Information for prospective students#
Prospective graduate students should read these pages, as well as look through the group’s website to learn about current directions of our research, and see where members of the lab who have graduated are currently working.
Which program should I apply to?#
The lab can take students through one of the following programs:
The Psychology Graduate Program. In particular, prospective students interest in joining the group should apply to the Cognition and Perception area and list Ariel among the faculty they are interested in working with.
The Graduate Program at the Allen School for Computer Science & Engineering.
Some FAQs:#
How does the lab handle flexibility in terms of work hours and location? For example, are there expectations for regular in-person lab presence, or is there flexibility for remote collaboration, particularly for computationally focused projects?
There are no specific requirements for in-office hours. I understand that different people have different needs and get their work done at different times/conditions. Everyone is expected to attend weekly group meetings and weekly journal clubs / code reviews (and to lead these occasionally). Some of these meetings, with a larger group of collaborators is always done remotely on zoom. In addition, I expect a weekly one-on-one meeting with me, to discuss progress on projects and work together. That said, I do expect people to revisit their goals and progress on a regular basis – typically quarterly – and see whether and what needs adjustment. That should help folks make progress towards their goals, and also prevent people from spinning their wheels. See also this page.
What is the office setup like?
Our group has its space in the Center for Human Neuroscience (established in 2021) that is on the 5th floor of Kincaid Hall. We are co-located with several other Cognition/Perception groups here. The floor has a lot of shared spaces: a break room and kitchen, as well as a big meeting room and some spaces for co-working. There’s plenty of desk space for students/postdocs/RAs and the Cog/Per faculty all have their offices within the same space, so there are a lot of opportunities for interaction, and we often hold joint group meetings.
I’ve heard of some labs where publishing anywhere other than Nature/Science is discouraged. Is this the case in your lab?
I definitely don’t discourage anyone from publishing in Nature/Science… I understand and can generally agree with a philosophy that encourages substantial and broadly impactful work (the kind that could be published in those journals). But I also think that there is a lot to be said for work that is specialized and more detailed than the kind of work that gets published in these journals. A lot of our work is rather technical in nature and is much more suitable for more specialized journals, and we will almost certainly continue to do work that gets published in such journals (and that we’ll be proud of). I also don’t think there is anything wrong with work that is “incremental”. One day maybe we’ll publish a Nature/Science article, but that hasn’t happened yet (see the group’s publications page) to see where and what we have published.
Do you have a specific number of papers you expect students to publish?
In the best case, I think that a PhD needs to be substantial and coherent. It’s not so much about a particular number of papers, but more about making a significant and consistent set of contributions. I think that a successful PhD project in neuroinformatics would encompass methodological and technical contributions (perhaps a conference paper in a machine learning or scientific computing conference), as well as their application to data (that’s the Science/Nature paper, I guess?) and possibly a followup that tests a model presented by the original application in a different context or dataset (another Science/Nature paper right there). That would probably result in approximately 3 lead-author papers. I’d also encourage students to collaborate with others and contribute as non-lead authors on another couple of papers, either within the group or with other groups (we collaborate a lot). I think that one outcome of this kind of trajectory is the development of \(\pi\)-shaped people (not literally! see this post by Jake Vanderplas for some more details). So, the goal is not so much about the number of papers per se (though publishing is definitely a good thing!), but more about the kind of training that this kind of PhD would result in.
How do you support students who aren’t making progress on their goals?
I think that the main thing to do in these situations is to figure out why they are not making progress. If it is because they need more feedback (which is often the case), I try to figure out mechanisms to touch base on a more regular basis (more than the weekly one-on-one) and to provide the needed feedback. Sometimes it’s about the project itself, in which case it’s worth rethinking how to approach the questions we’re interested in answering. If motivation is flagging, it is worth thinking about how they can renew their motivation. For some people, having more than one project that they are working on is useful, because it allows them to switch between things, and the probability of multiple projects being stuck at the same time is low. For others, really focusing in on one thing is the right thing. It’s also good to find a balance between work that is collaborative with others and work that is purely on one person’s shoulders and where they can run ahead. For some people, it’s good to zoom out and to evaluate their long-term goals on a regular basis, and to think strategically about how what they are currently doing leads to where they want to go. Without being too paternalistic, I think that it’s important to make sure that students are taking good care of themselves and finding balance in their lives. It’s easy to think that you will make more progress by digging in, but often what you really need is a weekend of walking in a forest, or doing something else you enjoy (other than research, that is).
What is the general exam process like at UW?
The “General Exam” is typically taken in the third year. It has to cover both depth on the student’s research topic, as well as breadth of literature from related fields. There is a written portion, which can take one of several forms: a review paper, a grant proposal, and/or answers to a set of questions created by the committee, and an oral portion, which is basically a discussion with the committee based on the written portion. From what I have seen, the written and oral portions are usually taken close in time, while material from the written portion is still fresh in the student’s mind. I think that it’s a good opportunity to step away from the nitty-gritty of the work the student is doing, get some perspective, and get well acquainted with relevant literature. It’s also an opportunity to soak up some knowledge from committee members other than the advisor. There is a lot more information about the program in the program manual (see page 21 for information about the general exam, and page 4 for a table of all the program milestones).
Where have your former students gone after completing their PhDs, and how do you guide them in building careers in computational neuroscience, either in academia or industry?
Lab alumni (together with their current position, if known) are listed on this page. As you can see there, career trajectories vary with some academic researchers and faculty and some who have gone to industry. I believe (hope?) that the training that folks can get in the lab serves as a good preparation for either kind of career (and maybe others?). I think that the core of the training in the lab is in solid engineering, with an appetite to innovative methods and technologies, and I think that translates pretty to doing rigorous scientific work in a range of environments. To facilitate more choice, I tend to also encourage trainees to seek out opportunities to experience other kinds of environments, e.g., through industry internships.
With ongoing interdisciplinary work, how do you approach integrating techniques from other fields, such as computational neuroscience or data science, into your lab’s projects? What skills do you see as most valuable for students contributing to these collaborations?
Because the work is so interdisciplinary, it is abundantly clear that people from a variety of different backgrounds can contribute, so long as they are willing to make the (sometimes substantial) effort to learn things that are outside of their comfort zone. For example, a person with a strong background in statistics and/or computing will have a lot to learn about brain structure and function and about the kinds of questions we can ask with human neuroscience data. A person with a strong background in biology will have to learn some software engineering and statistics. Overall, because the core of our work is neuroinformatics, across all projects we highly prioritize software engineering skills. This does mean a degree in computer science or in engineering (I don’t have either of these myself…), but the equivalent of “bench skills” in our work is the ability to write computer software that is robust, functional and does what it purports to do.