PhD Studentship: Disentangling and modelling behaviourally-relevant visual and semantic dimensions of visual cognition in the human brain (Kamila Maria Jozwik lab, the University of Cambridge)

Start date: October 2026

Project:

Seeing the world allows most of us to experience the surrounding reality. To interact with the world, we must recognise what is in our environment. The proposed research aims to define the dimensions - here, characteristics of objects (”curved”, ”pink”, ”having eyes”, “being animate”, ”having agency”, or ones that are hard to name) that are crucial for our everyday functioning. This research is only possible now thanks to 1) advances in imaging techniques that fuel a more detailed understanding of the brain, 2) tools from artificial intelligence that enable building better computer simulations of the brain. 

The lab will leverage these advancements to disentangle and model behaviourally-relevant visual and semantic dimensions of visual cognition in the human brain, while increasing the ecological validity of experiments (including mobile EEG and immersive technologies), in the light of three aims:

1) characterise behaviourally-relevant visual and semantic dimensions of visual cognition by the use of large-scale brain imaging datasets of responses to images and model these representations with AI models (deep neural networks (including topographical), multimodal models, LLMs),

2) define and model dimensions related to the perception of animacy when interacting with objects and people using videos (behaviour, fMRI, and MEG),

3) determine to what extent these brain representations and dimensions change when humans are immersed in the environment (mobile EEG and VR).

The PhD student is welcome to work on one (or more) of the three aspects of the research programme.

The lab studies visuo-semantic cognition combining cognitive science, neuroscience, and computational modelling. We combine experimental behavioural tasks, brain imaging (fMRI and M/EEG), and, through collaborations, macaque electrophysiology. We use machine learning techniques for data analysis and computational modelling with a special interest in biologically-inspired deep learning and AI models (NeuroAI). Please find prior work here: (Google Scholar: https://scholar.google.com/citations?hl=en&user=oEifmSgAAAAJ&view_op=list_works&sortby=pubdate). We also started exploring characterization of visual, semantic, and affective dimensions in mental health conditions and modelling neurotechnology interventions. I would be happy to discuss PhD projects related to these research spaces as well. Please contact me via email before applying using the link below.

Ideal candidate:

• extensive experience in programming in Python or Matlab and data analysis (essential, beyond introductory coursework)
• experience with neuroimaging data design/collection/analysis (fMRI, M/EEG) and behavioural data collection and analysis
• experience in machine learning and AI
• highly motivated and curious

Contact:

Kamila Maria Jozwik (Royal Society University Research Fellow and an Assistant Research Professor)

jozwik.kamila@gmail.com or kj287@cam.ac.uk

Please send in the initial email:

• your CV
• information about the relevant research experience
• journal and conference publications, preprints (if applicable), and research thesis
• please also ask 2-3 of your referees, ideally with whom you worked on research projects, to send me their reference letters (I understand I may not receive all of them before the application deadline)

This is a 4-year, fully funded position for UK home students. However, international students are welcome to apply and secure funding through other studentships. I will also be recruiting a postdoc in the similar research space soon, in case you know researchers who may be interested in applying.

Advert information