, the PNC program will be especially attractive to students coming from majors in math, computer science, statistics, engineering or physics who want to apply approaches from these disciplines to the understanding of brain function.
Neuroscientists are applying new technologies to acquire and analyze large data sets, and more and more are using quantitative models to understand the great complexities of neurobiological systems. As a result, quantitative methods have become centrally important in the field of neuroscience. At the same time, the number of investigators with requisite skills who are actively engaged in this domain of research is relatively small. There is a widely recognized need for increased training in the application of computational, mathematical, and statistical methods to biology and medicine, and to problems in neuroscience in particular. These points have been emphasized in recent articles about the field of neuroscience and in recent NIH calls for grant applications.
Neuroscience has historically been heavily influenced by quantitative approaches. There have been important advances through the use of quantitative methods in neurophysiology, and there has been a continuing stream of related work within mathematics and applied physics. More recently, engineers, computer scientists, and statisticians have contributed to the field, expanding further the definition of computational neuroscience. We believe that the environment at Carnegie Mellon and the University of Pittsburgh have much to offer to students interested in these approaches and thus we have founded a PhD Program in Neural Computation here at Carnegie Mellon, in collaboration with colleagues at the University of Pittsburgh. This program is designed to attract students with strong quantitative backgrounds and to train them in quantitative disciplines relevant to neuroscience and also to provide them the essential background in experimental neuroscience. In doing so we would bring to bear the special strengths of our institution and the unique neuroscience community here in Pittsburgh. Training faculty and courses will be drawn both from CMU and Pitt as described. The PNC PhD program is designed to capture students with backgrounds in computer science, physics, statistics, mathematics, and engineering who are interested in computational neuroscience, particularly with an emphasis on quantitative methods from computer science, machine learning, statistics and nonlinear dynamics.
The program consists of the following core activities:
Coursework and program milestones are described elsewhere.
Advising and Student Evaluation
Students are assigned early in their first year an advisor to guide the student in selecting courses and beginning his or her initial research project. By the end of the summer following the first year students must identify a thesis advisor. This faculty member also will serve as the student’s academic advisor.
The Program in Neural Computation is supervised by a faculty steering committee appointed by the CNBC education committee. The Co-chairs of the CNBC education committee will serve as directors of the graduate program. Graduate students meet with this committee for approval of their curriculum, particularly for elective selections. This committee also approves thesis committees and project committees. Twice each year, the faculty steering committee reviews the progress of each student in all aspects of the program. The results of this evaluation will be communicated to the student by the Co-directors of the CNBC.
Ph.D. in Statistics and Neural Computation:
This program allows students to pursue a Ph.D. that combines Ph.D.-level training in statistics with a solid understanding of the elements of neuroscience, as in the PNC. Students complete the requirements for a Ph.D. in Statistics while also fulfilling the core requirements for the PNC Ph.D. by taking courses in cellular, cognitive, and systems neuroscience, as well as computational neuroscience, and gaining exposure to methods of data collection in at least one experimental laboratory. Because there is some overlap of requirements, this joint Ph.D. program should not take very much longer than either single program taken separately.
Ph.D. in Neural Computation and Machine Learning :
This program allows students to pursue a Ph.D. that combines Ph.D.-level training in machine learning with a solid understanding of the elements of neuroscience, as in the PNC. Students complete the requirements for a Ph.D. in Machine Learning while also fulfilling the core requirements for the PNC Ph.D. by taking courses in cellular, cognitive, and systems neuroscience, as well as computational neuroscience, and gaining exposure to methods of data collection in at least one experimental laboratory. Because there is some overlap of requirements, this joint Ph.D. program should not take very much longer than either single program taken separately.
M.D.-Ph.D. Program with the University of Pittsburgh School of Medicine:
The Program in Neural Computation participates in a combined M.D.-Ph.D. Program with the University of Pittsburgh School of Medicine, to offer M.D. degree from the University of Pittsburgh and Ph.D. from Carnegie Mellon University. The aim is to allow physicians to blend computational research and clinical perspectives in treating patients.
Prospective students should apply directly to the University of Pittsburgh School of Medicine, indicating an interest in the Ph.D. Program in Neural Computation at Carnegie Mellon University. During the first semester of the second year of medical school, the student should submit an application to the Ph.D. program, which may include supporting documents previously submitted to the University of Pittsburgh School of Medicine.
Students formally enter the Ph.D. program after completing their second year of medical school, although research may start as soon as the summer before the first semester of medical school and during the subsequent two summer semesters. This allows the student to gain a total of six months of research before officially entering the Ph.D. program. As such the student will enter the PNC program as a 2nd year PhD student and the course requirements for experimental neuroscience will be waived. Completion of the Ph.D. program is targeted at 3-4 years. The student then returns to the University of Pittsburgh School of Medicine to completes the last two years of M.D. training.
Other program activities:
PNC students participate with CNBC certificate students in the following co-curricular activities.
The CNBC colloquium series is a student-run speaker series that brings eminent scientists to Pittsburgh. Students have played a major role in the selection and hosting of speakers throughout the years; faculty provide input on speaker selection, but the students do all the voting and interact extensively with the speakers during their visits.
The Brain Bag research seminars meet approximately bi-weekly throughout the academic year on Monday evenings. At each Brain Bag, a student gives a brief talk describing research in progress.
The CNBC retreat has been an important part of the process of creating an integrated inlellectual community. The retreat provides a venue for informal discussions of important topics of general interest to members of the community, and introduces the students and faculty to recently added CNBC faculty members through a series of 1/2 hour talks, held on a Saturday and the following Sunday morning. Another important element of the retreat is a set of interdisciplinary evening discussions. Participants break up into small groups cutting across levels of analysis and research methods to discuss topics of interdisciplinary interest.
CNBC Friday Seminars are an occasional seminar series at which in-house and outside speakers present in an informal and interactive setting.