We propose a deep learning-based spike sorting method for extracellular recordings. For analysis of extracellular single unit activity, the process of detecting and classifying action potentials called “spike sorting” has become essential. This is achieved through distinguishing the morphological differences of the spikes from each neuron, which arises from the differences of the surrounding environment and characteristics of the neurons. However, cases of high structural similarity and noise make the task difficult. And for manual spike sorting, it requires professional knowledge along with extensive time cost and suffers from human bias. We propose a deep learning-based spike sorting method on extracellular recordings from a single electrode that is efficient, robust to noise, and accurate. In circumstances where labelled data does not exist, we created pseudo-labels through principal component analysis and K-means clustering to be used for multi-layer perceptron training and built high performing spike classification model. When tested, our model outperformed conventional methods by 2.1% on simulation data of various noise levels, by 6.0% on simulation data of various clusters count, and by 1.7% on in-vivo data. As a result, we showed that the deep learning-based classification can classify spikes from extracellular recordings, even showing high classification accuracy on spikes that are difficult even for manual classification.

Anxiety disorders, such as post-traumatic stress disorder (PTSD), are thought to occur by dysfunction in the fear and anxiety-related brain circuit, however, the exact mechanisms remain unknown. Recent human studies have shown that the right anterior insular cortex (aIC) activity is positively correlated with the severity of PTSD symptoms. Understanding the role of the aIC in fear and anxiety may provide insights into the etiology of anxiety disorders. We used a modified shock-probe defensive burying behavioral test, which utilizes the natural propensity of rodents to bury potentially dangerous objects, to test the role of aIC in fear. Mice exposed to restraint stress exhibited burying of the restrainer-resembling object, indicative of defensive behavior. Electrolytic ablation of the aIC significantly diminished this defensive burying behavior, suggesting the involvement of the aIC. Single-unit recording of pyramidal neurons in the aIC showed that a proportion of neurons which increased activity in the presence of a restrainer-resembling object was significantly correlated with the defensive burying behavior. This correlation was only present in mice exposed to restraint stress. These results suggest that altered neuronal representation in the aIC may regulate fear and anxiety after exposure to a traumatic event. Overall, our result demonstrates that the aIC mediates fear and anxiety and that it could be a potential target for treating anxiety disorders.

Impaired social abilities are characteristics of a variety of psychiatric disorders such as schizophrenia, autism spectrum disorder, and bipolar disorder. Studies consistently implicated the relationship between the anterior insular cortex (aIC) and social ability, however, how the aIC involves in processing specific subtypes of social ability was uninvestigated. We, therefore, investigated whether the absence or presence of the aIC affects the social behaviors of mice. We found that electrolytic lesions of the aIC specifically impaired mice’s ability to recognize a novel stranger mouse, while the sociability of the aIC-lesioned mice was intact. Interestingly, the aIC-lesioned mice were still distinguished between a mouse that had been housed together before the aIC lesion and a novel mouse, supporting that retrieval of social recognition memory may not involve the aIC. Additional behavioral tests revealed that this specific social ability impairment induced by the aIC lesion was not due to impairment in olfaction, learning and memory, locomotion, or anxiety levels. Together our data suggest that the aIC is specifically involved in processing social recognition memory, but not necessarily involved in retrieving it.

The ventral midline thalamus, consisting of the reuniens and rhomboid nuclei (RE/Rh), is a thalamic structure interconnected with the limbic systems including the hippocampus. Recently, many studies have revealed that this structure plays distinctive roles in spatial learning and memory in collaboration with hippocampal functions. However, what aspects of spatial information process are influenced by the RE/Rh is not clearly known. To elucidate the roles of RE/Rh in spatial information processing and its effects on hippocampal activity, specifically with the manipulation of spatial contents, we measured hippocampal-dependent spatial memory performance and hippocampal place cell activities after RE/Rh lesion using male C57BL/6J × 129/SvJae hybrid mice. We found that the lesion altered the behavioral aptitude in recognizing locational changes of an object. Furthermore, CA1 place cells in the lesion group showed different spatial representation patterns in recognizing the environment with cue locational changes compared with the control group. Interestingly, the patterns of CA1 place cells in recognizing the same environment previously visited were not disrupted in the lesion group compared with the control group. These findings demonstrate that the ventral midline thalamus (RE/Rh) is important in recognizing the spatial relationships, especially when spatial rearrangement of cue position was introduced.SIGNIFICANCE STATEMENTThe ventral midline thalamic nuclei (reuniens and rhomboid) interact with the hippocampus to influence various cognitive functions requiring spatial memories, yet what aspects of spatial information process are influenced by these nuclei is not clearly known. Here, we reveal that these nuclei play a crucial role in modulating hippocampal properties only with locational rearrangement of cues, not with the familiar arrangement. These nuclei are distinctively involved in cue-dependent spatial information processes of CA1 place cells. In particular, we suggest that these nuclei modulate spatial information processing on discrete components, especially when the spatial cue relationship is modified.

Animals seeking survival needs must be able to assess different locations of threats in their habitat. However, the neural integration of spatial and risk information essential for guiding goal-directed behavior remains poorly understood. Thus, we investigated simultaneous activities of fear-responsive basal amygdala (BA) and place-responsive dorsal hippocampus (dHPC) neurons as rats left the safe nest to search for food in an exposed space and encountered a simulated ‘predator.’ In this realistic situation, BA cells increased their firing rates and dHPC place cells decreased their spatial stability near the threat. Importantly, only those dHPC cells synchronized with the predator-responsive BA cells remapped significantly as a function of escalating risk location. Moreover, optogenetic stimulation of BA neurons was sufficient to cause spatial avoidance behavior and disrupt place fields. These results suggest a dynamic interaction of BA’s fear signalling cells and dHPC’s spatial coding cells as animals traverse safe-danger areas of their environment.