Decoding neurotransmitter and genetic contributions to abnormal neuronal signal variability in Anti‑N‑Methyl‑D‑Aspartate receptor encephalitis: Implications for targeted therapies

The molecular mechanisms linking brain function alterations to gene expression in anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis remain unclear. We analyzed the coefficient of variation of blood oxygenation level dependent signal (CV BOLD ) and functional connectivity (FC) in 30 healthy controls and 42 patients, with classification via 5 machine learning models. Transcriptomic profiles from the Allen Human Brain Atlas and neurotransmitter density maps from positron emission tomography were integrated. Partial least squares (PLS) regression determined gene expression relevant to the CV BOLD /FC changes. Multivariate linear regression evaluated neurotransmitter contributions. Anti-NMDAR encephalitis patients exhibited increased CV BOLD in the right superior parietal gyrus, right fusiform gyrus, right lingual gyrus, left fusiform gyrus and left paracentral lobule meanwhile disrupted FC mainly in default mode and salience networks. PLS analysis revealed 2,320 genes significantly associated with CV BOLD /FC ( p bonferrni < 0.05), enriched in synaptic signaling (MAPK, cAMP), metabolic regulation (insulin resistance), and neurodegeneration pathways. Hub genes PPARGC1A (positive correlation with CV BOLD /FC) and UBA52 (negative correlation) were validated in key brain regions. Neurotransmitter analysis showed norepinephrine (NAT) strongly contributed to CV BOLD (weight = 0.57, p FDR < 0.001), meanwhile serotonin (5HT4), cannabinoid (CB1), noradrenaline (NAT), and glutamate (NMDA) influenced FC. This study identifies a transcriptional signature that is spatially associated with CV BOLD /FC abnormalities and neurotransmitter distributions in anti-NMDAR encephalitis, thereby generating hypotheses about molecular targets that may be relevant for future mechanistic studies and precision medicine. • Anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis patients showed increased the coefficient of variation of the blood oxygenation level dependent signal (CV BOLD ) in specific brain regions and disrupted functional connectivity (FC) in default mode and salience networks. • Using five machine learning models based on regionally significant CV BOLD and FC values to assess discriminative ability for anti-NMDAR encephalitis and found the higher classification accuracy localized to the left paracentral lobule. • The 2,320 genes linked to CV BOLD /FC changes enriched in synaptic, metabolic, and neurodegenerative pathways. • Hub genes PPARGC1A (positive correlated with CV BOLD /FC) and UBA52 (negatively correlated) validated in key brain regions. • Norepinephrine (NAT) strongly influenced CV BOLD , while serotonin (5HT4), cannabinoid (CB1), norepinephrine (NAT), and glutamate (NMDA) revealing neurotransmitter imbalances in disease pathogenesis.