Heritability of the olfactory bulb and its associated brain network

The olfactory bulb (OB), the first central relay of the olfactory pathway, plays a critical role in odor perception and exhibits remarkable structural plasticity shaped by environmental influences. This raises a fundamental question about the extent to which OB volume and its integration within the wider olfactory network is genetically determined. While both OB volume and function have been implicated in a range of neurodegenerative and neuropsychiatric disorders, such as Parkinson's disease, schizophrenia, and depression, all of which involve known genetic risk factors, the heritability of the OB structure remains poorly understood. Here, we investigated the heritability of OB volume and the broader olfactory network architecture in a large sample of healthy young adults (n = 941; aged 22-35 years), including monozygotic and dizygotic twin pairs. Using a deep learning-based segmentation model, we first automatically segmented OB volumes and then employed a support vector machine framework to classify zygosity based on within-pair morphological similarity. The OB volume alone showed weak classification performance in distinguishing between monozygotic and dizygotic twins (accuracy = 60%, p = .15), indicating that OB volume by itself may not strongly capture genetically driven similarity. This is consistent with the OB's high responsiveness to environmental exposures, which can lead to noticeable within-pair differences even among monozygotic twins. In contrast, combining OB volume with morphometric features from olfactory-associated brain regions, including the hippocampus, parahippocampal gyrus, entorhinal cortex, and medial orbitofrontal cortices, substantially improved classification performance (accuracy = 82%, p = .004). This effect was specific to the olfactory network, underscoring the distributed nature of genetic influence in this system. Overall, these findings indicate that the OB, despite being an environmentally responsive structure, is embedded within a genetically coordinated neural network.