Decoded Neurofeedback Toward Bias and Racism Mitigation (2021-2022)


Racial bias is a prevailing problem that demands actions. From the tragic death of 17-year-old Trayvon Martin in 2012 to the 8 minute and 46 second kneeling on George Floyd in 2020, understanding why these events occurred is increasingly crucial in promoting social equality. Although racism is a multidimensional issue that is culturally and socially embedded, the root of it lies in categorical thinking, which distorts perceptions and causes ingroup bias. 

Ingroup bias is a pattern of behavior or mental process that favors members of one’s ingroup vs. outgroup. It has been identified as the leading cause of intergroup discrimination. As such, understanding the neural basis and mechanisms of ingroup bias is a crucial first step to address racial tension. 

While advanced neuroimaging technology has shed light on specific brain regions that mediate biases, no specific treatments are currently available to downregulate this complex phenomenon at the system neuroscience level. However, a novel neural intervention method, neurofeedback, which induces specific changes in brain activity patterns, has proven effective in mitigating obsessive compulsive disorder, autism spectrum disorder, depression and schizophrenia; subconsciously modulating confidence; and reducing phobia for common fears. This project aims to apply this proven line of neurofeedback technique to downregulate ingroup bias and alleviate racist tendency. 

Project Description

This project’s goals are to understand the neural basis of ingroup bias through advanced functional neuroimaging; validate the effectiveness of decoded neurofeedback on the downregulation of neural representation underlying ingroup bias; developing a proof-of-concept protocol for clinical applications; extend fMRI decoded neurofeedback methods to “modulate” racial biases; and raise public awareness that biases could be modulated.

The team will decode neural correlates of bias in individuals’ brains and combine it with neurofeedback to modulate such neural representation. The intervention efficacy will be interpreted with social theories toward refining our understanding of the interaction between social cohesion and individual’s cognition.

Neural decoder construction: Participants’ ingroup biases will be parameterized with the Implicit Association Test, which provides measurements of the strength of associations between concepts (e.g., Black people) and evaluations (e.g., good, bad). A series of tasks, such as drawing correlation of positive words with respective groups, will be carried out to induce ingroup bias. Team members will then acquire functional magnetic resonance images (fMRI) while the participants view images of faces that differ in racial features (Asian vs. African) inside the scanner. The team will obtain sets of brain patterns that represent a biased vs. an unbiased state, which will be used to train a neural pattern decoder to classify these two brain states for that individual.

Neurofeedback intervention: The goal of neurofeedback training is to foster “unbiased” brain activities by persistently associating such activities with external rewards. Visual stimuli (e.g., a disk) will be presented, and participants will be instructed to mentally maximize the size of a subsequently presented disk. The size of the disk will be controlled online by brain pattern similarity comparison with the two previously constructed decoders. When the brain patterns get more similar to the “unbiased” state, the disk will be enlarged, which leads to bigger monetary reward for the subject. After a prolonged period of training, the subject will be induced to prefer the ideal brain patterns that are associated with the “unbiased” state. 

Validation and wider application: The team will evaluate the efficacy of the neurofeedback by comparing implicit bias test scores before and after. Team members will refine the bias-reduction neuromodulation model, replicate it in larger samples and make necessary changes to attain a more long-lasting effect.

Learn more about this project team by viewing the team's video.

Anticipated Outputs

Decoded neurofeedback model applicable for reducing in-group bias; symposia at Duke University and Duke Kunshan University; using results for public education; journal publications


Summer 2021 – Summer 2022 

  • Summer 2021 (optional): Consolidate literature review; design and finalize experiment protocols; obtain Institutional Review Board ethics approval
  • Fall 2021: Conduct pilot experiments and preliminary data analysis; prepare for final experiments
  • Spring 2022: Conduct full experiment and control experiments (if any)
  • Summer 2022 (optional): Analyze data; write manuscript(s); disseminate findings (e.g., organizing symposia, submission of abstract to conferences, clinical settings)

This Team in the News

During Pandemic, Duke and Duke Kunshan Students Find a Home at the Other Campus

Neurofeedback Taking on In-group Bias

See related team, Decoded Neurofeedback Toward Bias and Racism Mitigation (2022-2023).


Image: fMRI Image of Preteen Brain, by NIH Image Gallery, licensed under CC BY-NC 2.0

fMRI Image of Preteen Brain.

Team Leaders

  • Sze chai Kwok, Duke Kunshan University
  • James Moody, Arts & Sciences-Sociology
  • Joseph Quinn, Arts and Sciences-Sociology-Ph.D. Student

/undergraduate Team Members

  • Anu Aggarwal
  • Alexandra Bayer
  • Morgan Biele, Neuroscience (BS)
  • Arunangshu Chakrabarty, Biomedical Engineering (BSE), Electrical & Cmputr Egr (BSE2)
  • Greta Cywinska, Neuroscience (BS)
  • Abigail Groth
  • Imani Hall
  • Dunhan Jiang
  • Jenny Li, Sociology (AB)
  • Yushi Li
  • Mackenzie Martinez, Biology (BS)
  • Catherine Mbata
  • Eliana Shapiro, Neuroscience (BS)
  • Reah Syed
  • Casey Szilagyi, Electrical & Computer Egr(BSE), Computer Science (BSE2)
  • Chenyu Wang
  • Sihan Wang
  • Zhou Xia
  • Anqi Xie

/yfaculty/staff Team Members

  • Yudian Cai, Duke Kunshan University

/zcommunity Team Members

  • Shanghai Key Laboratory of Magnetic Resonance
  • Advanced Telecommunications Research Institute International
  • Jianqi Li, East China Normal University