Cell . . . Zebrafish whole-brain calcium imaging records reveal cerebellum participation in cognitive coding.

Chen Wenqiang (Postdoctoral Fellow of Harvard Medical School) wrote that animals respond to their external environment through goal-directed approach, which can continuously optimize their behavioral response through learning.goal oriented behavior involves motor planning before action. Therefore, the research on the preparatory activity of neurons before motor decision-making can help us understand the neuron coding rules behind goal-directed behavior.in the past decade, it has been found that there are multiple brain regions in the neuronal population dealing with preparatory activities, such as motor cortex, pre motor cortex and parietal cortex [1-2]. These findings are of great significance for us to understand the neuronal coding of complex behaviors.ideally, neuroscientists expect to simultaneously measure the activity of the same neuron population in different experiments on a whole brain scale. However, due to the technical challenges, this scientific question has been difficult to answer.on January 16, 2020, the alipasha Vaziri research group of Rockefeller University published the title of cerebellar neurodynamics predict decision timing and outcome on the single trial online in cell magazine Level's research paper, through the use of volumetric calcium imaging technology to record the whole brain neuronal activity of zebrafish juvenile involved in the operant conditioning task, found the dynamic activity characteristics of neurons before motor selection at the whole brain level, and further found the important role of the activity of neurons in the small brain granule cells in the process, thus revealing the zebrafish small The brain encodes important functions of cognition.the researchers used juvenile zebrafish as a model animal, because zebrafish brains are transparent and highly homologous with mammals in physiology and genetics.first of all, the researchers used high-speed volumetric calcium imaging recording technology based on light field microscope to synchronously detect the activity of neural network in the whole brain at the single cell level, and combined with a new type of operating conditioned reflex paradigm based on zebrafish juveniles, tail flick to reduce aversion stimulation (release of aversive stimulus by turn, Toast, Fig. 1) to study the neurobiological mechanism of goal-directed action selection.through the learning of multiple trials, zebrafish can learn to swing tail in a specific direction to reduce the stimulation.Figure 1. Figure a shows the behavior paradigm of this paper: tail flick reduces aversion.the mild aversion stimulation (thermal stimulation) caused by infrared laser irradiation on juvenile zebrafish with fixed head can be alleviated by swinging tail in a specific direction.figure B shows the wrong tail flick direction (purple) and the correct tail flick direction (green).the toast behavior paradigm can not only study operational learning, but also provide an ideal behavior paradigm to study the mechanism of action selection and pre motor activity. This is because of the decision time between the beginning of heat stimulation and the beginning of the first tail flick animal, DT) can be used to learn the direction of tail flick and the mechanism of neuronal dynamic activity at the beginning of movement.the researchers used transgenic zebrafish labeled with calcium indicator to conduct a 700 × 700 × 200 experiment The activity of nearly 5000 active neurons was tracked by micro volume whole brain calcium imaging, and the spatial distribution of these active neurons was analyzed through synchronous thermal stimulation time points. The researchers found that these active neurons were distributed in the whole forebrain, midbrain and anterior part of the posterior brain (Fig. 2).Figure 2 After more than 4000 neurons have been successfully manipulated and learned, the linear dimension reduction based on PCA can not analyze the dynamic activity of brain state with the change of task. Therefore, researchers use the nonlinear dimensionality reduction method based on t-sne to analyze the time history of large brain state. It is found that there are six groups of neurons, not only in the neural activity network It has functional characteristics and specific anatomical characteristics (Fig. 3).Fig. 3 T-sne was used to analyze the space of neurons in the whole brain. It was found that six groups of neurons with significant functional characteristics had brain anatomical characteristics that overlapped with specific anatomical regions. Based on the functional and anatomical characteristics of these six groups of neurons, researchers proposed a hypothesis that decision-making behavior under goal-directed behavior depends on the coordination and integration of information in different regions of the whole brain, However, there is no experimental evidence for this hypothesis.the researchers conducted task-dependent event-related analysis of the neuronal activity of the anterior motor, and found that the correlation value of the neuronal activity in the ipsilateral cerebellum and the activity of the ipsilateral cerebellum habenular nucleus were highly correlated with the accuracy of the task.then, the correlation between preparatory activity and motor response was analyzed by mixed principal component analysis. The researchers found that each neuron group can encode different motor related information, and the fourth group code can participate in sensory motor transformation, indicating that cerebellum has the strongest prediction signal for decision direction (Fig. 4). Figure 4 Group 4, composed of cerebellum and Artr, has the strongest prediction signal of decision direction. In order to detect the contribution of major cell types (Purkinje cells, granulosa cells and eurydendroid cells) in cerebellum, researchers expressed calcium indicator in transgenic zebrafish driven by promoters of different cell types, and found that cerebellar granule cells showed the strongest anterior motor activity (Fig. 5a) This is consistent with the results of cell and nature in mammals [3-4] (both reported by the lockley group of Stanford University). A. cerebellar granule cells showed the strongest anterior motor activity. A new model of sensory information integration, motor learning, motor decision-making and motor selection in cerebellum and other brain regions. so far, the whole brain volume calcium imaging recording of zebrafish juveniles not only supports the traditional view of cerebellum in motor control, but also indicates the role of cerebellum in cognitive function. the researchers believe that the presence of cerebellar preparatory activity suggests that intrinsic parameters in the cerebellum can participate in motor decision-making. at the same time, our findings also provide a new model of sensory information integration, motor learning, motor decision-making and motor selection in the whole brain (Fig. 5b). it is worth mentioning that due to the lack of cortical structure in zebrafish brain, this conclusion should be extended to mammalian. in addition, combined with a science article on cerebellum and spatial cognition shared by the author on the journal club in 2011, interested readers can pay attention to the work of mouse cerebellar structure encoding spatial information of hippocampal location cells. similarly, the circuit from L5 layer cells of neocortex to cerebellum reported by luoliqun research group in cell magazine [3] in 2019 is also worth reading (see bioart report: cell | luoliqun et al. Used dual calcium imaging to record the dynamic changes of neurons in different brain regions during learning process). if readers are interested in the plasticity of granulocytes in motor learning, they can extend their reading of the review published in 2012 by Nature Reviews Neuroscience [6]. original link: references 1. Gao, Z., et al. (2018). A cortico cerebellar loop for motor planning. Nature 563, 113 – 116.2. Li, N., et al. (2015). A motor cortex circuit for motor planning and movement. Nature 519, 51 – 56.3. Wagner, M.J, et al. (2019). Shared Cortex-Cerebellum Dynamics in the Execution and Learning of a Motor Task. Cell 177, 669–682.4. Wagner, M.J., et al. (2017). Cerebellar granule cells encode the expectation of reward. Nature 544, 96–100.5. Rochefort, C., et al. (2011) Cerebellum shapes hippocampal spatial code. Science 334, 385-389.6. Gao, Z., et al., (2012) Distributed synergistic plasticity and cerebellar learning. Nat Rev Neurosci. 13, 619-635.