WIP Special

Identifying the neuroimaging fingerprint of hippocampal insulin resistance in depression: a pilot study

Presented by Antoine Cherix

Abstract: Depression is a leading cause of disability worldwide and a major public health issue but little is known about the etiological factors and its underlying pathophysiology. Recently, metabolic syndrome (MetS) and insulin resistance (IR) in particular, have been proposed as potential causes of a subset of depression due to their high co-morbidity. The hippocampus is a well-recognised insulin-sensitive tissue which might be key in this pathophysiological process, but tools are lacking to assess insulin signalling efficiency non-invasively in the brain. Disrupting brain insulin signalling in mice allowed us to identify a neuroimaging fingerprint (using ¹H, ³¹P and ¹³C-MRS) of insulin resistance in preclinical models of depression. We observed that a reduction of hippocampal lactate (end-product of glycolysis) and reduction of excitatory/inhibitory (E/I) balance (Glu/GABA) was linked with animal’s depressive-like behaviour.

We are now investigating whether a similar hippocampal ‘trace’ can be identified in a pilot group of participants with depression and known disruption of insulin signalling. To do so, we will conduct a cross-sectional pilot study using edited MRS at 7 Tesla in the hippocampus of people with depression and controls. We will test whether low levels of lactate and Glu/GABA can be detected in people with depression and whether they correlate with ‘systemic’ insulin resistance (HbA1C) and/or with specific cognitive/memory functions.

Overall, we anticipate that our findings will provide preliminary data to capture the role of IR in depression and bring forward a useful imaging fingerprint with high clinical potential in the field of psychiatry.

Learning Reward Distributions

Presented by Jan Grohn

Abstract: Recent work has shown that neural activity of tracks the full distribution of reward rather than just some summary statistics of it such as the mean and variance. We have devised tasks that allow us to reconstruct these learned distributions from behaviour and plan on linking this behaviour to 7T fMRI data.