The UW Computational Neuroscience Center and the Shanahan Fellowship program have an exciting paid opportunity for pre-doctoral researchers at the interface of data and neuroscience. These roles are open to both current undergraduate and post-baccalaureate students.
Student fellows will be mentored by a current Shanahan Fellow, and they will gain hands-on research experience in neural computation, neural networks, and computational modeling/method development. Students will join a vibrant interdisciplinary research community with the opportunity to work with researchers at all levels at University of Washington and Allen Institute for Brain Science. As appropriate, fellows will be supported in opportunities to present and publish their work.
The fellowship will be full time during the summer quarter starting in June 2024, with opportunities to continue into the academic year. The salary range for this role is $20.19 – $26.98/hour and is eligible for benefits.
To apply, please send your CV and cover letter to cncadmin@uw.edu. Please include a description of your career goals, research interests, and which projects interest you in your cover letter. Applications will be reviewed on a rolling basis.
Current Fellows and Project Descriptions
Sven Dorkenwald
Bio:
Sven Dorkenwald joined the Allen Institute and the University of Washington as a Shanahan Fellow in September 2023. He received his undergraduate degree in Physics and a Masters degree in Computer Engineering at the University of Heidelberg in Germany. While in Heidelberg, he worked on automated image analysis in connectomics with Jörgen Kornfeld in the department of Winfried Denk at the Max Planck Institute for Medical Research. Sven received his Ph.D. in Computer Science and Neuroscience from Princeton University, where he worked with Sebastian Seung and Mala Murthy. During his PhD, he developed approaches for the reconstruction and analysis of neuronal circuits from Electron Microscopy images and spearheaded the FlyWire consortium effort that produced a synapse-resolution connectome of an adult Drosophila brain. As a student researcher at Google Researcher, Sven devised a self-supervised machine learning approach for efficient annotation of cell reconstructions. As a Shanahan Fellow, Sven is working on the analysis of new and emerging connectomics datasets while pursuing new ways to integrate them with neuroscience datasets from other domains to enable multi-modal analyses.
Project Description
Neurons in the mammalian cortex are summarized by a vast diversity of cell types, each with characteristic anatomical, molecular and functional properties. Neurons of each type exhibit distinct connectivity rules but studying them has been challenging due to difficulties in identifying those synaptic connections. Recent advances in the imaging and analysis of neuronal tissue now enable the reconstruction of synaptic wiring diagrams of entire insect brains and large chunks of mammalian cortices. These wiring diagrams reveal unprecedented insights into the organization of neuronal circuits and provide new opportunities to study the brain at a synaptic resolution.
We are seeking a research assistant for one year to pursue a project aimed at untangling connectivity patterns using cutting-edge connectomics datasets. This project is at the interface of neuroscience and data science. It is suitable for candidates that either have a strong background in neuroscience and are eager to advance their data science skills, or candidates with a strong data science background that want to enter into neuroscience, and everyone in between. The proportions of the data science and neuroscience aspects will be tailored to the interests of the candidate.
A neuroscience-focused version of this project might study the input distributions onto excitatory cells from various cell types (see related studies below). Existing predictions of synapse types would further enable the study of the organization of inhibitory and excitatory inputs on individual dendrites.
Data science tools, and machine learning in particular, are indispensable for the analysis of connectomics datasets because of their size and complexity. A data science-focused project could develop new machine learning models that accelerate proofreading of connectomics datasets (see FlyWire paper on proofreading below) or annotate cell structures (see SegCLR paper below) to enable new analyses.
Required
- Proficiency in Python. i.e, comfortable working with datasets using pandas, NumPy, and performing data visualization with libraries such as Matplotlib and Seaborn. Understanding of object-oriented programming principles in Python
- Background in neuroscience and/or data science
- Willingness to learn and excitement about the project
Examples of connectomics datasets that can be used in this project:
- https://www.microns-explorer.org/cortical-mm3
- https://www.microns-explorer.org/phase1
- https://flywire.ai/
Relevant publications:
- MICrONS dataset: https://www.biorxiv.org/content/10.1101/2021.07.28.454025v2
- Whole-brain connectome of the fruit fly: https://www.biorxiv.org/content/10.1101/2023.06.27.546656v2
- SegCLR: https://www.nature.com/articles/s41592-023-02059-8
- FlyWire proofreading: https://www.nature.com/articles/s41592-021-01330-0
- Study of cell type specific connectivity motifs: https://www.biorxiv.org/content/10.1101/2023.01.23.525290v1
Yiliu Wang
Bio:
Yiliu Wang joined the Allen Institute in 2023 as a Shanahan Foundation Fellow and a member of the modeling, analysis and theory team. She received a Master’s degree in Mathematics and Statistics from the University of Oxford. She earned her PhD in Statistics from the London School of Economics (LSE), supervised by Prof. Milan Vojnovic. During her PhD research, she studied inference, optimization, approximation and selection problems in relational learning, a subdivision of high-dimensional statistics. Specifically, she developed a stochastic valuation function system to evaluate sets of items and learn about individual representations through group outcomes. In her current role, Yiliu is looking forward to leveraging her expertise in mathematics and statistics to bring methodological advances in neuroscience and bridge the gap between data science and neuroscience.
Project Description:
Cell types are important neuronal properties that contribute to different cell functions and connectivities. Through the Allen Brain Cell Atlas, single cell data from human and mouse is acquired, which enables researchers to identify cell types with their spatial locations and gene expressions. It is of interest to understand the spatial distribution of cell types and how cell types interact with each other. It also remains an open question to unify a complete cell type notion between human and mouse. Furthermore, the Cell Atlas provides accurate spatial annotation and colocalization of cell types in the middle temporal gyrus of donors from the SEA-AD aged cohort that span the spectrum of Alzheimer’s disease. It is not well understood how the cell type distribution changes as the disease progresses. It is thus of interest to study the change in cell type neighborhood and explore the underlying mechanisms of the disease.
Depending on the interests and skill sets of the selected applicant, the project can be tailored to focus more on the theoretical or applied side.
Skills and qualifications:
Required: Bachelor’s degree or current student in science or computational related degree. Interests in interdisciplinary research. Experience in computational biology is not required.
Preferred:
- Knowledge in statistical inference and analysis.
- Programming proficiency in R or Python, with experience in data manipulation and machine learning frameworks.
- Ability to work both independently and in a collaborative, multi-disciplinary environment.
Iris Stone
Bio:
Iris Stone joined the Allen Institute and the University of Washington as a Shanahan Foundation Fellow in October 2023. She received her Ph.D. in Quantitative and Computational Neuroscience and a graduate certificate in Statistics and Machine Learning from Princeton University. During her graduate studies, she worked with Prof. Jonathan Pillow and Prof. Ilana Witten to develop latent variable models for characterizing the dynamic structure underlying complex behaviors in mice, such as decision-making and exploration. Her past experience also includes an internship at a biotech startup focused on using cerebral organoids to develop treatments for neurological diseases, as well as a B.S. in Physics from George Mason University, where she researched the use of organic and nanomaterials in biomedical applications. Iris is enthusiastic about working on open, team-oriented science and supporting the Allen Institute’s mission of unlocking treatments and cures for human health
Project description:
Identifying neural signatures of learning during a mouse foraging task using computational models of behavior and neuromodulator activity
What choices should we make to maximize the possible reward (e.g., coffee consumption) that we obtain? This can be a tricky question to answer, as in many environments there is uncertainty about the “rules” that govern how rewards are distributed (Where are the best coffee shops? Which one is easiest to access? Is low effort or high quality more important? Do the factors we care about change over time?). All of these choices are broadly lumped under the rubric of “foraging behavior”. But to become expert coffee foragers, we must first learn how to forage, slowly picking up on the things that matter most as we engage with the coffee-drinking world. Exactly what is happening in the brain during this process is an open question. In this project, we will investigate a rich fiber photometry dataset containing recordings of four different neuromodulator concentrations (dopamine, serotonin, norepinephrine, acetylcholine) at various location in the brain of a mouse while it is foraging, with a focus on how the animal learns different tasks (no coffee for the mice, but they do enjoy a good drink of water!). As this is a new dataset, work will likely include plenty of fundamental analyses of the raw data and an open exploration process to select and apply relevant computational models that can identify connections between the animals’ learning behavior and neuromodulator activity. The ideal candidate will be open-minded about exploring modeling from multiple angles and has an interest in fundamental data analysis.
Day-to-day experience
The ideal research assistant will work semi-independently. On the one hand, work will be very collaborative when it comes to big-picture thinking, with regular meetings with me (and potentially other Allen Institute scientists) on topics like project ideas, literature reviews, progress tracking, troubleshooting, and interpretation of results. However, smaller day-to-day tasks such as writing and executing code, analyzing data, and learning new skills as they become necessary will be done somewhat independently (of course with opportunities for some help when needed). This will require the RA to be a self-starter and have a sense of self-direction in order to make progress and manage their time well.
Skills and qualifications
Applicants looking to expand their interdisciplinary expertise are encouraged to apply.
Required: 1 year experience in python (writing and/or using code), some neuroscience background, relevant math courses (calculus, linear algebra, and/or probability and statistics), and especially a willingness to learn and excitement about the project
Preferred: 2+ years experience writing python code; direct research experience working with neural and/or mouse behavioral data; advanced math courses (higher-level calculus, machine learning, computational neuroscience).
Start date is flexible (June to August). There is also flexibility for candidates to work either full- or part-time.