Neural energy coding patterns of dopaminergic neural microcircuit and its impairment in major depressive disorder: A computational study

Numerous experiments have found that the behavioral characteristics of major depressive disorder (MDD) animals are usually associated with abnormal neural activity patterns and brain energy metabolism. However, the relationship among the behavioral characteristics, neural activity patterns and brain energy metabolism remains unknown. In this paper, we computationally investigated this relationship, with a particular focus on how neural energy coding patterns change in MDD brains, in the VTA-NAc-mPFC dopaminergic pathway of the reward system based on our biological neural network model and neural energy calculation model. Interestingly, our results suggested that the neural energy consumption of the whole VTA-NAc-mPFC microcircuit in MDD group was significantly reduced, which was mainly attributed to the decreasing neural energy consumption in the mPFC region. This observation theoretically supported the view of low-level energy consumption in MDD. We also investigated the neural energy consumption patterns of various neuronal types in our VTA-NAc-mPFC microcircuit under the influence of different dopamine concentrations, and found that there were some specific impairments in MDD, which provided some potential biomarkers for MDD diagnosis. More specifically, we found that the actual neural energy consumption of medium spiny neurons (MSNs) in the NAc region was increased in the MDD group, whereas pyramidal neurons in the mPFC region exhibited higher actual neural energy consumption in the NC group. Additionally, in both neuron types, the actual neural energy required to generate an action potential was higher in the MDD group, suggesting that, given the same energy budget, these neurons in the MDD group tended to generate fewer action potentials. To further explore the relationship between neural coding patterns and neural energy coding patterns in the VTA-NAc-mPFC microcircuit, we in addition calculated P-V correlation for each neuronal type, defined as the Pearson's correlation coefficient between membrane potential and neural power. The results showed that the membrane potential and neural power were not perfectly correlated (P-V correlations ranged from 0.6 to 0.9), and dopamine concentration inputs affected the P-V correlations of the MSN, pyramidal neurons and CB interneurons in the mPFC region. These findings suggested that the joint application of the neural coding theory and neural energy coding theory will be superior to the application of any single theory, and this joint application could help discover new mechanisms in neurocircuits of MDD. Overall, our study not only uncovered the neural energy coding patterns for the VTA-NAc-mPFC neural microcircuit, but also presented a novel pipeline for the study of MDD based on the neural coding theory and neural energy coding theory.