Fig. 1: Human hypothalamic and corticolimbic activity is correlated with interstitial glucose dynamics in a diurnally mediated fashion.

A Clinical recording system comprising stereo-encephalography (sEEG) and continuous glucose monitoring (Dexcom G6) in patients undergoing invasive monitoring. This illustration was created with BioRender.com B Anatomical location of electrodes in the bilateral hypothalamus of Subject 1. C Interstitial glucose levels as a function of time over six recording days in Subject 1 with corresponding mean high-frequency activity (HFA; 70–170 Hz) from the hypothalamic electrodes. The black lines denote periods of sleep while the blue lines denote start and end of ad libitum meals. D Scatter plots of hypothalamic HFA and interstitial glucose variations with 0-h (top) and −2.8 h lags (bottom). Pearson’s correlation (R) and associated P-value are shown. N = 1588 datapoints. E Mean cross-correlogram between hypothalamic HFA and interstitial glucose variations. The error bar indicates the standard error. Cross-correlograms using shuffled glucose data with 2 standard deviations is shown in gray. N = 1588 datapoints. F Mean lag-corrected correlation between hypothalamic activity and interstitial glucose variations across conventional powerband (N = 7 hypothalamic channels, one-way ANOVA; F(5,36) = 17, P < 0.001). G Lag-corrected hypothalamic HFA-glucose correlation stratified by sleep and wake states (N = 7 hypothalamic channels, one-sample T-test: t(6) = 23, P < 0.001). H Subcortical and corticolimbic lag-corrected HFA-glucose correlation stratified by sleep and wake states across three subjects. All tests were Paired t-test with statistics and P-value shown per region. All error bars indicate standard error of the mean (SEM). Source data are provided as a Source data file. *P < 0.05, **P < 0.01, ***P < 0.001.