How reliable are the results from functional magnetic resonance imaging?

Abstract

Functional magnetic resonance imaging (fMRI) is one of the most important methods for in vivo investigation of cognitive processes in the human brain. Within the last two decades, an explosion of research has emerged using fMRI, revealing the underpinnings of everything from motor and sensory processes to the foundations of social cognition. While these results have revealed the potential of neuroimaging, important questions regarding the reliability of these results remain unanswered. In this paper, we take a close look at what is currently known about the reliability of fMRI findings. First, we examine the many factors that influence the quality of acquired fMRI data. We also conduct a review of the existing literature to determine if some measure of agreement has emerged regarding the reliability of fMRI. Finally, we provide commentary on ways to improve fMRI reliability and what questions remain unanswered. Reliability is the foundation on which scientific investigation is based. How reliable are the results from fMRI?

Apport des « neurosciences de la décision » à l’étude des comportements alimentaires et de l’obésité

“…D'autre part, les résultats des études expérimentales ne sont pas nécessairement robustes, c'est-à-dire facilement reproductibles. Une publication récente [4] a effectivement montré que la reproductibilité des résultats d'expérimentations en IRMf n'excédait pas 50 %, même dans les cas où les mêmes tâches et les mêmes stimulus étaient présentés au même groupe de participants ! 1 1 Notons toutefois que la variabilité due à la technique d'imagerie cérébrale, aux caractéris-tiques de l'instrument de mesure et au traitement des données ne sont pas les seules à entrer en compte dans ce cas là.…”

Neuroimaging in Psychiatric Pharmacogenetics Research: The Promise and Pitfalls

“…Within fMRI, for example, there is tremendous variability in experimental paradigms and analytic approaches that contributes to false positive results (Carpentier et al, 2012). The issue of test-retest reliability is another consideration; some cognitive fMRI measures (such as the n-back working memory task) provide very good within-subject reproducibility, whereas others (such as the facial matching tasks) are more variable (Bennett and Miller, 2010;Plichta et al, 2012). Further, comorbidities in psychiatric populations can pose a significant challenge, and variability in screening criteria across studies may limit the reproducibility of findings.…”

Food-induced brain responses and eating behaviour

“…Basic measures could be, e.g., relevant subject characteristics, beyond already common ones such as sex, age, and BMI. Furthermore, for neuroimaging experiments standardisation of data acquisition, pre-processing and analysis would help to increase consistency (100) . In addition, it would be helpful if the same (validated) fMRI tasks or the same food stimuli were used for measuring responses to food stimuli or other relevant brain responses such as those related to reward sensitivity or impulsivity.…”