Use of Ketosis in Modulating Metabolic Pathways in Bipolar Disorder

Primary outcomes

1. Stabilization of brain networks (general brain functioning)

   Time frame: Within a month of enrollment completion

   Baseline network stability will be measured using resting-state fMRI. Brain network stability (unitless metric) is a biomarker derived from fMRI scan activity that quantifies the degree to which regions that are active together at one time point continue to remain active together throughout the scan. It has been shown by prior work to be a biomarker sensitive to both aging and metabolic effects and is thus a primary measure as this study examines its validity in the context of bipolar disorder. Given this prior work, it is anticipated that relative to comparison subjects, individuals with bipolar disorder will show greater network instability (increased instability score - unitless) consistent with metabolic dysregulation.

2. Relative stabilization or destabilization of brain networks in response to metabolic bolus

   Time frame: Within a month of enrollment completion

   Using the resting-state fMRI acquired after either the glucose or ketone bolus, brain network stability will be quantified as described above and compared to the baseline outcome. Based on prior work using this metric, it is anticipated that in both healthy individuals and individuals with bipolar disorder will exhibit network stabilization in the presence of ketones (decreased instability score - unitless) and network destabilization in the presence of glucose (increased instability score - unitless), consistent with the effects seen in metabolic regulation in an aging population.

3. Prefrontal-limbic circuit regulation

   Time frame: Within a month of enrollment completion

   Using task fMRI data for learning and matching tasks, the BOLD signal will be measured for the pre-frontal limbic circuit (composed of the ventromedial prefrontal cortex, orbitofrontal cortex, hippocampus, amygdala, and thalamus). The relative signal correlations between these regions will be used to determine signal lag (measured in seconds) as a metric of circuit regulation. Prior work has shown this metric to be sensitive to changes in emotional regulation (specifically in generalized anxiety disorder), and thus it is hypothesized to provide a sensitive marker in comparing individuals with bipolar disorder to healthy individuals as well.

4. Cortico-striatal circuit regulation

   Time frame: Within a month of enrollment completion

   Using task fMRI data for learning and matching tasks, the BOLD signal will be measured for the cortico-striatal circuit (composed of the prefrontal cortex, striatum, thalamus, globus pallidus, subthalamic nucleus, and substantia nigra). The relative signal correlations between these regions will be used to determine signal lag (measured in seconds) as a metric of circuit regulation. Prior work has shown this circuit to be important in learning tasks like the ones used in this study, and thus provides an excellent control circuit for comparison to the pre-frontal limbic metrics above.

5. Concentration of neurometabolites measured by Magnetic Resonance Spectroscopy (MRS)

   Time frame: Within a month of enrollment completion

   Using 7T MRS allows us to sensitively measure the concentrations of several neurometabolites sensitively and simultaneously. The following metabolite concentrations (in mmol) will be quantified both before and after the energy bolus consumption: Neural glucose and D-βHB (ketone), Taurine, Lactate, Ascorbate, Phosphocreatine, Aspartate, Phosphoethanolamine, Gamma-Aminobutyric Acid (GABA), Scyllo-Inositol, Myo-Inositol, Phosphocholine and Glycerophosphocholine, Glutathione, N-Acetylaspartate, Creatine and Phosphocreatine, N-Acetylaspartate and N-Acetylaspartylglutamate, Glutamate, Glutamine, N-Acetylaspartylglutamate