A sinister subject: Quantifying handedness‐based recruitment biases in current neuroimaging research

• Published in: The European journal of neuroscience Vol. 51; no. 7; pp. 1642 - 1656
• Main Authors: Bailey, Lyam M., McMillan, Laura E., Newman, Aaron J.
• Format: Journal Article
• Language: English
• Published: France Wiley Subscription Services, Inc 01.04.2020
• Subjects: adextral; Bias; generalisability; Handedness; left handed; Medical imaging; Neuroimaging; recruitment practices; generalisability; adextral; recruitment practices; left handed
• ISSN: 0953-816X; 1460-9568; 1460-9568
• DOI: 10.1111/ejn.14542

Approximately ten per cent of humans are left‐handed or ambidextrous (adextral). It has been suggested that, despite their sizable representation at the whole‐population level, this demographic is largely avoided by researchers within the neuroimaging community. To date, however, no formal effort has been made to quantify the extent to which adextrals are excluded from neuroimaging‐based research. Here, we aimed to address this question in a review of over 1,000 recent articles published in high‐impact, peer‐reviewed, neuroimaging‐focused journals. Specifically, we sought to ascertain whether, and the extent to which adextrals are underrepresented in neuroimaging study samples, and to delineate potential trends in this bias. Handedness data were available for over 30,000 research subjects; only around 3%–4% of these individuals were adextral—considerably less than the 10% benchmark one would expect if neuroimaging samples were truly representative of the general population. This observation was generally consistent across different areas of research, but was modulated by the demographic characteristics of neuroimaging participants. The epistemological and ethical implications of these findings are discussed. One in ten individuals are left‐handed or ambidextrous. We sought to assess whether modern neuroimaging studies are representative of this population. A review of over 1,000 high‐impact studies revealed that around 97% of research participants are right‐handed, suggesting a pervasive handedness bias in contemporary research. This bias was consistent across different imaging techniques and research areas, but was less prevalent in studies recruiting from certain demographic populations.