The “infinite brain” refers to systems that can connect to the brain and replicate its function. These have been variously called the digital or electronic brain, such as by Turing, or the cybernetic brain, by the cybernetician William Grey Walter. The desire to build such systems in the past has been central to the founding of computer science, and also led to the development of brain-computer interfaces (BCIs)—devices that can read and stimulate the electrical activity of neurons in the brain, cognitive and artificial intelligence (AI) frameworks, and more.
While we have not yet realized the infinite brain, we have made tremendous progress. BCIs, paired with computational cognitive and AI frameworks, have been crucial in understanding brain function and dysfunction, treating debilitating diseases, restoring lost function, and even augmenting human capability. Moving forward, we find that computer architecture and system design is going to be the key driver to support the growth of these technologies, and eventually realize the ambitious vision of the infinite brain.
This course will cover (1) basics of BCIs, example applications and computational demands, (2) fundamental tradeoffs in computer architecture and system design encountered in developing processors for BCIs, (3) some advanced computer architecture concepts relevant to BCI processor design tradeoffs like super-scalar processing, caching, accelerators etc., building on prior architecture undergrad courses (e.g., COMP 311, 541), (4) open problems, (5) future directions like brain-inspired computer architectures, and (6) logistics permitting, a gentle introduction to concepts in chip design for BCI data processing and computer-aided design (CAD) tools that are used to develop BCI processors. The course is targeted for early graduate students and advanced undergraduates.
The course will consist of lectures, reading and discussion of research papers, homework assignments, processor chip design lab problems, and an applied course project. Prepare to read about two papers per week, additional lighter reading, on BCIs, and demonstrate a working team project by the end of the semester. Paper summaries, homework assignments and the course project along with the report will be graded.
Background in one or more of computer architecture, organization, digital design, systems (e.g., COMP 311, 541) is recommended, but students with alternative backgrounds in relevant topics (e.g., neuro-engineering, neural/digital signal processing, biomedical IoT/CPS) are also welcome to take the course by discussing with the instructor.
You will understand what BCIs are, gain exposure to the kinds of applications they run, understand the challenges in designing BCI processors and systems, and develop the ability to critique computer architecture and system designs for BCIs.
Project (30%): presentation (15%) + report (15%), homeworks (20%), processor design labs (20%), technical paper reviews (10%), lead presentation (10%), participation and discussion (10%). If logistics for CAD tool exercises don't pan out, the weightage would be assigned to the project, split equally between the presentation and the report.
Undergraduates will receive standard letter grades based on the university scale (A, A-,...), and graduates will receive grades based on the department scale (H+, H,...). Grading will use both objective (homework scores) and subjective measures (like quality, significance and novelty of project submissions used in assessing technical work, and class participation).
Course projects are expected to involve one or more of (but not limited to) FPGA or microcontroller prototyping, software languages, compilers, emulators, tools, runtimes or kernels, application design and optimization, or as relevant to computer systems for BCIs, broadly construed. Highly innovative projects could also lead to publishable manuscripts.
Students can propose course projects individually or as teams of two. There will be a mid-term review of progress. Team members hold collective responsibility for the project.
Project report: The report will be a pdf that is 10-12 pages long, with a 12-point font, and 1-inch margins. It must include a title, team member names and emails, brief description of the contributions of team members, any overlap with projects in other courses or research, and an abstract (< 200 words). The report should contain separate sections, with appropriate subsections as follows:
Beyond this content, include the following sections (these don’t count toward the page limit):
The project reports will be published online for future students. If you have confidential material, e.g., work that you're planning to use for a publication or patent, please send me an email to withhold publication and duration. You can also request not publishing the report for other reasons. Please send me an email if you wish so.
Grading will consider technical contribution, completeness, i.e., a working system, and novelty, integration of systems design principles in analysis and design, experimental/analytical rigor, depth of discussion, literature survey, and organization.
Project presentation: The duration is 25 minutes including at least 5 minutes for questions. Grading will consider content (30%), delivery and presentation (30%), organization (20%), and response to questions (20%).
Review content (40%), delivery and presentation (30%), organization (20%), and discussion (10%).
Demonstrate overall understanding of the paper and the context (20%), identify main technical contributions (20%), identify critical limitations (20%), identify underlying systems design principles or implication, as appropriate (20%), significance and impact (10%), critique and discussion points (10%).
This includes both BCI data processing exercises and exposure to chip design tools. BCI data processing exercises are in MATLAB, and tutorial handouts will be provided. Training materials for CAD tools for chip design will be provided through video recordings or after-hour sessions. Team-based lab exercises test the ability to use industry-standard CAD tools to specify simple BCI processing designs in hardware description languages, simulate and verify them, synthesize into netlist, run physical design including layout, floor-planning, and timing, run design verification checks, and produce the final tapeout-ready file.
Regular presence (40%), responsiveness (30%), insightful engagement (30%).
Homeworks with written assignments due in class are to be submitted within the first ten minutes of the class start. Homeworks with problems that are to be submitted online (e.g., on Gradescope) or paper reviews (e.g., on Piazza), are due by specified times. Anything later is considered a late submission, and a letter grade or 5% of the points will be dropped per day. The project proposal, presentation and report cannot be late.
Exceptions include pre-arranged accommodations, unexpected emergencies including illness.
There is no required textbook, but readings are recommended from the following books as appropriate.
Note that most books on BCIs so far, if not all, have not aimed at the computer scientist or engineer. So, if the math in the signal processing, or the neuroscience is overwhelming, don’t worry. We will cover the basics in the class as needed for our purposes. Our goal is to understand at least enough math and neuroscience so that we can reason about how to best build computer systems for them (you're most welcome and even encouraged to go as further as you find intellectually stimulating). Additionally, unless specifically stated, the readings are best approached after the corresponding lecture, and are self-paced.
You can discuss your initial thoughts on the homeworks with other student in course, but each of you must answer separately. Do not use or read another students work before you turn yours in. You should not violate the university honor code. Violations are serious, and can end a career.
Please be respectful and maintain the sanctity of the classroom to ensure everyone has a positive experience with the course. For any student concerns, reach out to UNC Care, or to UNC Safe.
The Heels Care Network website is a place to access the many mental health resources available at Carolina. CAPS is the primary mental health provider for students, offering timely access to consultation and connection to clinically appropriate services. Go to the CAPS website or visit their facilities on the third floor of the Campus Health building for an initial evaluation to learn more. Students can also call CAPS 24/7 at 919-966-3658 for immediate assistance.
Any student who is impacted by discrimination, harassment, interpersonal (relationship) violence, sexual violence, sexual exploitation, or stalking is encouraged to seek resources on campus or in the community. Reports can be made online to the University Compliance Office (UCO), or by contacting the University’s Title IX Coordinator, or the Report and Response Managers in UCO. Please note that I am designated as a Responsible Employee, which means I must report to the UCO any information I receive about the forms of misconduct listed in this paragraph. If you’d like to speak with a confidential resource, those include Counseling and Psychological Services, the University’s Ombuds Office, and the Gender Violence Services Coordinators. Additional resources are available at safe.unc.edu.
Please contact the UCO Accommodations Team to forward any requests you have for this course to me.
You can use AI tools to understand concepts, find related work, but no course materials (lecture slides, homeworks etc.) and works of other students (reviews etc.) should be shared with those tools.
The instructor reserves the right to make changes to the syllabus including due dates, content, rubrics, and grading weightage and policies. These changes will be announced as early as possible.
Materials (including, but not limited to documents, slides, images, audio, and video) used in connection with this course may be subject to copyright protection. They are only for the use of students enrolled in this course, for purposes associated with this course, and may not be retained for longer than the class term. Unauthorized retention, duplication, distribution, commercialization, or modification of copyrighted materials is strictly prohibited by law. The materials cannot be shared outside of the course without my approval.
I developed all the materials of the course, sometimes utilizing or building on works of others for fair, educational use (appropriately credited and/or licensed when needed). Such works and their copyrights belong to respective owners. Usage of product or brand names, images, or trademarks are for identification and educational purposes only, and do not imply endorsement. If you believe your work has been used in the materials and requires attribution, please email.
| Date | Topic | Notes | Reading |
|---|---|---|---|
| 8/17/2026 | Introduction to BCIs | HW 1 released (Gradescope) Paper list released (Piazza) |
B1: Part 1, Part 2. Ch 1-3, 6 B2: Ch 1-5, 9-11, 19-23 |
| 8/19/2026 | BCI signal processing | Signal processing (SP) tutorial released |
B1: Ch 4, 5 B2: Ch 7, 8 Paper: Signal Processing for Brain–Computer Interfaces: A review and current perspectives |
| 8/24/2026 | Computer architecture and goals | HW 1 due |
B3: Ch 1 Paper: Hints and Principles for Computer System Design (Sections 1, 2, 3.1) |
| 8/26/2026 | Paper: An Integrated Brain-Machine Interface Platform with Thousands of Channels Paper: Power-saving Design Opportunities for Wireless Intracortical BCIs |
SP Lab 1 released |