The overarching goal of this R16 SuRE project is to optimize the validity of the Multidimensional Assessment of
Interoceptive Awareness (MAIA) for evaluating mind-body interventions. Interoception is defined as one's
ability to adaptively interpret and respond to signals from within the body. Many interoceptive responses are
autonomically driven; others are within the realm of subjective experience. The subjective facet is termed
`interoceptive sensibility' which is measured using self-report questionnaires—of which the MAIA is the most
comprehensive and broadly adopted. Interoceptive sensibility has wide-ranging implications on human health.
Deficits in interoceptive sensibility are a well-established transdiagnostic feature of eating disorders and
psychopathology in general. Better understanding this trait has implications for phenotyping eating disorders
and other mental health conditions. Targeting interoceptive sensibility in mind-body interventions is a growing
area of research. However, operationalizing this construct has been limited by measurement inconsistencies
and problems with validity. We recently conducted preliminary psychometrics and validity testing of the
MAIA's underlying common factor model (N=1294) and supported the original eight-dimension factor
structure. Our preliminary data also suggested measurement non-invariance across demographic groups and
potential unobserved population heterogeneity, which limit our ability to use the tool to evaluate interventions,
as intended. This proposal aims to address these concerns. We will collect survey data from a US Census-
matched sample of adults (N~2000) using an online platform. Our first aim will apply moderated nonlinear
factor analysis (MNLFA) to create person-specific factor scores that optimize the validity of using the MAIA to
make group comparisons, predict distal outcomes, and measure change. The MNLFA procedure identifies
which questions on the MAIA have differential item functioning across a group of relevant background and
psychosocial characteristics, including biological sex, age, body mass index, disordered eating, negative body
image, and traumatic stress. MNLFA adjusts the scoring algorithms for the factors of the scale to account for
population heterogeneity. Our second aim will apply latent profile analysis (LPA) to identify a typology of
interoceptive sensibility by clustering unobserved groups of people who tend to score similarly on the multiple
dimensions of the scale. LPA applies a person-centered approach instead of the more dominant variable-
centered approaches (such as confirmatory factor analysis) as an alternative dimension reduction technique.
The resulting typology will serve as prototypes of interoceptive sensibility with which we will generate and test
hypotheses about how the components of the MAIA interact to predict disordered eating and the associated
risk factors of negative body image and traumatic stress. In summary, one of the most impactful long-term
contributions of this project is to increase the rigor of the MAIA measurement model to expand the field's
capacity to evaluate the next generation of targeted, accessible, and broadly effective mind-body interventions.
