Neuronal dynamics of the anterior cingulate cortex during working memory and serotonergic manipulation

Psilocybin, a serotonergic agonist, shifts the anterior cingulate cortex (ACC) of awake mice into a desynchronized state resembling REM sleep, marked by increased network activation, reduced low-frequency oscillations (delta, theta, alpha), and a moderate rise in gamma power. In a working memory delayed response task, ACC neurons encoded both low-level stimulus information and high-level choice and reward anticipation, with population activity predicting task outcomes faster than behavioral responses. Optogenetic perturbation of ACC neurons during the delay period effectively altered behavior and disrupted neural encoding of working memory information.

Study at a glance

Abstract

The prefrontal cortex (PFC) plays a strong role in the execution of executive tasks such as working memory. It is thought to exert top-down control over sensory regions, and integrate information from both external stimuli, and internal computation, such as choice and expectation (1-3). The anterior cingulate cortex (ACC) is a region of the PFC known to receive strong projections from serotonergic neurons of the raphe nuclei (4). In the work presented here, the influence of the hallucinogenic compound, and serotonergic agonist, psilocybin, was investigated in the neural circuits of the ACC of awake, head-fixed mice. Extracellular, multi-unit recordings were performed to examine the effects of the systemic administration of psilocybin. Psilocybin was found to transition the network into a desynchronised state, reminiscent of REM sleep. This state was characterised by an increase in activation in the network, combined with a reduction in power of low frequency LFP oscillations in the delta, theta and alpha band, coupled with a moderate increase in power in the gamma frequency band. The encoding mechanisms in the ACC in working memory were examined by conducting similar recordings in mice performing a delayed response task. The neuronal dynamics revealed strong encoding of low-level (stimulus information) and high-level (choice, reward anticipation) information on both a single cell and population level. The population code appeared to have strong predictive properties of task outcome that surpassed the speed at which the behavioural information could correctly classify the upcoming behavioural response. An effective method to test these hypotheses was provided in the optogenetics experiments, which perturbed the neuronal activity during the delay period of the task using channel-rhodopsin 2 (ChR2). This method proved effective and versatile in inducing a behavioural response and perturbing the neural response profiles of cells encoding information key to working memory.