The Computational Visual Neuroscience Laboratory (cvnlab) is located at the Center for Magnetic Resonance Research (CMRR) at the University of Minnesota. We are a highly collaborative lab that seeks to integrate broad interdisciplinary insights to understand brain function. If you are interested in collaborating or working in the cvnlab, contact Kendrick and also read the cvnlab advising/mentoring/collaboration statement.

Fundamental questions of interest to cvnlab include:

• Can we develop better models of information processing in visual cortex?
• Can we exploit data-driven approaches to gain insight into brain function?
  
• Can we build an integrated understanding of how diverse stimuli and tasks are processed in the visual system?
• Can we design techniques to improve the quality and quantity of information extracted from fMRI?
• Can we expand the spatiotemporal limits of fMRI?
  

Research in the cvnlab is vertically integrated --- spanning theory, modeling, data, and analysis --- and lies at intersection of three fields:

• Cognitive neuroscience: We are interested in the representation and processing of visual images by the brain. We also study the anatomical and functional topography of human cortex.
• Functional MRI methods: We specialize in analysis methods for fMRI data. This includes advanced statistical and analysis methods (encoding, decoding, multivariate analyses) as well as methods for improving the quality of fMRI data (e.g. denoising methods). We are currently interested in characterizing and assessing ultra-high-resolution fMRI measurements and developing associated analysis methods.
• Computational neuroscience: We use experimental data to develop models of neural information processing. We build functional (or computational) models of information processing in both low- and high-level cortex. We are also interested in the principles of modeling and associated philosophical issues.
  

The cvnlab likes high-quality data, clear and interpretable analyses, and quantitative models. We are interested in understanding the limitations of current methodology and developing ways to overcome these limitations. In the spirit of reproducible research, we make freely available software tools and data resources developed in the course of our research.

Brain art

News

February 2024

• Pre-print Recent statistics shift object representations in parahippocampal cortex now on bioRxiv.

January 2024

• Manuscript Driving and suppressing the human language network using large language models published in Nature Human Behaviour.
• Manuscript Better models of human high-level visual cortex emerge from natural language supervision with a large and diverse dataset published in Nature Machine Intelligence.
• Manuscript Re-expression of CA1 and entorhinal activity patterns preserves temporal context memory at long timescales published in Nature Communications.

July 2023

• Pre-print What can 1.8 billion regressions tell us about the pressures shaping high-level visual representation in brains and machines? now on bioRxiv.
  

June 2023

• Manuscript Brain-optimized deep neural network models of human visual areas learn non-hierarchical representations published in Nature Communications.

April 2023

• Pre-print Driving and suppressing the human language network using large language models now on bioRxiv.
• Manuscript Tasks and their role in visual neuroscience published in Neuron.

March 2023

• Pre-print BOLD Moments: A video fMRI dataset for modeling short visual events now on bioRxiv.
• Manuscript Engaging in word recognition elicits highly specific modulations in visual cortex published in Current Biology.

 T2*-weighted functional image, GE-EPI, 0.8-mm isotropic, 2.2-s TR (click for full image)