Poster #29 - Examining hair artifact in MRI: A case study on coarse hair texture and increasing inclusivity in neuroimaging

• In Event: Poster Session 06

Abstract

Neuroimaging techniques, such as magnetic resonance imaging (MRI), have significantly contributed to the fields of developmental psychology and neuroscience. These methods not only aid in clinical diagnoses but also in further understanding of complex systems of the body (e.g., the nervous system) and consequently are highly relied on in academic and medical settings. Presently, approximately 80% of psychology study samples are comprised of participants from Western, educated, industrialized, rich and democratic (WEIRD) populations, despite the fact that that demographic is in the minority globally (12%) (Henrich et al., 2010). This history of oversampling WEIRD populations perpetuates biases in data practices (Nielsen et al., 2017). At a time when there is growing recognition that increasing sample diversity generates more reliable and representative data, psychology and neuroscience fields should move towards more diverse and inclusive samples and investigate possible biases in data acquisition, analysis, and interpretation.
Imaging artifacts are visual anomalies that interfere with the quality of MRI data, affecting the interpretation of results and often resulting in clinical misdiagnoses and exclusion of otherwise useable data (Krupa & Bekiesińska-Figatowska, 2015). Historically, professionals have attempted to improve data quality by informing participants of ways to reduce the artifacts, like the removal of hair ties or jewelry; however, practices to improve data quality not consider imaging-related artifacts that disproportionately affect systematically and historically excluded populations.
The present study is a case report of MRI artifacts in 7- and 9-year-old sisters which may represent previously understudied artifacts specific to Black populations. The MRI artifacts were previously identified as “unknown” and are depicted in Figures 1 and 2. Initial preprocessing using the fMRIPrep tool failed to remove distortion with fMRIPrep’s susceptibility distortion correction (SDC) feature with field maps applied for offline distortion correction of the BOLD images (i.e., images showed no improvement). Images were therefore run without field maps. However, the structural and functional runs for both siblings still presented significant artifacts declared “unexplained” or “possibly explained by technical difficulties.” In this report, we suggest that the artifacts are attributable to the sisters’ shared hairstyle, locs. Prior studies have reported atypical imaging artifacts found in Black participants corresponded to their coarse hair textures, hair products and hairstyles which produced hair-related imaging artifacts resulting in lower quality data (Koulakian et al., 2016; McKinstry & Jarrett, 2004). With this case study, we aim to call attention to how hair-related artifacts may disproportionately affect those of Afro-descent whose hair texture, styling and maintenance have not been historically accounted for in psychophysiological and neuroimaging technology. In an effort to aid scientists in being more cognizant of this systemic issue during data collection. Future work should identify and formalize presentations of hair-related artifacts that perpetuate the exclusion of Black participants in neuroscience research in order to identify methods to eliminate exclusionary practices and obtain high quality data for all individuals, regardless of demographic makeup.

Authors

• Maresa A. Tate, University of North Carolina at Chapel Hill
  Presenting Author

• Summer Motton, University of North Carolina at Chapel Hill

• Lucy Lurie, University of North Carolina at Chapel Hill

• Emily Hammer

• Amber Leinwand

• Meredith Gruhn, University of North Carolina at Chapel Hill

• Margaret A. Sheridan, University of North Carolina at Chapel Hill