Anorexia nervosa and diffusion weighted imaging: An open methodological question raised by a systematic review and a fractional anisotropy anatomical likelihood estimation meta-analysis.

Abstract

Anorexia nervosa (AN) is characterized by white matter abnormalities in neuroimaging studies. Fractional anisotropy (FA) is a diffusion tensor imaging (DTI) index that is considered an instrument for the evaluation of white matter alterations. However, the literature has recently pointed out the role of the partial volume effect (PVE) as a confounding factor for the identification of juxtaposed tissues. Our goal was to review the DTI literature in AN and evaluate possible confounding factors linked to the reported results. A systematic review of the literature was conducted to identify Diffusion Tensor Imaging studies of individuals with AN and, subsequently, an anatomical likelihood estimation (ALE) meta-analysis was performed on studies published before March 18, 2019. Twenty-four studies (AN = 517, controls = 542) were included in the qualitative systematic review of the literature. Ten published studies underwent the ALE-analysis (AN = 210, controls = 229), plus data from an unpublished cohort (AN = 38, controls = 38). Two clusters of decreased FA were identified, namely in the left corona radiata, and in the left thalamus. Only one article took the PVE correction analysis into account. The alterations identified must be considered within the limits of a possible methodological bias regarding PVE and free water and re-analysis of the data may be recommended. The preliminary data showed that the alteration of white matter pathways between the limbic structures and brain cortex may be linked to the processing of somatosensory information that could play a key role in the psychopathology of the disorder.