Applications will be reviewed on a rolling basis.
We are seeking postdoctoral researchers, graduate or undergraduate student researchers for joint positions across UC Berkeley (our lab and the Jagust Lab), the UCSF Memory and Aging Center (NeuroAI Lab), and Columbia University.
Overview and Research. Our work spans multi-cohort modeling of disease pathways, network-level modeling of brain function, AI-assisted visual interpretation of structural MRI and amyloid/tau PET, multimodal modeling of electrophysiology and fMRI, and ML/AI methods for integrating longitudinal imaging, biomarker, and clinical data. Our current interests aim to:
- Understand how genetic risk, Alzheimer’s pathology, brain function, longitudinal cognitive trajectories, lifestyle, and physiological measures relate to one another, and how they can be modeled jointly rather than in isolation.
- Develop predictive models of individual cognitive and disease trajectories by integrating imaging, biomarkers, genetics, clinical information, and longitudinal assessments.
- Develop cross-modal models that map electrophysiological signals, including EEG and intracranial recordings, to fMRI/BOLD dynamics, using geometric brain representations and modern multimodal learning to recover underlying neural activity.
- Study how to reduce these models to sparse, potentially wearable EEG configurations while preserving accurate reconstruction of brain dynamics, with validation using paired electrophysiology–fMRI data.
Several distinct sub-projects sit under this umbrella, with the balance shaped around your strengths and interests. You would be joining an active group with regular access to mentorship across imaging, statistics, AI, and clinical neurology.
Data. We work with deeply phenotyped cohorts and electronic health records (EHRs) combining task-based and resting-state fMRI, amyloid and tau PET, structural MRI, longitudinal neuropsychological assessment, and genetics. Established collaborations and co-investigators at the UC Berkeley, UCSF Memory and Aging Center and Columbia University allow findings to be validated in independent clinical and neuropathological samples.
What we are looking for. We are interested in people with demonstrated ability in two or more of the following:
- fMRI analysis — fMRIPrep, connectomics, dynamic graph approaches, dynamic causal modeling, task and resting-state designs
- PET imaging — familiarity with FreeSurfer and standard PET processing, and comfort working from raw images
- Alzheimer's disease biology — the underlying pathology and the biomarker literature
- Machine learning/AI — convolutional networks for image classification, vision-language models
- Multivariate and longitudinal statistics — path and mediation modeling, and the analysis of repeated-measures data pooled across cohorts
We do not expect anyone to have all of these. If you bring expertise in at least two of these areas along with an enthusiasm to learn and grow into the others, we would love to hear from you!
Start date: Flexible.
Questions: jingshenwang@berkeley.edu