Liang Zhifeng's research group and Xu Min's research team of the Center for Excellence and Innovation in Brain Science and Intelligent Technology of the Chinese Academy of Sciences have established the photogenetic fMRI (opto fMRI) experimental paradigm in conscious mice, and used this method to reveal the whole brain BOLD activation specifically modulated by four types of neurons (VGLUT2, ChAT, PV, SOM) in the basal forebrain (BF) and the corresponding behavioral preferences. In addition, this study linked BF modulated BOLD activation, low dimensional brain networks, and behavioral preferences through a decoding model, revealing the potential functional network basis of mouse behavioral preferences. On February 15th, the research findings were published in Neuron, titled Cell type specific optogenetics fMRI on basal forebrain tumors functional network basis of behavioral preference.

BF is a highly complex brain region based on anatomy and neurochemistry, playing an important regulatory role in many brain functions. Although most studies have focused on the impact of cholinergic (ChAT) neurons in BF on functional networks and behavior, most neurons in BF are non cholinergic and have regulatory effects on functional networks and behavior. Meanwhile, previous studies have mostly been limited to small-scale loop levels, with few studies directly observing and comparing the regulatory effects of different types of BF neurons on brain networks and corresponding behaviors at the whole brain scale.

This study established an opto fMRI experimental paradigm in awake mice at 9.4T. Using this method, the study characterized the whole brain BOLD response induced by photogenetics in BF cell specific neuronal activation. Furthermore, based on the previously published results of cell specific primary projections of BF, it was found that the spatial similarity between the distribution of primary projections and BOLD activation was low, indicating that cell specific BOLD activation is not only contributed by the primary projections of the four types of BF neurons. Research suggests that the secondary projection of BF origin dominates the activation of BOLD mentioned above. The study performed non negative matrix factorization (NMF) on the first and second level projection matrices containing the origin of BF, resulting in six low dimensional structural networks (i.e. NMF components). At the same time, it was found that low dimensional structural networks can well explain the BOLD activation of the four types of neurons in BF, indicating that the second level projection of BF origin mainly contributes to the aforementioned BOLD activation.

At the same time, the study used the same photogenetic stimulation paradigm as opto fMRI to conduct behavioral tests on mice under free movement conditions, and found that the activation of four types of BF neurons caused different internal/external directional behavioral preferences. Among them, the activation of VGLUT2, ChAT, and PV neurons respectively favored motor behavior, new object exploration, and grooming behavior.

The study assumes that the whole brain activation map caused by photogenetic activation and the whole brain activation map caused by mouse behavior can be reconstructed by the same NMF low dimensional brain network. Based on this assumption, a novel decoding model was constructed to predict the feature vectors of four types of behaviors in the low dimensional space of NMF using the activation map induced by BF photogenetics as the independent variable and the behavioral results caused by BF photogenetics as the dependent variable. By combining the spatial distribution map of low dimensional structural networks, a whole brain activation map induced by mouse behavior was studied and constructed. Furthermore, fMRI experiments for simultaneous behavior monitoring and calcium fiber recording experiments for free movement behavior were conducted to verify the reliability of the predicted whole brain activation maps.

This study established a clear mouse opto fMRI experimental paradigm and used this method to reveal cell type specific global BOLD activation and corresponding behavioral preferences modulated by the four types of BF neurons; A novel decoding model was constructed, linking BOLD activation within nuclear magnetic resonance with behavioral preferences outside nuclear magnetic resonance, revealing the underlying functional network foundation behind mouse behavior, providing a new perspective for exploring mouse behavior from a whole brain perspective.

The research work was funded by the Ministry of Science and Technology, the National Natural Science Foundation of China, the Chinese Academy of Sciences, Lingang Laboratory and the China Postdoctoral Science Foundation, and supported by the MRI platform of Brain Imaging Center of Brain Intelligence Excellence Center, Brain Science Data and Computing Center and the laboratory mouse room.