Activity of medial prefrontal cortex excitatory neurons as the basis of a murine neurofeedback model for the modulation of defensive states

Emotional defensive states are elicited upon threat in order to orchestrate a response to thrive for survival. They encompass multiple physiological responses, such as autonomic, behavioral and hormonal changes, and depend on contextual, mnemonic and external and internal sensory cues. Multiple interconnected brain regions have been involved in the generation of emotional defensive states, such as the amygdala, the periaqueductal gray or the prefrontal cortex (PFC), which plays a major role in their orchestration and selection. Specifically, the medial PFC integrates internal and external cues, as well as previous experiences through mnemonic information, in order to select an appropriate response to the threat, and to send that information forward to lower-order brain regions to elicit said responses. The same mechanisms underlying the generation of emotional defensive states can become maladaptive, that is out-of-focus or out-of-proportion, becoming the basis for certain mental disorders, such as anxiety disorders. While several therapeutic tools are available to treat said disorders, such as cognitive-behavioral therapy or pharmacological treatments, they are unspecific and their efficacy is only partial. For these reasons, many circuit-therapy approaches are being developed, whose aim is to manipulate the maladaptive activity of the specific circuits underlying those disorders. One novel circuit-therapy approach is Neurofeedback (NFB), in which online feedback of the subject’s own neural activity is presented to allow self-regulation. Through this procedure, the neural substrates underlying an specific disorder are altered, consequently altering the maladaptive state. While NFB has had some clinical successes, there are multiple questions surrounding its implementation, such as why many patients are unable to self-regulate their own activity, or how exactly is that self-regulation taking place. Therefore, our aim was to use mPFC activity, which we recorded using population activity fiber photometry after delivery of a viral vector to express GCaMP in CaMKII+ neurons, as the basis for developing a murine NFB model to modulate defensive states. This would allow us to better characterize the basis of NFB training and which neural elements are better suited for the modulation of maladaptive defensive states. While both control and experimental mice managed to learned the paradigm and consistently performed successful trials during NFB training, we did not observe any modulation of the underlying mPFC activity, nor in the behavioral testing done before and after NFB training. In order to elucidate why our NFB protocol did not obtain the desired behavioral results we performed one-photon microendoscopy recordings of the same neuronal population, with concomitant ECG and behavioral recordings across multiple anxiety-related paradigms. We observed that the activity of different mPFC subpopulations correlated with specific cardio-behavioral microstates, such as defensive freezing or grooming; with the valence of CS-US paring across a modified fear conditioning; as well as with the threat level of different parts of the EPM. Finally, we observed a similar macrostate modulation on range of possible correlates that cells could have across recording time. In summary, we conclude that mPFC population activity recorded through fiber photometry was too heterogeneous to allow NFB training, since many subpopulation vectors could be identify within it, which correlated with variate defensive-state elements.